Clinical Research Coordinators

11-9121.01
Median wage $167,220/yr108,690 employed (US)Rank #399 of 923 scored · top 43% by substitution

Plan, direct, or coordinate clinical research projects. Direct the activities of workers engaged in clinical research projects to ensure compliance with protocols and overall clinical objectives. May evaluate and analyze clinical data.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution30
Exposure30
Augmentation63

Substitution — the headline: capability discounted by cost, barriers and adoption.

Exposure — technical capability alone, regardless of whether anyone deploys it.

Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.

Tasks on the substitution scale

33 rated tasks, binned by substitution score.

Position among all scored occupations

Distribution of 923 occupation scores; the marker is this occupation.

Tasks with substitution ≥ 70

0%

Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.

Why this score

The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.

Task automatabilityw 35%30

panel mean rating 2.2/5 → substitution pressure 30/100

Technical feasibility todayw 20%29

panel mean rating 2.2/5 → substitution pressure 29/100

Cost vs. human wagew 15%32

panel mean rating 2.3/5 → substitution pressure 32/100

Adoption barriersw 20%inverted — strong barriers lower the score29

panel mean rating 3.8/5 (barrier strength) → substitution pressure 29/100

Sector adoption velocityw 10%25

panel mean rating 2.0/5 → substitution pressure 25/100

Task breakdown (33 tasks)

Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.

Code, evaluate, or interpret collected study data.

65

CI 45–85 · exposure 70 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Clinical research organizations show moderate adoption of automated data coding and analysis tools, with pilots and partial deployments common in pharma and CROs. Full end-to-end automation of the coding and interpretation step remains slower than in non-regulated sectors due to audit and compliance requirements.
Sector adoption velocityclaude-sonnet-53/5Clinical research and pharma sectors are adopting AI tools for data management and analytics but remain cautious due to regulatory scrutiny, with pilots more common than full production deployment for interpretive tasks.
Augmentation potentialclaude-haiku-4-5-202510014/5AI strongly assists coordinators by pre-coding data, flagging anomalies, and generating initial interpretations, allowing human review and QA to focus on complex or ambiguous cases. This augmentation significantly raises coordinator productivity while preserving critical human oversight.
Augmentation potentialclaude-sonnet-54/5AI tools significantly speed up coding of adverse events, identification of data anomalies, and preliminary statistical summaries, meaningfully boosting coordinator productivity while humans retain final interpretive responsibility.
Task automatabilityclaude-haiku-4-5-202510015/5Modern AI systems can automatically code, categorize, and interpret structured clinical data (lab values, demographics, outcomes) with high accuracy and at least 50% time savings compared to manual review. Evaluation and basic interpretation of coded data against predefined protocols falls squarely within current AI capability for routine, standardized datasets.
Task automatabilityclaude-sonnet-53/5AI can assist with coding categorical data, running statistical analyses, and flagging patterns, but interpretation requiring clinical judgment and protocol-specific nuance still needs human oversight, so only partial automation meets the 50% threshold.
Adoption barriersclaude-haiku-4-5-202510012/5Few regulatory barriers specifically prevent AI automation of data coding and interpretation itself; oversight and human review are standard practice but not prohibitive. Data privacy (HIPAA, GDPR) requires safeguards on deployment, but these are implementable rather than blocking adoption.
Adoption barriersclaude-sonnet-54/5Clinical trial data integrity is heavily regulated (FDA 21 CFR Part 11, GCP), requiring auditable human sign-off on data interpretation, and errors carry significant liability and regulatory risk, creating strong barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510015/5API-based AI data coding and interpretation costs a fraction of clinical coordinator labor (typically $35–50k/year salary). Inference and validation overhead per record is negligible compared to loaded human cost, yielding at least 10:1 cost advantage.
Cost vs. human wageclaude-sonnet-53/5AI-assisted coding and analysis tools reduce time spent on data cleaning and preliminary analysis, but the need for validation, audit trails, and human review of interpretations keeps overall costs closer to comparable rather than order-of-magnitude cheaper.
Technical feasibility todayclaude-haiku-4-5-202510014/5Deployed products (e.g., clinical data validation platforms, automated coding engines, statistical analysis tools) reliably handle standardized data coding and interpretation in production settings. Minor gaps remain in handling ambiguous or unstructured narrative data, but the core task is production-ready at scale.
Technical feasibility todayclaude-sonnet-53/5Statistical software and NLP tools for coding adverse events or extracting structured data from clinical notes exist and are used in production, but full autonomous interpretation of study data is not yet standard practice at scale.

Participate in preparation and management of research budgets and monetary disbursements.

51

CI 30–71 · exposure 50 · augmentation 88 · importance 3.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Research institutions and clinical centers are adopting accounting automation and AI-driven expense management at a moderate pace, with pilots common in larger academic medical centers but slower uptake in smaller practices. The sector digitizes steadily but lags pure-information industries.
Sector adoption velocityclaude-sonnet-52/5Clinical research administration is a moderately digitized but compliance-heavy sector where AI adoption for financial/administrative tasks remains in early pilot stages rather than widespread production use.
Augmentation potentialclaude-haiku-4-5-202510015/5AI assistants for budget templates, real-time expense categorization, forecast modeling, and anomaly detection substantially augment coordinator productivity, allowing them to focus on compliance, strategic planning, and exception handling while the AI handles routine data processing and report assembly.
Augmentation potentialclaude-sonnet-54/5AI-powered spreadsheet tools, budget templates, and financial forecasting assistants can meaningfully speed up drafting and error-checking of research budgets while the coordinator retains final responsibility.
Task automatabilityclaude-haiku-4-5-202510014/5Budget preparation and monetary disbursement management involve structured data entry, calculations, reconciliation, and document generation—tasks well-suited to current AI systems with accounting software integration. While human judgment on cost allocation and budget strategy remains valuable, the execution of disbursement workflows, expense tracking, and budget report generation can achieve >50% time savings with existing automation tools.
Task automatabilityclaude-sonnet-52/5Budget preparation involves judgment about study-specific cost drivers, negotiation with sponsors, and institutional policy compliance that current AI cannot fully replicate end-to-end, though spreadsheet and drafting portions are automatable.
Adoption barriersclaude-haiku-4-5-202510013/5Institutional policies, audit trails, and segregation-of-duties controls require human oversight of disbursements and budget changes; many organizations mandate coordinator sign-off or approval chains that limit full automation. Regulatory requirements (FCOI, grant compliance) create friction but do not absolutely prohibit AI-assisted workflows.
Adoption barriersclaude-sonnet-53/5Financial disbursement and budget sign-off typically require institutional and grants-management authorization, creating moderate procedural and compliance barriers, though not strict professional licensure requirements.
Cost vs. human wageclaude-haiku-4-5-202510015/5Cloud-based accounting automation, invoice processing, and disbursement systems cost a fraction of loaded coordinator wages ($50–70k annually). Per-task AI inference and integration overhead for budget management is negligible compared to human labor for data entry, reconciliation, and report generation.
Cost vs. human wageclaude-sonnet-52/5AI can reduce time spent on repetitive calculations but still requires substantial human oversight for compliance, sponsor negotiation, and disbursement approval, so total cost savings versus a coordinator's wage are modest.
Technical feasibility todayclaude-haiku-4-5-202510014/5Mature accounting and research management software (e.g., REDCap, SAP, QuickBooks) coupled with AI-driven invoice processing and expense categorization are deployed in production across healthcare and research institutions. These systems reliably handle routine budget transactions, though complex financial decisions and audit-sensitive allocations still require human verification.
Technical feasibility todayclaude-sonnet-52/5AI tools (e.g., budgeting spreadsheet templates, financial software with AI features) can assist with calculations and forecasting, but no deployed product manages clinical trial budget preparation and disbursement autonomously in production.

Maintain contact with sponsors to schedule and coordinate site visits or to answer questions about issues such as incomplete data.

45

CI 30–60 · exposure 45 · augmentation 75 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research organizations have been slow to automate coordinator functions compared to information-sector adoption, with pilots on scheduling common but production agent deployment rare. Legacy systems, regulatory conservatism, and sponsor relationship norms inhibit rapid uptake.
Sector adoption velocityclaude-sonnet-52/5Clinical research operations are moderately digitized but still rely heavily on human coordinators for sponsor communications, with AI adoption mostly in pilot stages for data management, not communication tasks.
Augmentation potentialclaude-haiku-4-5-202510014/5AI assistants can significantly boost coordinator productivity by drafting responses, auto-flagging incomplete data fields, suggesting optimal visit windows, and pre-populating sponsor correspondence templates, allowing the human coordinator to focus on relationship management and complex issue resolution.
Augmentation potentialclaude-sonnet-54/5AI can significantly assist with scheduling, drafting responses, summarizing data discrepancies, and tracking communications, boosting coordinator productivity while humans remain the primary contact.
Task automatabilityclaude-haiku-4-5-202510014/5Current AI systems can autonomously schedule meetings via calendar integration, send templated follow-up emails, flag incomplete data in study databases, and route queries to appropriate personnel, achieving >50% time savings on routine coordination. The task involves structured communication and data management, which are well within AI capability, though complex sponsor negotiations may still require human judgment.
Task automatabilityclaude-sonnet-52/5This involves relational communication, negotiation of schedules, and clarifying ambiguous data issues that require judgment and rapport-building with sponsors, limiting full automation.; scheduling logistics alone could be automated but the substantive Q&A cannot.
Adoption barriersclaude-haiku-4-5-202510013/5Regulatory frameworks (FDA, GCP) and sponsor contracts may require human sign-off on protocol-related communications and data discrepancies, and institutional policies often mandate human review of sponsor responses. However, routine scheduling and data flagging face less regulatory friction than clinical assessments.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement, but sponsor relationships, trust, and regulatory documentation standards (GCP) create organizational friction against full delegation to AI.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI-driven scheduling and email routing cost pennies per interaction, while a coordinator's loaded wage (salary + benefits + overhead) for equivalent contact management work runs tens of dollars per hour, representing a cost advantage of 10–100x for routine tasks.
Cost vs. human wageclaude-sonnet-52/5Scheduling tools are cheap, but resolving data queries with sponsors still requires coordinator time and judgment, so overall cost savings are modest, not order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510013/5Email automation and calendar scheduling tools are deployed at scale, but end-to-end contact management with institutional context (study protocols, sponsor preferences, data integrity issues) typically requires human oversight in production clinical research settings. Products exist but material error rates in data interpretation or missing nuanced sponsor relationship dynamics limit full autonomy.
Technical feasibility todayclaude-sonnet-52/5AI scheduling assistants exist and are used in some clinical operations, but answering sponsor questions about incomplete data requires contextual judgment that current products don't reliably handle in production.

Prepare study-related documentation, such as protocol worksheets, procedural manuals, adverse event reports, institutional review board documents, or progress reports.

