Interviewers, Except Eligibility and Loan
43-4111.00Median wage $45,920/yr148,060 employed (US)Rank #75 of 923 scored · top 8% by substitution
Interview persons by telephone, mail, in person, or by other means for the purpose of completing forms, applications, or questionnaires. Ask specific questions, record answers, and assist persons with completing form. May sort, classify, and file forms.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution51
Exposure46
Augmentation64
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
16 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
31%
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%46
panel mean rating 2.8/5 → substitution pressure 46/100
Technical feasibility todayw 20%48
panel mean rating 2.9/5 → substitution pressure 48/100
Cost vs. human wagew 15%60
panel mean rating 3.4/5 → substitution pressure 60/100
Adoption barriersw 20%inverted — strong barriers lower the score62
panel mean rating 2.5/5 (barrier strength) → substitution pressure 62/100
Sector adoption velocityw 10%45
panel mean rating 2.8/5 → substitution pressure 45/100
Task breakdown (16 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.
Locate and list addresses and households.
80CI 76–84 · exposure 75 · augmentation 63 · importance 4.4/5 · click for rater detail
Locate and list addresses and households.
80| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Survey firms, market research companies, and interviewing services have already integrated address databases and mapping tools into production workflows; adoption is well underway in the information and professional services sectors where interviewers operate. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Survey and research organizations that employ interviewers have moderate digitization; sampling frame construction increasingly uses automated tools but adoption varies by organization size and field type. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI systems assist interviewers by pre-populating address lists and verifying household information, reducing time spent on research and data entry. The human interviewer still reviews results and may manually resolve ambiguities or missing data, making this a meaningful productivity boost without full automation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and mapping/database tools substantially speed up and improve accuracy of address/household listing while humans still verify and use the lists in fieldwork. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can reliably locate addresses and compile household lists using public databases, maps APIs, and data aggregation tools with minimal manual intervention, achieving significant time savings. The task is largely data lookup and organization, which AI handles efficiently, though some edge cases (incomplete records, address ambiguities) may require verification. |
| Task automatability | claude-sonnet-5 | 4/5 | Locating and listing addresses/households is largely a data-lookup and compilation task that can be done via mapping APIs, public records databases, and geocoding tools with significant time savings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Privacy and data protection regulations (GDPR, CCPA) and client confidentiality policies create some friction, but no hard legal barrier prevents automated address collection if properly authorized and compliant. Most organizations can adopt with routine data governance oversight. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There is no licensing, liability, or human-contact requirement for compiling address lists; it's a purely administrative/clerical function. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated address lookup and household listing via APIs and databases cost pennies per record, far below the loaded hourly wage of an interviewer conducting manual research and compilation. AI cost is at least an order of magnitude cheaper than human labor for this task. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated address lookup and household listing via APIs and databases costs a small fraction of a cent per record compared to a human manually compiling such lists. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Multiple deployed products (mapping services, data enrichment platforms, address verification APIs) reliably perform address location and household listing at scale in production systems. These tools are widely integrated into business workflows, though they occasionally require human review for complex or ambiguous cases. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Mature geocoding, GIS, and address-verification products (Google Maps API, USPS databases, data enrichment tools) are widely deployed in production for exactly this kind of list-building today. |
Compile, record, and code results or data from interview or survey, using computer or specified form.
78CI 72–84 · exposure 75 · augmentation 88 · importance 4.2/5 · click for rater detail
Compile, record, and code results or data from interview or survey, using computer or specified form.
78| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Survey and market-research firms, HR departments, and customer-insight teams are already adopting RPA and AI-driven data capture in production. This is standard practice in digitized organizations across information, professional services, and research sectors. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Survey research and market research firms have adopted digital data capture and NLP coding tools at moderate pace, with mixed penetration depending on organization size and legacy paper-based methods. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at assisting interviewers by auto-populating forms in real time, suggesting codes as data is entered, and flagging inconsistencies—significantly raising the speed and accuracy of data handling while the human reviews and validates. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools for transcription, auto-coding of open-ended responses, and data cleaning substantially speed up this task while humans still verify and finalize outputs. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Compiling, recording, and coding structured interview/survey data is highly amenable to automation. Current AI can extract information from interviews (via transcripts or notes), populate forms, and apply coding schemas with minimal human intervention, achieving >50% time savings on this largely mechanical data-entry and categorization task. |
| Task automatability | claude-sonnet-5 | 4/5 | Compiling, recording, and coding structured interview/survey data into a computer or form is a well-defined data entry and classification task that current AI (especially with OCR, transcription, and text-classification models) can largely automate, though edge cases and open-ended coding may need review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few hard regulatory or licensing barriers exist for automating this administrative data-processing task. Main friction is organizational inertia and quality-assurance requirements (human review), not legal mandate; no licensed professional sign-off is required. |
| Adoption barriers | claude-sonnet-5 | 1/5 | This is a back-office clerical task with no licensing, liability, or human-contact requirements blocking automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference and form-automation costs are orders of magnitude cheaper than the loaded hourly wage of an interviewer doing manual data entry and coding, especially at volume. The integration overhead is minimal for standard survey workflows. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated data capture and coding via software is dramatically cheaper per unit than manual transcription and coding by human interviewers or data clerks. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products reliably perform data extraction, form-filling, and basic coding/categorization at scale (e.g., RPA tools, LLM-based data extraction, survey-software integrations). Production systems exist, though accuracy depends on data quality and schema complexity; occasional oversight is needed but the capability is mature. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Survey platforms and CAQDAS/CATI tools already offer automated data capture, transcription, and coding features in production use, though some manual verification remains common for quality control. |
Ask questions in accordance with instructions to obtain various specified information, such as person's name, address, age, religious preference, or state of residency.
77CI 76–79 · exposure 75 · augmentation 63 · importance 4.6/5 · click for rater detail
Ask questions in accordance with instructions to obtain various specified information, such as person's name, address, age, religious preference, or state of residency.
77| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Information and service sectors have rapidly deployed conversational AI for intake, surveying, and data collection over the past 3–5 years. Public adoption data shows widespread IVR and chatbot use for routine interviews in healthcare, finance, government, and customer service. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Adoption varies by sector; many organizations still use human interviewers or paper/phone processes for structured intake, though digital self-service forms and chat-based intake are increasingly common. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants can draft interview guides, validate responses in real time, suggest follow-up questions, and automatically populate forms—significantly boosting human interviewer productivity while keeping humans in control of nuanced or sensitive interactions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can pre-fill, transcribe, or prompt interviewers with next questions, improving efficiency, but the core task is simple enough that augmentation value is moderate rather than transformative. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Modern AI systems can reliably ask structured questions and capture responses at high speed with minimal human intervention. The task is inherently formulaic—following a script to collect demographic and preference data—which AI can execute end-to-end with >50% time savings, though quality control and handling edge cases may still require human oversight. |
| Task automatability | claude-sonnet-5 | 4/5 | Structured question-asking to collect specified demographic/contact data is a highly scriptable, conversational task that chatbots and voice AI systems already handle at scale (e.g., intake forms, IVR systems, survey bots).rapid |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal barriers exist for automated information collection; no license or human sign-off is typically required. The main friction is organizational inertia, customer preference for human contact in some contexts, and potential data privacy/consent requirements—modest but not prohibitive barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | Minimal licensing requirements for this specific data-collection task, though some contexts (e.g., census, legal intake) may require verification or have privacy/regulatory handling requirements around sensitive data like religion. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference and integration cost is orders of magnitude cheaper than the loaded wage of an interviewer, especially at volume. A single deployment can handle thousands of structured interviews with minimal marginal cost per interaction. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated form/voicebot systems cost a small fraction of a human interviewer's wage per completed interview, especially at scale. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed conversational AI and survey automation tools routinely perform this task in production. Chatbots and IVR systems collect biographical and preference information reliably at scale in many organizations, though accuracy rates on nuanced responses or complex follow-ups remain slightly below perfect human performance. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed conversational AI and automated intake systems (surveys, IVR, chatbot forms) reliably collect this kind of structured personal information in production today across many industries. |
Review data obtained from interview for completeness and accuracy.
73CI 67–79 · exposure 70 · augmentation 75 · importance 4.3/5 · click for rater detail
Review data obtained from interview for completeness and accuracy.
73| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Information, financial services, and survey/polling organizations have widely adopted automated data validation in production pipelines. Customer-facing and compliance-heavy sectors show slower but still measurable adoption. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Survey and interview-based sectors (market research, healthcare intake, government surveys) are adopting automated validation tools steadily, but many still rely on manual review as a norm. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered validation tools strongly augment human reviewers by flagging anomalies, suggesting corrections, and auto-populating obvious errors, allowing interviewers to focus on complex or sensitive inconsistencies requiring judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can flag inconsistencies, missing data, and outliers instantly, substantially speeding up human reviewers who then confirm or correct flagged issues. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI can systematically check interview data for missing fields, format errors, and logical inconsistencies (e.g., contradictory responses) at scale with minimal human oversight, achieving >50% time savings. However, nuanced accuracy assessment and context-dependent validation may require human judgment on edge cases. |
| Task automatability | claude-sonnet-5 | 4/5 | Reviewing interview data for completeness and accuracy against structured fields or checklists is well within current NLP/data-validation capabilities, especially for structured or semi-structured survey data.dequate for most cases though edge cases need human judgment.), |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal regulatory barriers exist for automating data quality checks; however, organizational workflows often require human sign-off and some sectors (financial, legal) impose downstream liability pressures that slow adoption despite technical feasibility. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this specific review step, though organizational quality-control policies and liability for data errors create some friction before fully removing human review. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated data validation via APIs or integrated modules costs pennies per review versus $15–$30+ loaded labor cost per interview validation check, representing >10× cost advantage. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated validation rules and AI-based checks run at a fraction of the cost of manual review, especially at scale, though some human spot-checking remains needed. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed NLP and rule-based validation systems in production CRM and survey platforms routinely perform data completeness and consistency checks. Error rates on well-structured forms are low, though performance degrades on unstructured or ambiguous responses. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products exist (survey platforms, CRM validation tools, automated QA checks) that flag missing fields or inconsistencies, but nuanced accuracy checks requiring contextual understanding of open-ended responses still have material error rates. |
Perform office duties, such as telemarketing or customer service inquiries, maintaining staff records, billing patients, or receiving payments.
71CI 66–75 · exposure 67 · augmentation 63 · importance 4.3/5 · click for rater detail
Perform office duties, such as telemarketing or customer service inquiries, maintaining staff records, billing patients, or receiving payments.
71| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Healthcare, finance, and telecom sectors—major employers of interviewers—have rapidly deployed payment automation, billing systems, and chatbots; production AI-driven customer service and telemarketing are widespread, though voice interviewing remains more cautious. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Administrative/office support functions in healthcare, finance, and customer service sectors show fast, broad adoption of automation tools like chatbots and billing software. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI assists with routing inquiries, suggesting responses, and flagging payment/billing errors, meaningfully raising human productivity on customer service and administrative tasks, but does not transform the full scope of complex negotiation or intake interviewing. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly boost productivity for staff handling billing, records, and customer inquiries by automating repetitive steps while humans manage exceptions and complex cases. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Routine office tasks like billing, payment receipt, and staff record maintenance are largely automatable with existing systems (accounting software, CRM, payment processors), but telemarketing and handling varied customer service inquiries still require judgment and adaptation. Overall, roughly 50–60% of the bundled task can be automated at equal quality. |
| Task automatability | claude-sonnet-5 | 4/5 | Many of these sub-tasks (billing, telemarketing scripts, payment processing, basic record maintenance) are highly structured and already handled by chatbots, IVR systems, and billing software with substantial time savings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few hard regulatory barriers exist for most office tasks; telemarketing is subject to compliance (Do Not Call, consent recording) but automation does not face licensing blocks. Customer preference for human contact and data privacy oversight create modest friction, but nothing prevents substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | Minimal licensing requirements for these clerical tasks, though patient billing may involve some compliance (HIPAA) and customer preference for human contact in service calls creates mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated billing, payments, and record maintenance are orders of magnitude cheaper than human operators; telemarketing and service chatbots cost far less per interaction than salaried interviewers, though per-inquiry quality variation may require human oversight. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated billing, payment processing, and scripted telemarketing/customer service systems cost a fraction of a human wage per interaction once deployed at scale. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature products (billing software, payment systems, CRM platforms, basic chatbots) reliably perform most of these subtasks in production, though handling complex customer inquiries still involves material error rates and manual oversight in deployed systems. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed products (CRM/EHR billing modules, telemarketing bots, payment portals) reliably perform these functions in production across many organizations today, though some tasks like nuanced customer service still need human escalation. |
Assist individuals in filling out applications or questionnaires.
69CI 59–79 · exposure 62 · augmentation 75 · importance 4.1/5 · click for rater detail
Assist individuals in filling out applications or questionnaires.
69| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Insurance, government benefits, financial services, and healthcare are actively deploying AI-assisted intake and form systems. Pilot and production adoption is visible across these sectors, with increasing deployment in customer-facing and internal workflows. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Adoption is growing steadily in customer service and government digital services but still uneven, with many agencies and organizations retaining human interviewers for accessibility and trust reasons. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly enhances interviewer productivity by auto-suggesting responses, flagging missing fields, checking conditional logic, and offering real-time guidance on complex sections, allowing humans to focus on clarification and validation rather than data entry mechanics. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can pre-fill forms, flag missing information, and draft responses that a human interviewer then reviews with the applicant, meaningfully speeding the interaction. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate parts of this task by guiding users through form-filling, pre-populating fields from available data, and detecting missing information. However, complex cases requiring clarification, judgment about eligibility nuances, or handling ambiguous responses still need human oversight, limiting end-to-end automation to roughly 50% time savings. |
| Task automatability | claude-sonnet-5 | 4/5 | Conversational AI can guide individuals through form fields, ask clarifying questions, and populate applications with high time savings, especially for standardized questionnaires. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No legal requirement mandates a licensed human perform form assistance in most sectors. Barriers are mainly organizational (quality preference, liability concerns, customer expectation of human contact) rather than hard regulatory or licensing restrictions. |
| Adoption barriers | claude-sonnet-5 | 2/5 | Some contexts (legal aid, sensitive medical/benefits applications) prefer human assistance for trust and accessibility, but most application assistance carries no licensing requirement. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-powered form assistance has very low marginal cost per interaction (minimal compute and integration overhead) compared to paying an interviewer's loaded wage. Large-scale deployment significantly reduces cost-per-application-processed. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated conversational form-filling costs a fraction of a cent to a few cents per interaction versus a paid interviewer's hourly wage for the same volume. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Chatbots and form-filling assistants are deployed in production (e.g., eligibility pre-screeners, onboarding bots), but they often struggle with edge cases, conditional logic, and nuanced follow-up questions. Material error rates and scope limitations prevent this from reaching reliable, fully autonomous operation at scale. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Chatbots and voice assistants are deployed at scale for intake forms, customer applications, and surveys across healthcare, government, and business, though edge cases still require human intervention. |
Prepare reports to provide answers in response to specific problems.
44CI 30–59 · exposure 38 · augmentation 63 · importance 3.6/5 · click for rater detail
Prepare reports to provide answers in response to specific problems.
44| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Interview and consulting roles remain in lower-digitization, human-contact-intensive sectors. Adoption of AI for report generation is slow; most firms still rely on human analysts to synthesize interview data and craft problem-specific responses. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Interviewing and case-reporting roles span sectors with mixed digitization; some professional services adopt AI drafting tools quickly while others in social services or public sector lag. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by drafting report sections, suggesting relevant findings, organizing data, and flagging patterns—useful support that raises human analyst productivity. However, the assistant role is limited because the human must substantially revise and validate the output. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can strongly assist by drafting report structure, summarizing information, and suggesting answers, significantly speeding up the human's writing process while they verify final content. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can generate drafted reports from structured data, this task requires understanding specific problems, analyzing context, and providing customized answers—activities that typically need human judgment. Current AI struggles with the problem-scoping and contextual decision-making phases, limiting time savings below the 50% threshold. |
| Task automatability | claude-sonnet-5 | 3/5 | LLMs can draft structured reports answering defined questions from provided data, but require accurate underlying data collection and domain-specific judgment to ensure correctness, limiting full automation without human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Interview-based work often involves client trust, judgment calls, and accountability for recommendations. While not strictly licensed, organizations typically expect human responsibility for problem diagnosis and report accuracy, creating moderate friction to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No strict licensing requirement for report writing itself, though organizational sign-off and accuracy verification create moderate friction before reports are finalized. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration of AI report-generation still requires significant human oversight, fact-checking, and customization for each unique problem presented. The total cost (inference, integration, review) often approaches or exceeds the cost of a human analyst drafting the report directly. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Once data is available, generating a report via AI is far cheaper than a human spending time drafting and formatting, though data gathering and verification costs remain. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some report-generation and template-filling tools exist, but deployed products remain limited for tasks requiring genuine problem analysis and tailored solutions. Most production implementations handle only narrow, highly structured scenarios; broader applicability remains in the pilot phase. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like AI writing assistants and report generators exist and are used to draft answers to specific problems, but reliability varies with complexity and accuracy of source data, requiring human editing in most deployed workflows. |
Explain survey objectives and procedures to interviewees and interpret survey questions to help interviewees' comprehension.
44CI 30–59 · exposure 38 · augmentation 63 · importance 3.6/5 · click for rater detail
Explain survey objectives and procedures to interviewees and interpret survey questions to help interviewees' comprehension.
44| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Survey and market research firms remain relatively slow in automation adoption; most high-stakes survey work (quality assurance, complex populations) still relies on trained interviewers. Pilots of automated interviews exist but are not widely deployed in production across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Market research and survey firms have adopted chatbot and IVR-based data collection at moderate pace, but many sectors still rely heavily on human interviewers, especially for complex or sensitive topics. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist human interviewers by suggesting clarifications, translating jargon in real-time, or flagging potential misunderstandings, thereby raising interviewer productivity. However, the assistant role is limited by the need for human judgment in final interpretation and respondent rapport. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can provide interviewers with real-time scripting support, translation, and automated clarification suggestions, meaningfully improving consistency and efficiency while humans still manage the interaction. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Explaining objectives and interpreting questions requires real-time adaptation to interviewee comprehension level, cultural context, and clarification needs. While AI could deliver scripted explanations, reliably detecting misunderstanding and dynamically adjusting language in live interaction remains difficult; this task cannot currently achieve 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 3/5 | Conversational AI systems can explain survey objectives and clarify questions via chat or voice interfaces, but nuanced in-person rapport building and adaptive clarification for diverse respondents still require human judgment in many contexts. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Survey research has ethical and regulatory guidelines (IRB oversight, respondent consent, informed participation) that expect human accountability. While not a hard legal barrier, organizational friction and the requirement that response quality be defensible create meaningful adoption friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing is generally required for survey interviewing, though some sensitive surveys (health, legal, government) mandate trained human interviewers or informed consent processes that favor human interaction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | High-quality survey systems still employ trained human interviewers; deploying conversational AI for this task requires significant integration and oversight, plus ongoing human validation of interpretation quality. The all-in cost of an adequate AI system does not yet undercut the loaded wage of a survey interviewer. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated survey administration via chatbots or voice AI is substantially cheaper per completed interview than paying human interviewers, though some setup and oversight costs remain. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Conversational AI (chatbots, voice agents) can deliver explanations in controlled settings, but deployed survey systems still rely heavily on human interviewers for nuanced interpretation and rapport-building. No mature production system reliably handles the full range of comprehension issues and clarification requests that arise in live survey interviews. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Chatbot-based and IVR survey systems are deployed at scale for market research and telephone surveys, but they show higher dropout and comprehension error rates than trained human interviewers, especially with vulnerable or low-literacy populations. |
Contact individuals to be interviewed at home, place of business, or field location, by telephone, mail, or in person.
43CI 35–50 · exposure 30 · augmentation 50 · importance 3.9/5 · click for rater detail
Contact individuals to be interviewed at home, place of business, or field location, by telephone, mail, or in person.
43| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Market research and survey organizations are increasingly piloting automated scheduling and reminder systems, but widespread production deployment remains partial; many firms still rely on human-driven outreach for response quality. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Survey and research organizations have piloted automated contact systems (IVR, text reminders) but field/in-person contact work remains predominantly human-driven with slow AI penetration. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by generating contact lists, drafting personalized messages, tracking response status, and suggesting follow-up timing, meaningfully raising productivity of human coordinators while they retain relationship management and judgment calls. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help schedule appointments, send reminders, and manage contact lists, improving efficiency for the human interviewer initiating outreach. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Contacting individuals for interviews can be partially automated (emails, SMS scheduling), but reliably reaching and securing commitments from reluctant or hard-to-contact respondents requires human judgment, adaptation, and rapport-building that current AI cannot fully replicate at 50% time savings with equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Contacting individuals via phone, mail, or in person to arrange interviews involves scheduling and rapport-building that AI can partially assist with but not fully replace, especially for in-person or resistant contacts.》 |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal or regulatory barriers exist to automating initial contact and scheduling; organizational preference for human touch and response rate concerns create moderate friction, but no hard licensing requirement prevents AI-assisted or fully automated contact. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but survey response rates and data quality concerns create organizational preference for human contact, and privacy/consent regulations add some friction to automated outreach. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated email/SMS outreach and scheduling tools have very low per-contact costs compared to human interviewers' loaded wages, though human oversight for relationship-building and confirmations is still needed. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While automated outreach (robocalls, bulk emails) is cheap, achieving comparable response rates to human-led contact often requires additional human effort, keeping blended costs closer to parity. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Tools exist to send contact messages and schedule appointments, but deployed systems handle only routine outreach; many interviews require personalized follow-up, handling objections, and verification that today's AI struggles with reliably in production environments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated dialers, chatbots, and email/SMS scheduling tools exist and are used for outreach, but reaching people at home or in the field reliably still requires human follow-up for non-responders and skeptical subjects. |
Collect and analyze data, such as studying old records, tallying the number of outpatients entering each day or week, or participating in federal, state, or local population surveys as a Census Enumerator.