44

CI 43–46 · exposure 50 · augmentation 75 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research remains moderately digitized with strong human-centric oversight cultures. Adoption of AI for documentation is in pilot phases at best; most sites still rely on human coordinators due to regulatory caution and institutional inertia.
Sector adoption velocityclaude-sonnet-52/5Clinical research/pharma is a heavily regulated, cautious sector; AI adoption for documentation is emerging via pilots but production use for compliance-critical documents remains limited.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can significantly augment coordinators by drafting templates, auto-populating structured fields, and generating first-pass report sections from clinical data, enabling humans to focus on review, compliance, and judgment-heavy elements of documentation.
Augmentation potentialclaude-sonnet-54/5AI is well-suited to drafting boilerplate sections, summarizing data into progress reports, and generating first drafts of manuals or worksheets, meaningfully speeding up coordinators' work while they retain final review responsibility.
Task automatabilityclaude-haiku-4-5-202510013/5AI can assist with generating template-based documentation, filling in structured fields, and drafting sections of reports from existing data or protocols. However, substantial human oversight is needed for accuracy, regulatory compliance, and domain-specific content that requires clinical judgment.
Task automatabilityclaude-sonnet-53/5AI can draft protocol worksheets, progress reports, and templated IRB documents from source data, but adverse event reports and final regulatory submissions require careful verification, clinical judgment, and compliance checks that limit full automation today.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory and institutional barriers are substantial: IRB submissions require authorized human signature and institutional review; adverse event reports have legal/safety implications requiring human clinical judgment; and many organizations have formal approval workflows that legally mandate human responsibility.
Adoption barriersclaude-sonnet-54/5Regulatory frameworks (FDA, GCP, IRB requirements) mandate qualified personnel to prepare and attest to accuracy of adverse event and protocol documentation, creating substantial compliance and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI inference costs for document generation are very low compared to the loaded wage of a clinical research coordinator (typically $40k–$60k+ annually). Even with overhead for integration and human review, the per-task cost ratio is favorable.
Cost vs. human wageclaude-sonnet-53/5AI drafting can cut time on templated documents significantly, but required human review, validation against source data, and regulatory sign-off keep overall cost savings moderate rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510013/5Products exist (LLMs, document automation tools) that can draft protocol worksheets and progress reports with reasonable accuracy, but material limitations remain in handling complex regulatory requirements (IRB forms) and ensuring consistency with institutional standards. Deployed use is growing but not yet production-standard in most organizations.
Technical feasibility todayclaude-sonnet-53/5Clinical documentation tools and LLM-based drafting assistants exist and are used in research settings, but adoption for regulated deliverables like IRB submissions and AE reports remains narrow due to accuracy and compliance concerns.

Review scientific literature, participate in continuing education activities, or attend conferences and seminars to maintain current knowledge of clinical studies affairs and issues.

44

CI 30–59 · exposure 42 · augmentation 75 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research remains a regulated, human-oversight-heavy sector with slow digitization of knowledge management. While some organizations use AI-assisted literature monitoring, widespread autonomous AI-driven knowledge curation in clinical trials is not yet standard practice.
Sector adoption velocityclaude-sonnet-53/5Clinical research and life sciences are moderately fast adopters of AI literature tools, but formal CE and conference attendance requirements remain largely unchanged by AI adoption trends.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered literature search, summarization, and alerting systems significantly augment a coordinator's ability to stay current on emerging issues and studies. LLMs and specialized tools can reduce time spent scanning journals while the human retains judgment on relevance and compliance implications.
Augmentation potentialclaude-sonnet-54/5AI significantly augments this task by rapidly filtering, summarizing, and surfacing relevant literature and trends, letting coordinators focus their limited time more effectively.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with literature retrieval and summarization, but cannot independently assess relevance to specific study contexts, evaluate source credibility across domains, or synthesize findings into actionable clinical knowledge. Human judgment on what constitutes current and applicable knowledge remains essential.
Task automatabilityclaude-sonnet-53/5AI can efficiently summarize and synthesize scientific literature and flag relevant new publications, saving substantial time on the literature review component, but attending conferences/seminars and continuing education activities are not automatable end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research is heavily regulated (ICH-GCP, FDA oversight). Coordinators must understand current regulatory guidance and study-specific protocols; relying on AI alone for maintaining knowledge of compliance-critical affairs creates audit and liability risk. Human oversight and institutional responsibility for knowledge currency are expected.
Adoption barriersclaude-sonnet-52/5No licensure is required for literature review or CE activities, but professional certification bodies may mandate specific accredited CE formats or attendance that AI cannot substitute for.
Cost vs. human wageclaude-haiku-4-5-202510012/5Literature review and knowledge-maintenance tasks are typically brief relative to a coordinator's day. AI tools (subscriptions, inference) cost is reasonable but not dramatically cheaper than allocating a coordinator's time to structured professional development, which is often employer-provided or mandated.
Cost vs. human wageclaude-sonnet-54/5AI-assisted literature scanning and summarization tools are dramatically cheaper than a coordinator manually reading and tracking journals, though human oversight is still needed for content quality.
Technical feasibility todayclaude-haiku-4-5-202510013/5LLMs and search tools can summarize recent papers and identify relevant studies, but production systems have documented limitations in comprehension of nuanced clinical content and liability concerns around independent knowledge curation for regulated research environments. Current products support the task incompletely.
Technical feasibility todayclaude-sonnet-53/5AI literature-summarization and alerting tools (e.g., PubMed AI assistants, research copilots) are deployed and used by professionals today, though they still require human verification for accuracy and clinical relevance.

Register protocol patients with appropriate statistical centers as required.

44

CI 25–62 · exposure 45 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research organizations remain heavily regulated and conservative in automating patient-facing administrative tasks; adoption of AI for patient registration is minimal, with most institutions preferring trained human coordinators to manage this compliance-critical function.
Sector adoption velocityclaude-sonnet-53/5Clinical research/pharma sectors are adopting digital trial management and automation steadily, but many sites still rely on manual entry and paper-to-digital transcription, keeping adoption moderate rather than fast.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by auto-populating forms from existing medical records and flagging eligibility issues, meaningfully reducing coordinator time on data entry and preliminary checks, while the coordinator retains decision authority over final registration.
Augmentation potentialclaude-sonnet-54/5AI-assisted data entry, eligibility checks, and auto-population of registration fields can significantly speed up coordinators' registration workflow while they retain oversight for compliance and accuracy.
Task automatabilityclaude-haiku-4-5-202510012/5Registration involves substantial data entry and form completion with clear requirements, but the task requires verifying patient eligibility against complex protocol criteria and handling exceptions that depend on contextual judgment about individual patient circumstances.
Task automatabilityclaude-sonnet-54/5This is a structured data-entry and submission task involving entering patient/protocol data into standardized registry systems, which is highly amenable to automation via forms-based AI agents or RPA integrated with EDC systems.rating reflects near-complete automatability with modest setup for validation rules.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical trial protocols and statistical centers typically require authorized human personnel to register patients; liability for incorrect patient enrollment is high, and regulatory frameworks (FDA, IRB) expect documented human accountability in patient registration processes.
Adoption barriersclaude-sonnet-53/5While registration itself isn't inherently requiring licensure, regulatory frameworks (GCP, IRB oversight) and sponsor requirements for accurate, auditable protocol enrollment create moderate procedural and compliance friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI systems require significant human oversight and error-correction for regulatory-sensitive patient registration, making total cost (including human review) comparable to or potentially higher than direct human performance.
Cost vs. human wageclaude-sonnet-54/5Automated registration via existing CTMS/EDC integrations is far cheaper per-transaction than coordinator time, though initial system setup and validation add cost, moderating full order-of-magnitude savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5While basic form-filling and data entry automation exist, deployed products rarely handle the nuanced eligibility verification and statistical center communication required for protocol patient registration at acceptable error rates.
Technical feasibility todayclaude-sonnet-53/5Clinical trial management systems and eRegulatory/EDC platforms already automate much of patient registration, but integration varies by sponsor/CRO systems and human verification is still commonly required to catch eligibility or data errors.

Dispense medical devices or drugs, and calculate dosages and provide instructions as necessary.

42

CI 6–78 · exposure 45 · augmentation 75 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Dosage automation and decision-support tools are already widely deployed in hospitals and large health systems; EHR integration is standard in information-dense healthcare settings, showing fast adoption in digitized institutional care, though slower in smaller practices.
Sector adoption velocityclaude-sonnet-52/5Clinical research and healthcare settings adopt AI cautiously for documentation and calculations, but direct dispensing tasks see minimal automation due to regulatory and safety constraints.
Augmentation potentialclaude-haiku-4-5-202510015/5AI dosage calculators and decision-support systems substantially augment coordinator productivity by automating lookup, reducing manual calculation errors, and flagging drug interactions, allowing coordinators to focus on patient communication and exception handling.
Augmentation potentialclaude-sonnet-53/5AI-assisted dosage calculators and clinical decision support tools can help coordinators verify calculations and reduce errors, though the human retains full responsibility for dispensing.
Task automatabilityclaude-haiku-4-5-202510015/5Calculating dosages and providing administration instructions for drugs and devices is a well-defined algorithmic task with deterministic inputs (patient weight, renal/hepatic function, drug interactions) and outputs (dose, timing, route). Current AI systems can reliably automate this with significant time savings, provided data is structured.
Task automatabilityclaude-sonnet-51/5This task requires hands-on dispensing of drugs/devices to patients and direct dosage calculation tied to physical administration, which current AI cannot perform end-to-end without physical embodiment or legal authority.
Adoption barriersclaude-haiku-4-5-202510014/5Significant regulatory and liability barriers exist: FDA oversight of decision-support accuracy, state pharmacy board rules requiring licensed pharmacist review/approval, malpractice liability for dosing errors, and institutional credentialing requirements mean human sign-off remains legally mandated in most jurisdictions.
Adoption barriersclaude-sonnet-55/5Dispensing drugs and devices in clinical trials is heavily regulated (FDA, IRB, GCP) and requires licensed, authorized personnel with liability accountability, making substitution legally prohibited.
Cost vs. human wageclaude-haiku-4-5-202510015/5AI inference and integration costs for dosage calculation and instruction generation are negligible (cents per calculation) compared to loaded clinical coordinator wages ($25–35/hour); automated systems pay for themselves within days.
Cost vs. human wageclaude-sonnet-52/5While dosage calculation software is cheap, the overall task requires licensed personnel to physically dispense and verify, so AI cannot substitute for the bulk of the labor cost.
Technical feasibility todayclaude-haiku-4-5-202510014/5Multiple production systems (pharmacy dispensing software, clinical decision support, EHR-integrated dosing calculators) perform dosage calculation and instruction generation reliably today. However, physical dispensing still requires human handling in most regulated settings, limiting full end-to-end automation to ~80% of the task scope.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product autonomously dispenses medications or devices to research subjects; this remains a human-performed clinical action with only decision-support tools available.

Develop advertising and other informational materials to be used in subject recruitment.

41

CI 30–51 · exposure 42 · augmentation 75 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research remains a regulated, risk-averse sector with slow AI adoption; institutions prioritize human accountability and legal defensibility over speed, and most coordinators still generate materials with human-centric workflows rather than AI agents.
Sector adoption velocityclaude-sonnet-52/5Clinical research and healthcare administration sectors are slower AI adopters overall due to compliance concerns, though marketing-adjacent tasks see more pilot use than core clinical work.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist by drafting initial copy, suggesting design layouts, and generating multiple message variants, enabling coordinators to iterate faster and focus effort on regulatory validation and audience refinement rather than starting from blank pages.
Augmentation potentialclaude-sonnet-54/5AI writing tools substantially speed up drafting of recruitment materials, letting coordinators generate initial versions and iterate faster while still reviewing for accuracy and compliance.
Task automatabilityclaude-haiku-4-5-202510012/5AI can generate draft copy and design concepts for recruitment materials, but the task requires regulatory compliance (FDA, IRB approval), medical accuracy verification, and tone calibration specific to target populations—elements that demand substantial human oversight and revision rather than end-to-end automation.
Task automatabilityclaude-sonnet-53/5AI can draft recruitment flyers, ads, and informational copy quickly, but compliance review, IRB approval language, and tailoring to specific protocols require human judgment, limiting full end-to-end automation.
Adoption barriersclaude-haiku-4-5-202510014/5Institutional Review Boards (IRBs) and regulatory bodies impose strict requirements on recruitment materials; organizations typically require human clinical research professionals to certify accuracy and compliance, creating a legal and procedural barrier to full automation.
Adoption barriersclaude-sonnet-53/5IRB/ethics review and regulatory compliance (e.g., FDA, HIPAA-adjacent language rules) require human sign-off, creating moderate friction even though drafting itself isn't restricted.
Cost vs. human wageclaude-haiku-4-5-202510012/5While generative AI reduces drafting time, the regulatory review, legal vetting, and iterative refinement required for clinical materials mean the all-in cost (inference + human oversight + compliance checking) remains close to or exceeds the cost of a coordinator creating materials from scratch.
Cost vs. human wageclaude-sonnet-54/5Drafting copy via LLMs is far cheaper than coordinator time spent writing from scratch, though human review and compliance checks still add cost.
Technical feasibility todayclaude-haiku-4-5-202510013/5AI writing and design tools can produce recruitment material templates, but no deployed system reliably handles the full pipeline of regulatory correctness, medical claims vetting, and IRB-compliant language without material manual intervention by experienced coordinators.
Technical feasibility todayclaude-sonnet-53/5Generative AI writing tools are commonly used to draft marketing/informational content, but no specialized deployed product handles clinical trial recruitment materials end-to-end reliably with regulatory compliance built in.

Schedule subjects for appointments, procedures, or inpatient stays as required by study protocols.

39

CI 30–47 · exposure 33 · augmentation 63 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research environments are traditionally conservative, digitization is uneven across sites, and adoption of AI agents in production remains pilot-stage. Most research coordinators still use manual calendar management with email and phone coordination.
Sector adoption velocityclaude-sonnet-52/5Clinical research sites are historically slow adopters of workflow automation due to regulatory environment, fragmented systems, and reliance on manual coordination despite growing CTMS digitization.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by suggesting available slots, sending automated reminders, flagging scheduling conflicts, and surfacing protocol constraints, meaningfully improving coordinator efficiency while the human retains control over subject communication and protocol compliance decisions.
Augmentation potentialclaude-sonnet-54/5AI-driven scheduling assistants and automated reminders substantially reduce administrative burden and errors, letting coordinators focus on complex protocol and patient-specific needs.
Task automatabilityclaude-haiku-4-5-202510012/5While calendar tools can be integrated with AI scheduling systems, this task requires real-time coordination with human subjects, study protocols with complex constraints, and handling of no-shows, rescheduling, and protocol deviations. Current AI struggles with the dynamic, exception-heavy nature of clinical trial scheduling and cannot achieve 50% time savings at equal quality end-to-end.
Task automatabilityclaude-sonnet-53/5Scheduling logic can be automated with calendaring/AI agents integrated to protocol visit windows, but complexity of protocol constraints, subject availability, and clinical resource coordination still requires human oversight for edge cases.'
Adoption barriersclaude-haiku-4-5-202510013/5Clinical research operates under IRB and regulatory oversight; scheduling must align with approved protocols and audit trails. While no licensing requirement mandates a human scheduler, liability and regulatory documentation requirements create moderate friction against full automation.
Adoption barriersclaude-sonnet-52/5No licensing requirement to schedule appointments, but protocol compliance, subject safety, and site policies create moderate procedural friction favoring human coordination.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI scheduling tools reduce some clerical work but still require human coordinators to manage protocol exceptions, subject communication, and compliance verification. The integrated cost of AI infrastructure, oversight, and remaining human effort likely exceeds the wage of a coordinator performing the task.
Cost vs. human wageclaude-sonnet-53/5Scheduling software/AI reduces coordinator time but still requires licensing, integration with EHR/CTMS, and human verification, making cost savings moderate rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed products exist for appointment scheduling (calendar integrations, basic reminder systems), but clinical research scheduling demands protocol compliance verification, subject eligibility checks, and regulatory documentation that require human oversight. No production system reliably handles the full scope without substantial manual intervention.
Technical feasibility todayclaude-sonnet-52/5Some clinical trial management systems offer automated scheduling modules, but few sites deploy fully autonomous AI agents for subject scheduling in production; most rely on coordinators using semi-automated tools.

Collaborate with investigators to prepare presentations or reports of clinical study procedures, results, and conclusions.

39

CI 30–48 · exposure 42 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research organizations adopt AI conservatively due to regulatory scrutiny and data sensitivity; while some larger CROs pilot LLM-assisted drafting, production deployment remains limited and heavily supervised. Adoption lags far behind tech and finance sectors.
Sector adoption velocityclaude-sonnet-52/5Clinical research and pharma/biotech sectors are cautious adopters of generative AI for regulated documentation, with pilots more common than widespread production use due to compliance concerns.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can substantially improve coordinator productivity by auto-generating first drafts, summarizing data, creating charts, and flagging inconsistencies—allowing coordinators to focus on investigator collaboration, interpretation, and compliance review rather than manual compilation work.
Augmentation potentialclaude-sonnet-54/5AI tools meaningfully speed up drafting, formatting, and summarizing complex study data into presentations or reports, letting coordinators and investigators focus on validation and interpretation.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with data compilation, figure generation, and draft report sections, but cannot independently synthesize study conclusions or ensure investigator alignment on clinical interpretation without significant human oversight. The collaborative, judgment-heavy nature of preparing presentations that communicate complex clinical findings limits automation to partial workflow support.
Task automatabilityclaude-sonnet-53/5AI can draft report sections, summarize study results, and generate slide outlines from structured data, but synthesizing accurate clinical conclusions and ensuring scientific validity still requires substantial investigator oversight and correction.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical study reporting is heavily regulated (FDA, ICH, protocol-specific requirements) and typically requires investigator signature and accountability; no automation removes the need for trained human review and sign-off. Liability asymmetry is high—errors in clinical reporting can trigger regulatory action, creating strong organizational and legal friction against full delegation to AI.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement mandates a human write these reports, but regulatory expectations (GCP, IRB, sponsor sign-off) and liability for inaccurate clinical claims create meaningful oversight friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted drafting and visualization reduce some coordinator labor, but integration overhead, review cycles with investigators, and the need for human fact-checking and regulatory sign-off mean the all-in cost remains close to or higher than the wage savings from partial automation.
Cost vs. human wageclaude-sonnet-53/5AI drafting can cut time on report/presentation preparation substantially, but the need for expert oversight, data verification, and compliance review keeps blended costs only moderately below fully human-authored reports.
Technical feasibility todayclaude-haiku-4-5-202510013/5LLMs and data visualization tools can generate report drafts and graphics from structured data, but no deployed system reliably handles the full pipeline of clinical study reporting—particularly the validation of statistical claims and regulatory compliance checks required in this domain. Products exist for components (text generation, figure automation) but not for end-to-end clinical reporting.
Technical feasibility todayclaude-sonnet-53/5LLM-based writing assistants and specialized clinical documentation tools are used in practice to draft reports and summaries, but accuracy issues with data interpretation and citations mean human review is still essential in production settings.

Review proposed study protocols to evaluate factors such as sample collection processes, data management plans, or potential subject risks.

38

CI 25–51 · exposure 45 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research and pharma remain relatively conservative and heavily regulated; AI adoption is mostly in back-office data tasks, not in core protocol evaluation. Pilot programs exist but production deployment remains limited.
Sector adoption velocityclaude-sonnet-52/5Clinical research and healthcare-adjacent administrative work adopts AI cautiously due to regulatory and safety concerns, with pilots more common than production deployment for this specific review task.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can rapidly surface risk flags, organize protocol components, and cross-check against templates and regulations, substantially accelerating a coordinator's review workflow. The human coordinator retains final judgment on novel risks and trade-offs, making this a high-value assistive use case.
Augmentation potentialclaude-sonnet-54/5AI tools can effectively assist coordinators by extracting key protocol elements, flagging missing data management details, or highlighting risk-relevant language for human review, meaningfully speeding up the process.
Task automatabilityclaude-haiku-4-5-202510014/5AI systems can extract, systematize, and cross-reference protocol elements (sample collection, data management, risk factors) against established standards and guidelines, automating a substantial portion of the evaluation checklist. However, complex risk assessment and novel study designs may require human judgment, preventing full end-to-end automation and limiting time savings to roughly 60–70%.
Task automatabilityclaude-sonnet-52/5AI can summarize and flag issues in protocols but the substantive risk evaluation and judgment calls about subject safety and regulatory compliance require human expertise, so full end-to-end automation with equal quality is not yet feasible.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory frameworks (FDA, IRB processes) and standard industry practices expect qualified human review of protocols; liability and error asymmetry are high if automation misses safety issues. While not an absolute legal bar, organizational and compliance friction is substantial.
Adoption barriersclaude-sonnet-54/5IRB/ethics review, regulatory compliance (FDA, GCP), and institutional policies typically require qualified human personnel to sign off on protocol risk assessments, creating strong liability and authorization barriers.
Cost vs. human wageclaude-haiku-4-5-202510013/5AI inference and integration costs are declining, but oversight by a qualified coordinator still dominates the all-in cost. The net cost is roughly comparable to partial human review, not a clear order-of-magnitude advantage.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply do first-pass summarization, but given the need for expert human verification and liability concerns, all-in cost including oversight is not dramatically lower than a trained coordinator's time.
Technical feasibility todayclaude-haiku-4-5-202510013/5AI products exist (contract review, compliance-checking tools) that can identify protocol deviations and flag standard risks, but reliability gaps remain in novel scenarios and subtle safety trade-offs. Deployed systems require meaningful human oversight and are not yet standard in clinical research environments.
Technical feasibility todayclaude-sonnet-52/5Some LLM-based tools assist with document review and highlighting inconsistencies, but no deployed product reliably performs full protocol risk/feasibility review in production clinical research settings today.

Maintain required records of study activity including case report forms, drug dispensation records, or regulatory forms.

36

CI 34–39 · exposure 50 · augmentation 75 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research is highly regulated and risk-averse; adoption of AI for autonomous record management is in early pilot stages. Most sites rely on traditional form-management systems with limited AI augmentation, and regulatory conservatism slows deployment of fully automated approaches.
Sector adoption velocityclaude-sonnet-52/5Clinical research organizations are historically slow adopters of new digital record systems due to regulatory validation requirements, though EDC adoption itself is mature; true AI-driven automation of compliance records is still in early pilot stages.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly assists coordinators by auto-populating forms from EHR data, flagging missing fields, performing compliance checks, and reducing manual transcription errors. These augmentations improve coordinator productivity and data quality substantially while the human retains verification and sign-off responsibility.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully assist coordinators by flagging discrepancies, auto-populating fields from source documents, and generating draft regulatory forms, substantially speeding up recordkeeping while humans retain final verification responsibility.
Task automatabilityclaude-haiku-4-5-202510013/5AI can automate significant portions of record-keeping—data entry, form population from structured sources, and compliance checking—but clinical research records often contain unstructured narrative, require domain judgment on what constitutes regulatory completeness, and demand human verification before submission. Roughly half the burden (routine form completion, data transcription) is automatable; the oversight and exception-handling require human involvement.
Task automatabilityclaude-sonnet-53/5Much of the record-keeping (data entry into case report forms, drug accountability logs) can be partially automated with EDC systems and AI-assisted data extraction, but accuracy verification, source document reconciliation, and regulatory sign-off still require significant human oversight.'
Adoption barriersclaude-haiku-4-5-202510015/5FDA regulations (21 CFR Part 11) and ICH-GCP guidelines require human accountability and signed attestation of record integrity. A licensed clinical research coordinator or investigator must legally verify and sign off on case report forms and regulatory submissions; AI cannot substitute for this human-authorized checkpoint.
Adoption barriersclaude-sonnet-54/5FDA/ICH-GCP regulations mandate accurate, auditable recordkeeping with accountability tied to qualified study personnel, and errors carry significant regulatory and patient-safety liability, creating strong compliance-driven barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Clinical research record-keeping involves compliance risk, audits, and regulatory consequences of error. AI tools reduce clerical overhead but require oversight infrastructure, validation workflows, and human sign-off, keeping all-in costs comparable to or exceeding a coordinator's time on routine data entry alone.
Cost vs. human wageclaude-sonnet-52/5EDC/eRegulatory software reduces some costs but still requires licensed coordinators for compliance oversight, monitoring, and data integrity, so all-in cost savings versus a human coordinator are moderate rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510013/5Deployed products exist for clinical trial data management and EHR-integrated forms (e.g., Medidata Rave, Oracle Clinical), but they typically require human operators to review, correct, and sign off on entries. True end-to-end automation without human gate-keeping is rare; most systems are semi-automated assistants rather than fully independent performers.
Technical feasibility todayclaude-sonnet-53/5Electronic data capture (EDC) systems and eSource tools are widely deployed in clinical trials, and AI-assisted data validation/query generation exists, but full automation of regulatory-grade recordkeeping with audit trail integrity remains narrow in scope and requires human verification.

Track enrollment status of subjects and document dropout information such as dropout causes and subject contact efforts.

36

CI 25–48 · exposure 38 · augmentation 63 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research operates in slow-moving, regulated sectors with conservative technology adoption. Most sites still rely on manual or semi-automated systems; pilot AI projects exist but production deployment remains rare due to validation and compliance requirements.
Sector adoption velocityclaude-sonnet-52/5Clinical research operations are digitizing steadily but slowly compared to tech/finance; CTMS adoption is common but AI-driven automation of dropout documentation is still nascent in practice.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by flagging potential dropouts from clinical data feeds, suggesting dropout categories, and auto-populating routine contact information, reducing clerical burden. However, the coordinator must verify classifications and make final compliance decisions, creating a moderate augmentation scenario.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully assist coordinators by auto-populating enrollment dashboards, flagging at-risk subjects, and drafting dropout summaries from notes, significantly speeding up documentation while a human verifies accuracy.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can extract and organize dropout data from structured forms and emails, the task requires contextual judgment about cause categorization, follow-up decisions, and protocol compliance that currently demands human oversight. Automation would require significant manual intervention to validate classifications and handle edge cases.
Task automatabilityclaude-sonnet-53/5Tracking enrollment counts and logging structured dropout data can largely be automated via CTMS/EDC systems and AI-assisted data entry, but capturing nuanced dropout causes and documenting contact efforts often requires human judgment and unstructured note review.'
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research is heavily regulated (FDA, ICH-GCP); dropout documentation is a critical regulatory record. Human coordinators must verify subject status, maintain compliance with study protocols, and ensure audit trails, creating a hard barrier to full automation without licensed oversight.
Adoption barriersclaude-sonnet-53/5No licensure is strictly required for this data tracking task, but GCP/regulatory documentation standards and audit trails create moderate compliance-driven barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI solutions for clinical data management are expensive to implement, integrate, and maintain, with substantial overhead for data privacy and regulatory compliance. The loaded cost of deploying such systems approaches or exceeds the wage of a coordinator performing manual tracking.
Cost vs. human wageclaude-sonnet-53/5Software-based tracking is cheap, but the portions requiring human outreach, judgment calls on dropout causation, and compliance documentation still need coordinator time, keeping overall cost roughly comparable to current staffing.
Technical feasibility todayclaude-haiku-4-5-202510012/5No production systems reliably automate this task end-to-end. Clinical trial data systems exist but require coordinators to manually enter and classify dropout information; AI assist tools are emerging but not yet deployed at scale for autonomous dropout tracking and documentation in regulated clinical settings.
Technical feasibility todayclaude-sonnet-53/5Clinical trial management systems already automate enrollment tracking and some dropout logging, but reliable extraction/summarization of dropout causes from varied sources (calls, emails, notes) is not yet a mature, widely deployed AI capability.

Arrange for research study sites and determine staff or equipment availability.

28

CI 25–30 · exposure 25 · augmentation 63 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research settings are conservative adopters of automation, with strong reliance on human oversight and established relationships with trial sites. While some pilot scheduling tools exist, production-level autonomous site arrangement and staff/equipment determination are rarely deployed in this highly regulated sector.
Sector adoption velocityclaude-sonnet-52/5Clinical research operations are moderately digitized but adoption of AI for site selection/logistics is still nascent, mostly pilot-stage CTMS integrations.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist by providing site databases, availability aggregation, equipment matching suggestions, and scheduling optimization. However, the human coordinator must still evaluate fit, negotiate contracts, and verify regulatory compliance, making augmentation supportive rather than transformative.
Augmentation potentialclaude-sonnet-54/5AI tools can significantly assist by analyzing site databases, predicting enrollment feasibility, and flagging equipment/staffing gaps, improving coordinator efficiency substantially.
Task automatabilityclaude-haiku-4-5-202510012/5This task requires negotiating with external sites, understanding complex scheduling constraints, and making judgment calls about resource fit. While AI could help gather site information and availability data, the actual arrangement and determination require human relationship-building and nuanced decision-making that current systems cannot reliably perform end-to-end.
Task automatabilityclaude-sonnet-52/5This involves logistics coordination, negotiation with sites, and judgment about suitability that requires human relationship management and decision-making not easily fully automated end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research is heavily regulated (FDA, IRB, protocol requirements), and site arrangements often involve legal agreements, licensing verification, and regulatory compliance checks. Human coordinators are typically required to sign off on site suitability and resource allocation, creating substantial institutional and liability barriers to full automation.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement for this specific task, but it requires institutional trust, contractual negotiation, and coordination with human parties that resist full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI tooling for scheduling and site-matching would require significant setup, oversight, and validation against regulatory requirements. The cost of deploying and maintaining such systems, plus human oversight, is likely comparable to or exceeds the cost of a coordinator's time on this task.
Cost vs. human wageclaude-sonnet-52/5Human coordinators are still needed for negotiation and relationship management; AI can assist with scheduling but doesn't replace the full cost of the task, so savings are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs the full task of arranging research study sites and assessing staff/equipment availability. AI can assist with data gathering and scheduling tools exist, but the coordination, negotiation, and contextual assessment required remains human-driven in production clinical research operations.
Technical feasibility todayclaude-sonnet-52/5Scheduling and coordination tools exist but no deployed product autonomously arranges research sites and negotiates staff/equipment availability at scale reliably.

Assess eligibility of potential subjects through methods such as screening interviews, reviews of medical records, or discussions with physicians and nurses.

25

CI 25–25 · exposure 25 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research coordinators remain concentrated in regulated, traditionally conservative healthcare and pharma sectors where adoption of AI automation for patient-facing eligibility decisions is slow; pilots exist but production deployment is limited.
Sector adoption velocityclaude-sonnet-52/5Clinical research and healthcare settings are historically slow adopters of AI-driven decision tools due to regulatory and safety concerns, with pilots more common than full production use.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist coordinators by automating medical record review, pre-screening questionnaires, and flagging relevant inclusion/exclusion criteria, reducing time spent on data gathering while the coordinator retains judgment on complex cases and clinical integration.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by pre-screening medical records, flagging potential matches, and summarizing patient history, significantly speeding up the coordinator's workflow while they retain final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with initial screening by extracting data from medical records and flagging potential eligibility criteria, but cannot reliably conduct screening interviews, synthesize physician/nurse discussions, or make final eligibility judgments that require clinical judgment and contextual understanding of complex medical histories.
Task automatabilityclaude-sonnet-52/5Eligibility assessment requires synthesizing unstructured clinical judgment, interviewing subjects, and coordinating with physicians—AI can assist with data extraction from records but cannot reliably perform the full end-to-end judgment call today.'
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory requirements (FDA, IRB oversight), patient consent procedures, physician sign-off on eligibility decisions, and the legal liability of incorrect screening create substantial barriers to full automation; human involvement is often mandated.
Adoption barriersclaude-sonnet-54/5Clinical trial protocols, IRB requirements, and regulatory frameworks (GCP, FDA) generally require qualified staff to confirm eligibility and document decisions, creating strong compliance and liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-based screening tools require significant integration, validation, and human oversight costs that approach or exceed the cost of a coordinator performing the initial assessment, especially when accounting for liability and regulatory compliance.
Cost vs. human wageclaude-sonnet-52/5AI tools can cheaply flag candidate records, but the full task still requires a human coordinator's time for interviews and physician discussions, so overall cost savings are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI tools exist for medical record extraction and basic eligibility flagging, no deployed system reliably performs end-to-end eligibility assessment across diverse protocols and clinical contexts; current products have significant error rates and require substantial human review.
Technical feasibility todayclaude-sonnet-52/5Some clinical trial matching tools exist (e.g., NLP-based cohort identification) but they are used as decision support, not as autonomous eligibility determination in production without human review.

Monitor study activities to ensure compliance with protocols and with all relevant local, federal, and state regulatory and institutional polices.

25

CI 25–25 · exposure 25 · augmentation 63 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research organizations adopt compliance software slowly due to regulatory conservatism, institutional inertia, and the high cost of errors. Pilot use of AI-assisted monitoring exists in some large research centers, but production displacement remains limited; most sites still rely on manual coordinator oversight.
Sector adoption velocityclaude-sonnet-52/5Clinical research and healthcare sectors are historically slow adopters of AI for compliance-critical functions due to regulatory caution, though eClinical systems are gradually incorporating automated alerts.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by automating routine data checks, flagging missing documentation, and summarizing protocol requirements, helping coordinators spend less time on manual review and more on interpretation. However, the assistive effect is moderate because the core task—ensuring regulatory and protocol compliance—requires sustained human judgment and signature authority.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by flagging protocol deviations, cross-referencing regulatory requirements, and automating documentation checks, significantly easing the coordinator's monitoring workload while they retain final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5Clinical research compliance monitoring involves interpreting complex, context-dependent regulations and protocols, then applying them to specific study scenarios. While AI could assist with flagging deviations in structured data or scheduling checks, the task requires human judgment about regulatory intent, institutional policy nuance, and case-by-case exceptions that current systems cannot reliably perform end-to-end at 50% time savings.
Task automatabilityclaude-sonnet-52/5Compliance monitoring requires judgment across evolving protocols, IRB requirements, and site-specific nuances that current AI cannot fully replicate end-to-end, though it can flag anomalies and check documentation completeness.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research is heavily regulated by FDA, IRBs, and institutional policy; coordinators often must document and sign off on compliance determinations. Liability and legal requirements create substantial friction: an AI system alone cannot satisfy regulatory requirements for human accountability and expert judgment in human-subjects research.
Adoption barriersclaude-sonnet-54/5Regulatory frameworks (FDA, IRB, GCP) generally require qualified human oversight and accountability for compliance monitoring, creating strong institutional and legal barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI tools (compliance checkers, data monitors) require significant integration, validation, and oversight by trained coordinators. The cost of building, maintaining, and supervising such systems is comparable to or exceeds the loaded wage of a junior compliance monitor, especially given liability and error costs in regulated research.
Cost vs. human wageclaude-sonnet-52/5AI tools can reduce time spent on document review and flagging, but the need for human verification, liability, and institutional oversight keeps blended costs closer to human-comparable rather than order-of-magnitude cheaper.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably monitors clinical research compliance autonomously. Audit and compliance software exists for data tracking and alerts, but evaluating protocol adherence across local, federal, and state rules requires expert interpretation that exceeds current production AI capabilities. Systems remain research-stage or require substantial human oversight.
Technical feasibility todayclaude-sonnet-52/5Some clinical trial management software includes automated protocol deviation alerts and compliance checklists, but comprehensive regulatory monitoring still relies heavily on trained human coordinators and auditors in production settings.

Direct the requisition, collection, labeling, storage, or shipment of specimens.

25

CI 25–25 · exposure 25 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While clinical research institutions adopt LIMS and inventory automation, these augment rather than replace coordinators. Adoption remains slow because human oversight is legally and practically non-negotiable in regulated research environments.
Sector adoption velocityclaude-sonnet-52/5Clinical research and healthcare settings adopt digital tracking tools but remain slow to automate physical lab logistics due to regulatory and safety constraints.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-driven LIMS, barcode tracking, automated alerts, and shipment logistics significantly boost coordinator productivity in managing specimen workflows, inventory, and compliance documentation while the coordinator retains final authority and physical oversight.
Augmentation potentialclaude-sonnet-53/5AI-enabled LIMS, barcode/label generation, and inventory tracking software meaningfully assist coordinators in organizing and monitoring specimen workflows, improving efficiency without replacing the physical/human oversight role.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist with labeling, storage tracking, and shipment documentation via computer systems, the core task involves physical handling and chain-of-custody oversight that requires human presence and responsibility. Current systems cannot reliably perform specimen requisition, collection, and physical handling end-to-end.
Task automatabilityclaude-sonnet-52/5This task involves substantial physical handling of biological specimens, coordination with lab personnel, and chain-of-custody documentation that requires hands-on human presence; AI can assist with tracking and paperwork but cannot perform the physical requisition/collection/shipment itself.'
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research specimens are heavily regulated (FDA, CLIA, institutional protocols); chain-of-custody and specimen integrity require documented human accountability. Regulatory frameworks and error-cost asymmetry (specimen loss or contamination) create strong legal barriers to full automation.
Adoption barriersclaude-sonnet-54/5Specimen handling is governed by strict regulatory, GCP, and chain-of-custody requirements often requiring certified/trained personnel and documented accountability, creating strong barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted tracking and inventory systems reduce administrative overhead, but the human coordinator remains essential for physical oversight, decision-making, and liability. Total cost savings are modest given the irreducible human component.
Cost vs. human wageclaude-sonnet-52/5Physical logistics, labeling accuracy, and regulatory chain-of-custody still require paid staff time and oversight; software assistance reduces some administrative cost but doesn't eliminate the human labor loop.
Technical feasibility todayclaude-haiku-4-5-202510012/5Laboratory information systems (LIMS) can automate some specimen tracking and documentation, but no deployed AI product reliably directs the full workflow of collection, labeling, storage, and shipment with the safety and compliance requirements clinical research demands.
Technical feasibility todayclaude-sonnet-52/5LIMS and lab tracking software exist for logging and labeling metadata, but no deployed AI product independently directs or performs the physical specimen workflow end-to-end.

Order drugs or devices necessary for study completion.

25

CI 25–25 · exposure 25 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research remains a heavily regulated, conservative sector with slow digitization of procurement workflows. Although some research organizations use electronic data capture systems, adoption of AI-driven ordering is minimal and largely limited to pilot projects rather than production deployment.
Sector adoption velocityclaude-sonnet-52/5Clinical research and pharma operations are moderately digitized but adoption of AI agents for regulated procurement tasks remains slow due to compliance requirements and cautious enterprise adoption in life sciences.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist coordinators by automating vendor lookups, flagging expired inventory, and drafting routine order requests, which would speed up the workflow. However, the regulatory review and final authorization remain coordinator-driven, limiting the transformative potential.
Augmentation potentialclaude-sonnet-53/5AI can assist by tracking inventory levels, generating reorder alerts, drafting purchase orders, and flagging compliance requirements, improving efficiency while the coordinator retains final responsibility.
Task automatabilityclaude-haiku-4-5-202510012/5Ordering drugs or devices requires navigating regulatory constraints (FDA approvals, GCP compliance), verifying inventory, and coordinating with multiple vendors—tasks with high domain specificity and liability. Current AI can assist with data entry or vendor lookup but cannot independently handle the regulatory verification and decision-making that 50% time savings with equal quality would demand.
Task automatabilityclaude-sonnet-52/5Ordering study drugs/devices involves procurement workflows, vendor coordination, and regulatory documentation that require judgment and accountability beyond simple transactional automation, though parts (form-filling, tracking) could be assisted.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research is heavily regulated; drug and device procurement is subject to FDA oversight, GCP guidelines, and institutional review board (IRB) requirements. Human coordinators are often required to sign off on orders and maintain documentation chains, creating legal and compliance barriers to full automation.
Adoption barriersclaude-sonnet-54/5Ordering investigational drugs/devices is tightly regulated (GCP, FDA, IRB oversight) and typically requires designated, trained personnel with accountability for chain-of-custody and regulatory compliance, creating strong barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI solutions for this task require significant setup, human oversight of regulatory compliance, and integration with specialized clinical trial management systems. The total cost of AI infrastructure and oversight likely approaches or exceeds the cost of a coordinator performing the task directly.
Cost vs. human wageclaude-sonnet-52/5Human coordinators still must verify protocol compliance, vendor contracts, and regulatory paperwork, so AI would mainly reduce administrative time rather than fully replace the labor cost, keeping cost savings modest given required oversight.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI product reliably performs end-to-end procurement of clinical trial materials. While general procurement software exists, the regulatory and compliance requirements specific to clinical research (chain of custody, expiration tracking, protocol alignment) are not handled at scale by existing AI systems.
Technical feasibility todayclaude-sonnet-52/5No deployed AI product autonomously manages clinical trial drug/device ordering end-to-end today; existing clinical trial management systems are largely rule-based software, not AI-driven agents performing procurement.

Contact outside health care providers and communicate with subjects to obtain follow-up information.

25

CI 25–25 · exposure 25 · augmentation 50 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research remains heavily regulated and risk-averse; adoption of fully automated subject follow-up is minimal. Most organizations still rely on coordinators for these interactions, with only limited use of AI-assisted scheduling or reminder systems in production environments.
Sector adoption velocityclaude-sonnet-52/5Clinical research and healthcare administration are historically slower to adopt AI-driven communication tools due to regulatory, privacy, and liability concerns, with pilots more common than production deployment.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist coordinators by drafting response templates, flagging subjects needing follow-up, and organizing provider contact information, moderately improving workflow efficiency. However, the human coordinator remains essential for handling complex cases, sensitive topics, and regulatory compliance.
Augmentation potentialclaude-sonnet-53/5AI can help draft outreach communications, schedule follow-ups, and summarize incoming information, meaningfully aiding coordinators without replacing the interpersonal and judgment-based aspects of the task.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can draft communication templates and schedule follow-ups, the task requires genuine rapport-building, handling sensitive health information, and responding to unpredictable subject concerns. Current systems cannot reliably manage the interpersonal nuance and medical context needed to obtain accurate follow-up information at ≥50% time savings with equal quality.
Task automatabilityclaude-sonnet-52/5This task involves relational outreach, scheduling, and probing for clinical follow-up details that often require judgment and rapport with subjects and providers; AI can assist drafting messages but cannot reliably conduct the full interaction end-to-end today.'
Adoption barriersclaude-haiku-4-5-202510014/5HIPAA and FDA regulations governing clinical trial communications, combined with liability concerns around autonomous health data collection and subject consent requirements, create substantial legal and regulatory barriers. Healthcare providers and trial subjects often prefer human contact for medical queries, adding organizational friction.
Adoption barriersclaude-sonnet-54/5Clinical trial coordination involves regulatory compliance (GCP, HIPAA), subject safety, and provider communication that often legally or procedurally requires qualified human staff to gather and verify follow-up medical information.
Cost vs. human wageclaude-haiku-4-5-202510012/5Automated systems require significant setup, compliance infrastructure, and human oversight to verify collected data and handle exceptions. When integrated with necessary oversight and error correction, the cost approaches or exceeds that of a coordinator handling routine follow-ups directly.
Cost vs. human wageclaude-sonnet-52/5While automated messaging is cheap, the need for human oversight, error correction, and handling of sensitive clinical data keeps effective all-in cost closer to human labor than a full order-of-magnitude cheaper.
Technical feasibility todayclaude-haiku-4-5-202510012/5Chatbots and IVR systems exist for appointment reminders, but no deployed product reliably conducts genuine clinical follow-up conversations with subjects or healthcare providers at scale. Error rates remain high for complex medical queries, and legal/privacy concerns around autonomous health communication limit real-world deployment.
Technical feasibility todayclaude-sonnet-52/5Some AI-driven patient outreach tools (chatbots, automated reminder/follow-up systems) exist in healthcare, but they are narrow in scope and not widely deployed for nuanced clinical trial follow-up requiring clinical judgment and provider coordination.

Participate in the development of study protocols including guidelines for administration or data collection procedures.

25

CI 25–25 · exposure 25 · augmentation 63 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research remains heavily regulated with strong human oversight requirements. Adoption of AI for protocol development is limited to pilot projects and preliminary drafting assistance, not production-scale displacement in typical research settings.
Sector adoption velocityclaude-sonnet-52/5Clinical research organizations are cautious adopters of AI in regulated documentation processes, with pilots emerging but production-scale protocol co-authoring by AI still uncommon.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist coordinators by generating protocol templates, flagging procedural inconsistencies, and suggesting data-collection language, raising efficiency on drafting phases. However, the human coordinator must remain in the loop for regulatory decisions and clinical design choices.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by drafting boilerplate sections, suggesting standard data collection procedures, and summarizing similar past protocols, significantly speeding up coordinator workflows while humans retain final control.
Task automatabilityclaude-haiku-4-5-202510012/5Protocol development requires deep domain knowledge, regulatory understanding, and creative problem-solving that current AI struggles with reliably. While AI can draft sections or suggest procedural language, the task demands expert clinical judgment and accountability that cannot be fully automated to achieve 50% time savings at equal quality.
Task automatabilityclaude-sonnet-52/5Protocol development requires clinical judgment, regulatory knowledge, and stakeholder negotiation that AI cannot fully replicate, though AI can draft sections or suggest standard language.a substantial portion still requires human expertise and accountability.
Adoption barriersclaude-haiku-4-5-202510014/5Research protocols require Institutional Review Board (IRB) review and approval, and regulatory compliance with FDA/Good Clinical Practice standards. Human experts must validate and take responsibility for protocols, creating substantial legal and liability barriers to full automation.
Adoption barriersclaude-sonnet-54/5Protocols must comply with IRB/ethics board requirements, GCP guidelines, and regulatory standards, requiring named qualified personnel to develop and sign off on them, creating strong institutional and legal barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5Human clinical research coordinators with domain expertise command significant loaded wages. AI tools provide limited value that does not yet offset integration, oversight, and quality-assurance costs for this specialized, high-stakes task.
Cost vs. human wageclaude-sonnet-52/5Because human oversight, regulatory review, and iterative stakeholder input remain necessary, AI reduces some drafting time but does not yet substantially undercut the loaded cost of skilled research coordinators and PIs.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably perform end-to-end protocol development with adequate clinical and regulatory validity. AI tools can assist with document drafting and formatting, but production systems do not yet independently author compliant research protocols.
Technical feasibility todayclaude-sonnet-52/5Some AI writing tools and clinical trial design software assist in drafting protocol language, but no deployed product independently produces compliant, sponsor-approved study protocols reliably.

Communicate with laboratories or investigators regarding laboratory findings.

25

CI 25–25 · exposure 25 · augmentation 63 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research organizations are cautious adopters of AI in regulated workflows due to compliance, audit, and liability concerns. While some use AI for draft assistance, production adoption of autonomous communication with laboratories remains minimal in the sector.
Sector adoption velocityclaude-sonnet-52/5Clinical research operations, especially at site level, are slow to adopt AI-driven communication workflows due to regulatory caution and fragmented systems.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by drafting communication templates, highlighting key findings from data reports, and organizing information for the coordinator to review and personalize. This raises coordinator efficiency without removing human judgment and accountability.
Augmentation potentialclaude-sonnet-54/5AI can help draft summaries, flag anomalies, and prepare communications for review, meaningfully speeding up the coordinator's workflow while they remain responsible for accuracy and judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can draft routine communications summarizing lab findings, clinical research contexts require nuanced interpretation of results, discussion of methodological issues, and coordination of complex next steps that demand human judgment and domain expertise. Current systems lack the contextual understanding and accountability required for reliable end-to-end handling.
Task automatabilityclaude-sonnet-52/5This involves relaying, interpreting, and discussing lab findings with investigators, requiring judgment about clinical significance and interactive back-and-forth that current AI cannot fully replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research is regulated (FDA, IRB oversight); communications become part of regulatory records and audit trails. GCP and regulatory compliance require that study coordinators—often trained and certified personnel—document and take responsibility for communications regarding findings, creating legal and liability barriers to full automation.
Adoption barriersclaude-sonnet-54/5Clinical trial communications involving lab data are subject to GCP/regulatory requirements and often need qualified personnel to interpret and relay findings accurately, creating strong compliance-driven barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI communication tools are cheap to run, but the task requires significant human oversight and often correction due to clinical context sensitivity, meaning total cost per reliable output remains comparable to direct human communication without major savings.
Cost vs. human wageclaude-sonnet-52/5Human coordination with liability-sensitive stakeholders still requires trained staff; AI assistance reduces some drafting time but oversight and integration costs keep the ratio close to human cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI systems can assist with templated communication drafting and data summaries, but no production systems reliably handle the full scope of clinical research communication—interpreting results, flagging anomalies, coordinating follow-up actions, and managing investigator relationships—at the fidelity required in regulated research environments.
Technical feasibility todayclaude-sonnet-52/5AI tools can draft summaries or flag abnormal values, but no deployed product autonomously manages the full communication loop between labs and investigators in clinical trials today.

Contact industry representatives to ensure equipment and software specifications necessary for successful study completion.

25

CI 20–30 · exposure 20 · augmentation 50 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research remains a traditionally regulated, cautious sector with low AI adoption in core protocol-critical functions. While digitization is increasing, substitution of AI for industry coordination remains rare due to regulatory and liability concerns.
Sector adoption velocityclaude-sonnet-52/5Clinical research operations are moderately digitized but adoption of AI for vendor liaison and equipment specification tasks is nascent, with most AI use concentrated in data analysis or documentation rather than external coordination.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by drafting communication templates, tracking contact information, and scheduling follow-ups, moderately raising coordinator productivity. However, the human must validate all technical specifications and maintain relationship continuity, limiting transformative potential.
Augmentation potentialclaude-sonnet-53/5AI can help draft communications, track specification requirements, and organize vendor information, providing moderate productivity gains while the coordinator still manages relationships and final decisions.
Task automatabilityclaude-haiku-4-5-202510012/5AI can draft emails and schedule meetings, but the task requires understanding nuanced equipment specifications, technical requirements, and relationship-building with industry contacts. Current systems lack the domain expertise and judgment needed to independently ensure specifications meet study requirements without substantial human oversight.
Task automatabilityclaude-sonnet-52/5This requires relationship-based communication, negotiation, and judgment about study-specific technical needs that current AI cannot fully replace, though drafting and information-gathering could be assisted.mphibian ideas.but full task requires human coordination and follow-through with vendors.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research is heavily regulated (FDA, IRB oversight), and study protocol compliance requires human accountability. Industry contacts often require authorization from licensed research coordinators or PIs, and liability for equipment specification errors rests with the organization, not AI systems.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement for this specific task, but clinical research oversight, GCP compliance, and accountability for equipment/software validation create moderate organizational and regulatory friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools for email automation and scheduling are inexpensive, but the task's critical nature (ensuring study-critical equipment meets specifications) requires expensive human oversight and verification. The cost of AI-induced errors (wrong specifications, regulatory issues) far exceeds the savings from automation.
Cost vs. human wageclaude-sonnet-52/5AI could reduce time on drafting emails or summarizing specs, but the core relationship management and verification still requires human labor, limiting cost savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can send emails and retrieve contact information, no deployed product reliably handles the relationship negotiation, technical validation, and contractual coordination this task entails. Most implementations require significant human verification of equipment specifications and regulatory compliance.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously manages vendor contact and specification verification for clinical trial equipment; this remains a manual coordination task performed by humans.

Inform patients or caregivers about study aspects and outcomes to be expected.

24

CI 23–25 · exposure 25 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Clinical research operates in a heavily regulated, risk-averse environment with slow digitization and strict compliance requirements. Current adoption of autonomous AI for patient communication is minimal; sectors dominating adoption (finance, tech) are irrelevant to clinical trial management.
Sector adoption velocityclaude-sonnet-52/5Clinical research and healthcare settings are cautious adopters of patient-facing AI due to regulatory, ethical, and trust concerns, resulting in slow deployment for this specific task.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by drafting personalized information sheets, summarizing study protocols, or flagging key points for the coordinator to cover, meaningfully easing preparation. However, the human coordinator remains the necessary point of contact, and augmentation is limited to content generation and organization rather than transforming the core communication task.
Augmentation potentialclaude-sonnet-54/5AI can generate patient-friendly explanations, summarize study protocols, and support multilingual communication, meaningfully aiding coordinators even though a human must lead the actual interaction.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate informational content about study aspects, the task requires personalized communication adapted to individual patient understanding, concerns, and context—elements that demand human judgment and empathy. Current AI systems lack reliable capability to handle the full two-way dialogue and relationship-building intrinsic to informed consent in clinical settings.
Task automatabilityclaude-sonnet-52/5Explaining study details and expected outcomes requires tailored communication, empathy, and responsiveness to patient questions that current AI cannot fully replace end-to-end without human oversight, especially given informed consent obligations.rules.rules.rules
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory and legal barriers are substantial: informed consent and patient communication are central to IRB oversight and good clinical practice (GCP) standards; liability exposure is high if patients report being informed by an automated system alone. Human coordinators are quasi-mandated by the regulatory framework governing clinical trials.
Adoption barriersclaude-sonnet-54/5Informed consent processes are heavily regulated (IRB, GCP, FDA) and typically require documented human-led interaction, creating strong compliance and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI for this task requires substantial setup (compliance review, validation, oversight infrastructure), and the cost of errors—misinforming patients or failing to document consent adequately—is asymmetrically high. The cost to deploy safely likely exceeds the wage savings from partial automation.
Cost vs. human wageclaude-sonnet-52/5While AI-generated educational materials are cheap, the need for a qualified human to verify understanding and answer nuanced questions keeps overall cost comparable to human-led processes.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed system reliably handles this task end-to-end in production clinical environments. Chatbots exist for basic FAQ delivery, but regulatory and ethical requirements around informed consent, liability, and documentation demand human coordinators; AI is at most a draft-assist tool with significant oversight burden.
Technical feasibility todayclaude-sonnet-52/5Chatbots and AI assistants exist for patient education but are not reliably deployed to independently conduct informed consent conversations in clinical trials; human coordinators remain the standard interface.

Identify protocol problems, inform investigators of problems, or assist in problem resolution efforts, such as protocol revisions.

24

CI 23–25 · exposure 25 · augmentation 63 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Clinical research remains a conservative, heavily regulated sector with slow AI adoption. Most sites still rely on manual protocol review by qualified coordinators and investigators; AI-driven problem identification is rare in production despite pilot projects.
Sector adoption velocityclaude-sonnet-52/5Clinical research is a highly regulated, moderately digitized sector where AI adoption for operational/compliance tasks is still in pilot stages rather than widespread production use.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist coordinators by flagging potential inconsistencies, cross-referencing regulatory requirements, and drafting revision suggestions, moderately raising their review speed. However, the human coordinator must still make the final judgment on clinical and regulatory significance.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by flagging inconsistencies, summarizing protocol deviations, and drafting revision language, significantly speeding up the coordinator's review and communication work while they retain final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist in identifying certain protocol problems through document analysis and pattern recognition, but protocol oversight requires nuanced understanding of clinical context, regulatory nuance, and investigator intent that current systems handle inconsistently. End-to-end automation would require reliable judgment calls on protocol safety and compliance—areas where AI falls short of the ≥50% time-saving threshold.
Task automatabilityclaude-sonnet-52/5Identifying subtle protocol problems requires clinical judgment, regulatory knowledge, and contextual understanding of trial operations that AI can partially support but not fully replace; drafting revisions can be assisted but the core diagnostic/judgment work resists full automation today.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory and legal barriers are substantial: IRB and regulatory agencies typically require a qualified human (the coordinator or investigator) to formally sign off on protocol compliance and problem resolution. Liability for missed safety issues creates strong protection against full automation.
Adoption barriersclaude-sonnet-54/5Clinical trial oversight is heavily regulated (GCP, IRB requirements), and protocol problem resolution typically requires sign-off by qualified investigators or coordinators, creating strong compliance and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI solutions for clinical protocol review require significant domain customization, human oversight, and integration effort; the total cost per identified problem often exceeds the cost of a coordinator's focused review time, especially given liability concerns.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply scan documents for inconsistencies, but the need for expert human review, investigator communication, and regulatory-compliant resolution keeps overall costs comparable to or only modestly below human-only workflows.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI tools can flag document inconsistencies and suggest revisions, no deployed product reliably performs the full task of identifying clinically meaningful protocol problems and working with investigators on resolution in production settings. Existing tools are narrow (e.g., checklist-based), not holistic problem identification.
Technical feasibility todayclaude-sonnet-52/5Some AI tools flag data anomalies or protocol deviations, but no deployed product reliably identifies nuanced protocol design problems and drives resolution end-to-end in production clinical research settings.

Record adverse event and side effect data and confer with investigators regarding the reporting of events to oversight agencies.

23

CI 20–25 · exposure 25 · augmentation 63 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research remains a regulated, compliance-heavy sector with slow digital transformation. While electronic data capture systems are standard, full AI-driven automation of safety reporting is rare in production; most sites rely on coordinators with software assistance rather than autonomous systems.
Sector adoption velocityclaude-sonnet-52/5Clinical research and pharma are adopting AI for data management but remain cautious and slow in regulated safety reporting workflows due to compliance risk.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by flagging potential adverse events from clinical notes, suggesting categorization, and organizing data for review, improving coordinator efficiency in data collection and initial triage. However, the judgment-heavy final determination of reportability and regulatory communication still requires the human in the loop.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully assist by extracting, coding, and flagging adverse events from clinical notes, helping coordinators work faster while humans retain final judgment and reporting responsibility.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with data entry and initial categorization of adverse events from structured sources, the task requires judgment about causality, severity assessment, and regulatory interpretation that demands human expertise. End-to-end automation falls short of the 50% time-saving bar because regulatory reporting demands human accountability and investigator judgment.
Task automatabilityclaude-sonnet-52/5Recording structured adverse event data can be partially automated, but conferring with investigators requires clinical judgment, contextual interpretation, and regulatory decision-making that current AI cannot reliably perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory agencies (FDA, EMA) require human accountability and signed attestation for adverse event reporting; oversight bodies mandate investigator review and approval. These legal and compliance requirements create hard barriers to full automation without human sign-off, and liability for incorrect or delayed reports falls on responsible parties.
Adoption barriersclaude-sonnet-55/5Adverse event reporting to regulatory bodies (e.g., FDA, IRBs) is legally mandated and requires qualified personnel and investigator sign-off, making this a heavily regulated, liability-sensitive task.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI tools (NLP for event extraction, data management systems) reduce some clerical burden but do not eliminate the need for a human coordinator to review, interpret, and decide on reporting. The all-in cost of oversight and error correction prevents significant cost savings versus the loaded wage of a clinical research coordinator.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply assist with data entry and coding, but the human oversight, liability review, and investigator discussions remain costly and necessary, keeping overall cost savings modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI systems can extract and organize adverse event data from clinical notes and databases, but deployed products do not reliably perform the full task of determining reportability, assessing causality, or making regulatory compliance decisions. This remains largely manual with only partial automation support in clinical trial management software.
Technical feasibility todayclaude-sonnet-52/5Some clinical trial software includes AI-assisted adverse event coding and triage suggestions, but no deployed product autonomously handles the full workflow including investigator consultation and regulatory reporting decisions.

Instruct research staff in scientific and procedural aspects of studies including standards of care, informed consent procedures, or documentation procedures.

21

CI 16–25 · exposure 17 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research remains highly regulated with embedded human oversight; adoption of AI for instruction is limited to early-stage content generation and supplementary support, not production-scale replacement of coordinators' teaching responsibilities.
Sector adoption velocityclaude-sonnet-52/5Clinical research settings are cautious adopters of AI due to regulatory scrutiny and compliance risk, with pilots more common than full production deployment for training functions.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by generating instructional materials, creating quiz templates, or drafting protocol summaries that coordinators review and deliver, modestly raising efficiency without removing the coordinator from the instruction loop.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by drafting training materials, creating FAQs, summarizing protocol changes, and providing on-demand answers to staff questions, enhancing coordinator efficiency.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires interactive instruction, judgment about what specific staff need to learn, and real-time adaptation to questions and comprehension—capabilities that current AI systems cannot perform end-to-end. While AI can generate training materials, it cannot replace the hands-on, responsive instruction that ensures research staff understand critical compliance and ethical requirements.
Task automatabilityclaude-sonnet-52/5Training staff involves live interaction, judgment calls, and adaptive instruction that current AI cannot fully replicate end-to-end, though AI can support content creation and Q&A.
Adoption barriersclaude-haiku-4-5-202510014/5Research institutions typically have strict regulatory requirements (IRB, GCP standards) that mandate human oversight and sign-off on research staff training and competency verification. Documentation of who trained whom and on what topics is often a legal/compliance requirement, creating hard barriers to full AI automation.
Adoption barriersclaude-sonnet-54/5Regulatory and institutional requirements (IRB oversight, GCP training certification, documentation accountability) generally require qualified human trainers to ensure compliance and liability coverage.
Cost vs. human wageclaude-haiku-4-5-202510012/5The cost of AI-generated training materials and oversight still requires a human instructor to validate, refine, and deliver the instruction, so total cost remains comparable to or higher than direct human instruction, particularly when error costs (regulatory/compliance failures) are factored in.
Cost vs. human wageclaude-sonnet-52/5Human coordinators combine domain expertise, institutional knowledge, and interpersonal instruction that would require substantial human oversight even if AI assists, limiting cost savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably delivers comprehensive instruction on research standards, informed consent, and documentation procedures in production environments. AI can assist with generating training content or answering FAQs, but organizations still rely on human coordinators to verify understanding and ensure compliance with institutional protocols.
Technical feasibility todayclaude-sonnet-52/5AI tools exist for generating training materials or answering procedural questions, but no deployed product reliably conducts full staff instruction on study standards and compliance without human oversight.

Prepare for or participate in quality assurance audits conducted by study sponsors, federal agencies, or specially designated review groups.

21

CI 16–25 · exposure 17 · augmentation 63 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research is highly regulated with slow digitization of audit processes. Organizations rely on established audit procedures and human relationships with sponsors; AI adoption in this domain remains in pilot phases, not production deployment.
Sector adoption velocityclaude-sonnet-52/5Clinical research and pharma sectors adopt digital tools cautiously due to regulatory compliance concerns; AI use in audit-related contexts remains in pilot stages rather than deep production adoption.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist coordinators with audit-readiness tasks—document organization, compliance checklist automation, preliminary gap analysis—but the human's regulatory knowledge and judgment remain central to successful audit outcomes.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully assist by organizing trial master files, flagging missing documentation, and pre-populating audit-readiness checklists, improving coordinator efficiency significantly while humans remain accountable for compliance interactions.
Task automatabilityclaude-haiku-4-5-202510011/5Quality assurance audits require human judgment, interpretation of regulatory nuance, and accountability for compliance findings. Current AI cannot independently conduct audits or make regulatory determinations with the rigor and legal defensibility audits demand.
Task automatabilityclaude-sonnet-52/5Preparing documentation and organizing records for audits requires judgment about compliance context and can be partially assisted, but actual participation in audits (answering auditor questions, defending study conduct) requires human presence and accountability. Only a fraction of the task (document retrieval, checklist prep) is automatable today.
Adoption barriersclaude-haiku-4-5-202510014/5Audits are regulated activities where sponsors, federal agencies, and IRBs expect direct human accountability and communication. A human coordinator must represent the institution; liability and regulatory expectation strongly discourage full substitution.
Adoption barriersclaude-sonnet-54/5Regulatory frameworks (FDA, ICH-GCP) require qualified, credentialed staff to interact with auditors and attest to data integrity, creating strong barriers against full automation or substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI could reduce administrative burden (scheduling, document gathering, preliminary report drafting), but the coordinator's oversight, audit participation, and sponsor communication remain necessary, limiting cost displacement to perhaps 20–30% of the role.
Cost vs. human wageclaude-sonnet-52/5AI can cut some document-prep time cheaply, but human oversight, verification, and audit participation remain necessary, keeping overall cost savings modest relative to a coordinator's time.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can assist with document review and data extraction for audit preparation, no deployed product reliably performs the full audit coordination task—which requires navigating sponsor expectations, regulatory relationships, and corrective action decisions—end-to-end.
Technical feasibility todayclaude-sonnet-52/5Some products assist with clinical trial document management and compliance checklists, but no deployed AI system independently prepares for or represents the coordinator in a live sponsor/FDA audit.

Perform specific protocol procedures such as interviewing subjects, taking vital signs, and performing electrocardiograms.

16

CI 7–25 · exposure 13 · augmentation 50 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research is a regulated, human-centered domain with slow digitization of core task execution. While electronic data capture and AI-assisted analysis are spreading, the direct subject-facing procedural work remains largely manual across the sector.
Sector adoption velocityclaude-sonnet-52/5Clinical research is a heavily regulated, in-person-oriented field where adoption of AI for physical data collection tasks is slow, though administrative and documentation aspects see more uptake.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist with post-task documentation (auto-generating summaries from interview notes), vital sign flagging and interpretation, and ECG preliminary analysis, raising coordinator efficiency on data processing and reporting components while they retain responsibility for subject interaction and protocol compliance.
Augmentation potentialclaude-sonnet-53/5AI can assist with structuring interview questions, transcribing responses, or interpreting ECG waveforms as a decision-support tool, but the physical execution remains human-driven.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist with data recording and vital sign interpretation, the core task requires direct human contact (interviewing, physical examination, ECG placement), which cannot be automated end-to-end. Current systems cannot reliably perform 50% time-saving at equal quality on the interactive and hands-on components.
Task automatabilityclaude-sonnet-51/5This task requires physical presence to attach ECG leads, take vital signs, and conduct in-person interviews with study subjects, none of which current AI systems can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong regulatory and human-contact barriers protect this role: clinical research protocols require documented human interaction and informed consent, liability for incorrect vital signs or ECG placement falls on the coordinator, and regulatory bodies mandate direct human assessment of subject safety and protocol adherence.
Adoption barriersclaude-sonnet-54/5Clinical trial protocols typically require trained, often certified personnel to perform and document these procedures for regulatory compliance (GCP/FDA), and liability for incorrect vital signs or ECG readings is high.
Cost vs. human wageclaude-haiku-4-5-202510012/5The hands-on nature of this task (taking vitals, placing ECG leads, building rapport with subjects) makes automation economically impractical today. AI assistance for documentation might reduce overhead marginally, but the loaded human wage remains lower than the cost of reliable automation plus oversight.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the physical labor involved, so there is no meaningful cost comparison; a human coordinator remains necessary for the hands-on components.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs the full task; AI excels at documentation and interpretation of results post-hoc, but clinical research coordinators must conduct interviews and physical assessments themselves. Some narrow components (e.g., ECG interpretation) have validated tools, but the integrated protocol execution remains human-dependent.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product autonomously performs hands-on clinical measurements like ECGs or vital signs in research settings; this remains physically embodied clinical work.

Oversee subject enrollment to ensure that informed consent is properly obtained and documented.

16

CI 11–20 · exposure 17 · augmentation 63 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research operates in a highly regulated environment with slow AI adoption. While some sites use digital consent tools and documentation aids, the pace of automation remains cautious due to regulatory scrutiny and the critical gatekeeping role of consent in human subjects research.
Sector adoption velocityclaude-sonnet-52/5Clinical research is a heavily regulated, cautious sector where AI adoption for consent processes remains in pilot stages due to compliance and liability concerns.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by generating consent form drafts, flagging missing data fields, tracking enrollment status, and alerting coordinators to protocol deviations—useful productivity gains while the human coordinator remains responsible for the actual consent interaction and sign-off.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist with consent form generation, tracking documentation completeness, flagging missing signatures, and simplifying language for readability, improving coordinator efficiency.
Task automatabilityclaude-haiku-4-5-202510011/5Obtaining and documenting informed consent requires direct human-to-subject interaction, legal/ethical judgment, and verification of subject comprehension—tasks that demand human presence and accountability. AI cannot reliably conduct the consent process or assume responsibility for this legally mandated function.
Task automatabilityclaude-sonnet-52/5Consent conversations require human judgment, empathy, and legal accountability for capacity assessment and comprehension checks, though AI can draft consent forms and track documentation status.The core oversight function resists full automation.
Adoption barriersclaude-haiku-4-5-202510015/5Informed consent is a legal and regulatory requirement with explicit human accountability mandates under 21 CFR Part 50, ICH-GCP, and institutional policy. A qualified human must conduct and sign off on consent; liability and compliance barriers prevent full automation.
Adoption barriersclaude-sonnet-55/5Informed consent is a strictly regulated, legally mandated process (FDA, IRB, GCP) requiring qualified human oversight and signature; liability for improper consent is severe.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools for consent management and documentation support exist but require significant human oversight and integration costs. The loaded cost of a clinical research coordinator remains lower than implementing and maintaining compliant AI systems with necessary human review.
Cost vs. human wageclaude-sonnet-52/5Documentation tracking tools are cheap, but the judgment-heavy oversight of proper consent still requires a trained coordinator, keeping overall costs comparable to human labor.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can draft consent forms or flag missing documentation, no deployed product reliably performs end-to-end informed consent oversight with the required legal and ethical accountability. Pilot systems exist for form completion, but regulatory and liability requirements limit production deployment.
Technical feasibility todayclaude-sonnet-52/5Some clinical trial software offers e-consent workflows and document tracking, but no deployed AI product independently 'oversees' consent adequacy or comprehension verification at scale.

Solicit industry-sponsored trials through contacts and professional organizations.

15

CI 5–25 · exposure 8 · augmentation 38 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Clinical research is a regulated, human-relationship-dependent sector with slow digital transformation in business development. Trial sponsorship remains a high-touch sales function; there is no evidence of automation in this specific workflow.
Sector adoption velocityclaude-sonnet-52/5Clinical research administration is a moderately digitized but relationship-heavy sector where AI adoption for business development activities remains nascent.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist marginally by identifying potential trial sponsors via database searches or drafting outreach emails, but the core task—relationship building and persuasion—remains human-centric. Assistance is limited and peripheral.
Augmentation potentialclaude-sonnet-53/5AI tools can help draft outreach emails, identify potential sponsors, track industry contacts, and prepare proposals, meaningfully aiding the coordinator without replacing them.
Task automatabilityclaude-haiku-4-5-202510011/5Soliciting industry-sponsored trials requires relationship building, negotiation, and judgment about trial fit—work that is fundamentally social and requires sustained human credibility and trust. Current AI cannot autonomously establish new business relationships or close sponsorships.
Task automatabilityclaude-sonnet-52/5This is a relationship-driven business development task requiring trust-building, networking, and negotiation that AI cannot autonomously execute end-to-end today.','rating_note':
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: industry sponsorship decisions depend on human trust and relationships with research institutions, regulatory approval of trials involves licensed physicians and IRBs, and pharmaceutical companies typically require direct negotiation with credentialed staff. Substitution is not legally or practically feasible.
Adoption barriersclaude-sonnet-53/5No strict licensing barrier, but sponsor trust, institutional reputation, and personal relationships create significant organizational friction against substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI cannot meaningfully reduce cost here because the task is not automatable; any AI tooling would require expensive oversight and the core work—relationship development—must still be done by humans. The all-in cost of AI would exceed the human wage.
Cost vs. human wageclaude-sonnet-52/5AI can support lead identification or CRM tasks cheaply, but the actual soliciting and relationship management still requires paid human time, so overall cost savings are limited.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably performs business development or trial solicitation in clinical research contexts today. This requires human-to-human outreach, understanding of regulatory landscapes, and negotiation capabilities beyond current LLM or agent systems.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product independently solicits and closes industry-sponsored clinical trial contracts through professional networking; this remains a human relationship function.

Interpret protocols and advise treating physicians on appropriate dosage modifications or treatment calculations based on patient characteristics.

13

CI 6–20 · exposure 13 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Clinical research organizations move cautiously on automation due to regulatory requirements and safety concerns. Most adoption to date is limited to assistive tools (protocol search, data entry validation) rather than autonomous dosage recommendation, and production deployment remains sparse.
Sector adoption velocityclaude-sonnet-52/5Clinical research and healthcare are historically slow to adopt autonomous AI decision-making for patient safety-critical calculations, despite growing use of AI in trial administration generally.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can substantially assist coordinators by rapidly searching and summarizing relevant protocol sections, extracting patient parameters, and suggesting dosage calculations for physician review. This augmentation significantly accelerates the workflow while the physician retains final authority and responsibility.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by flagging protocol deviations, calculating dose ranges, and cross-referencing patient data, substantially speeding up the coordinator's and physician's review process.
Task automatabilityclaude-haiku-4-5-202510012/5Interpreting protocols and advising on dosage modifications requires deep contextual understanding of patient characteristics, medical history, and nuanced protocol language. While AI can retrieve and summarize protocol text, the judgment call on appropriate dosage modifications for specific patients involves clinical reasoning and liability that current systems cannot reliably perform end-to-end without substantial human oversight.
Task automatabilityclaude-sonnet-51/5This requires clinical judgment integrating patient-specific factors, protocol nuance, and safety accountability; current AI cannot reliably perform this end-to-end with equal quality and safety assurance.
Adoption barriersclaude-haiku-4-5-202510015/5Clinical trial protocols are heavily regulated under FDA, IRB, and GCP guidelines; treating physicians must sign off on dosage modifications, and any error has direct patient safety consequences. Liability, regulatory oversight, and the legal requirement for physician authorization create hard barriers to full automation.
Adoption barriersclaude-sonnet-55/5Dosage modification advice is a clinical/regulatory activity requiring qualified personnel (often under GCP and IRB-approved protocols) with physician sign-off, making this a hard-barrier task.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted systems reduce coordinator time on routine calculations and protocol lookups, but the infrastructure, integration, validation, and mandatory physician oversight add cost. The effective savings per task remain modest compared to the loaded wage of a research coordinator, especially when liability and QA costs are included.
Cost vs. human wageclaude-sonnet-52/5Even if AI assisted with dosage calculations, the required human verification, liability review, and clinical oversight keep costs comparable to or only modestly below human-only performance.
Technical feasibility todayclaude-haiku-4-5-202510012/5Although LLMs can draft dosage guidance from protocols and patient data, no deployed product reliably performs this task independently in clinical settings. Products exist for clinical decision support, but they operate as assistants under physician review rather than autonomous agents, and error rates on edge cases remain material.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously advises physicians on dosage modifications in production clinical trial settings; this remains research-stage decision support at best.

Confer with health care professionals to determine the best recruitment practices for studies.

11

CI 5–16 · exposure 8 · augmentation 38 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Clinical research remains a human-intensive, regulated field with low automation culture for professional judgment tasks. Organizations have adopted IT tools for scheduling and documentation, but not AI agents making or leading recruitment strategy decisions.
Sector adoption velocityclaude-sonnet-52/5Clinical research operations are moderately digitized but still rely heavily on personal coordination and site relationships; AI adoption for this specific interpersonal task is nascent.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with summarizing best-practice literature or organizing meeting agendas, but the core task—conferring with professionals to reach consensus—is interpersonal and requires the coordinator's voice and authority. Assistance is minimal and peripheral.
Augmentation potentialclaude-sonnet-53/5AI can help coordinators prepare talking points, analyze past recruitment data, or draft materials to support these conversations, meaningfully aiding preparation even though the interaction itself is human-led.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires real-time dialogue, domain judgment, and interpersonal negotiation with healthcare professionals to assess and align on recruitment strategies. AI cannot credibly substitute for the live conferencing and relationship-building essential to determining 'best practices' through professional consensus.
Task automatabilityclaude-sonnet-52/5This is a collaborative, judgment-based conversation requiring relationship-building and contextual clinical knowledge; AI cannot conduct this interpersonal consultation end-to-end today. Only preparatory research or summarization could be offloaded.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical research is heavily regulated, and recruitment strategy decisions typically require sign-off from research coordinators, PIs, and IRBs. The human coordinator's professional judgment and accountability are often embedded in protocol compliance and documentation requirements.
Adoption barriersclaude-sonnet-54/5Clinical trial recruitment involves regulatory compliance (IRB, GCP), professional judgment, and trust relationships with physicians, creating strong organizational and quasi-regulatory barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The task is fundamentally human-to-human professional interaction. AI assistance (e.g., scheduling, note-taking) is marginal; replacing the coordinator conducting these conferences would either fail or require human oversight that negates cost savings.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot perform the core interpersonal negotiation task, there is no viable AI-only cost basis to compare against the human wage.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system reliably conducts independent professional conferences with healthcare teams to establish recruitment protocols. This requires authentic two-way professional engagement, not document analysis or scripted responses.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously confers with healthcare professionals to negotiate recruitment strategy; this remains a human relationship-driven activity with no production system replacing it.

Organize space for study equipment and supplies.

7

CI 0–15 · exposure 0 · augmentation 25 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Clinical research settings show minimal adoption of physical automation for space organization; the sector remains conservative with strong emphasis on human oversight and institutional control over critical research infrastructure.
Sector adoption velocityclaude-sonnet-51/5Physical facility organization in clinical research settings shows negligible AI adoption; this is a low-digitization, hands-on activity.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide limited assistance through spatial planning visualization or inventory management tools, but the core physical organization task requires human presence and judgment in a controlled clinical environment, limiting augmentation value.
Augmentation potentialclaude-sonnet-52/5AI could help generate checklists, inventory trackers, or spatial layout suggestions, but offers limited direct assistance for the physical act of organizing space.
Task automatabilityclaude-haiku-4-5-202510011/5Organizing physical space requires sensorimotor capabilities, real-time spatial reasoning, and decision-making about placement and accessibility that current AI systems cannot perform autonomously. This task involves manipulation of objects in three-dimensional environments, which remains far beyond current capabilities.
Task automatabilityclaude-sonnet-51/5This is a physical, spatial organization task involving arranging equipment and supplies in a real-world clinical space, which current AI systems cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5This task requires direct physical presence in a regulated clinical research environment where equipment placement affects study protocol compliance, safety, and data integrity. Regulatory and liability concerns create strong barriers to autonomous automation.
Adoption barriersclaude-sonnet-52/5No licensing requirement specifically for organizing space, but the physical nature and need for on-site presence and safety compliance create practical barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems capable of physical organization would require expensive robotics infrastructure, specialized sensors, and custom integration. These costs would far exceed the labor cost of a clinical coordinator performing this task manually.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the physical labor and spatial judgment required, so there is no viable AI cost comparison—human labor is the only option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably organizes physical laboratory spaces end-to-end. While computer vision can identify objects, current systems cannot autonomously manipulate equipment, make context-aware placement decisions, or adapt to dynamic laboratory constraints.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product organizes physical lab or clinical study space; this remains a manual, hands-on logistics task.

Related occupations — Management

How to read this

A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.

What would change this score

New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.