43CI 37–48 · exposure 45 · augmentation 75 · importance 3.6/5 · click for rater detail
Collect and analyze data, such as studying old records, tallying the number of outpatients entering each day or week, or participating in federal, state, or local population surveys as a Census Enumerator.
43| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Government agencies and census bureaus move slowly on automation; many still rely on manual tallying and paper records. While survey technology has digitized, actual displacement of interviewers and enumerators remains limited in practice. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Government statistical agencies are historically slow adopters of AI-driven fieldwork, though back-office data analysis is being modernized more quickly. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI tools can significantly assist interviewers by auto-populating fields from records, flagging data anomalies, and generating preliminary reports, allowing the human to focus on analysis, follow-up, and quality control. This is a strong augmentation scenario where AI handles routine data prep while the interviewer retains judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly help analyze collected data, flag anomalies, and automate tallying, greatly boosting productivity even though the human is still needed for actual data collection and rapport-building. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Data collection from existing records and tallying can be partially automated (scanning, OCR, database entry), but analysis of diverse data sources and integration with survey protocols requires human judgment. Automation could plausibly save 40–60% of time with setup, reaching the threshold for some workflows but not consistently. |
| Task automatability | claude-sonnet-5 | 3/5 | Data tallying and record analysis are highly automatable with software, but survey administration (especially in-person Census enumeration) requires human interaction that current AI cannot fully replace. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Census enumerators and survey data collectors are regulated positions with legal requirements for chain of custody, data integrity, and quality standards. Liability for data errors in federal/state surveys and strict compliance auditing create strong barriers to full substitution without human sign-off. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Government census work often has procedural, legal, and privacy requirements around who can collect and certify data, plus public trust considerations favoring human enumerators. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Automated data ingestion and tallying can be cheaper than manual entry per unit, but when factoring in setup, QA, and the human analyst required for survey quality assurance, total cost is roughly comparable to or slightly better than a full-time interviewer. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated data tallying is cheap, but the interview/survey collection component still requires human labor, keeping blended costs roughly comparable to human-only approaches. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist for data extraction (OCR, parsing), basic aggregation, and survey digitization, but error rates on unstructured records and the need for human validation in population surveys mean production systems still require significant oversight and manual correction. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Data aggregation tools are mature, but no deployed AI product independently conducts door-to-door census enumeration or equivalent field survey interviews at production scale. |
Perform patient services, such as answering the telephone or assisting patients with financial or medical questions.
42CI 30–54 · exposure 38 · augmentation 63 · importance 4.1/5 · click for rater detail
Perform patient services, such as answering the telephone or assisting patients with financial or medical questions.
42| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Healthcare and financial services sectors are piloting AI call-handling agents, but most organizations retain significant human capacity for patient-facing roles; adoption remains in the pilot-to-early-production phase rather than deep sector-wide displacement. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare administrative functions are adopting AI unevenly; many clinics and hospitals still rely heavily on human staff for patient-facing communication due to compliance and trust concerns. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist interviewers by pre-screening calls, summarizing patient histories, or drafting standard responses to common questions, meaningfully raising throughput; however, the complexity of medical and financial queries limits how much the human can offload to the system. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can efficiently pre-screen calls, draft responses to financial/insurance questions, and surface relevant patient data, significantly boosting interviewer productivity while humans retain final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Only narrow parts of this task—simple FAQ responses or call routing—could be automated end-to-end; the human judgment required for medical/financial questions and the relationship-building aspect of patient services prevent reaching 50% time savings at equal quality with current systems. |
| Task automatability | claude-sonnet-5 | 3/5 | AI chatbots and voice agents can handle routine scheduling, FAQ-type financial/insurance questions, and call triage, but complex or emotionally sensitive patient interactions still require human judgment and empathy, limiting full end-to-end automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Regulations like HIPAA govern handling of patient medical information, creating compliance overhead; patients also prefer human contact for sensitive financial or medical inquiries, adding organizational friction beyond pure technical capability. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement for this role, but HIPAA compliance, liability for medical misinformation, and patient preference for human contact create meaningful friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While basic AI phone agents have low per-call costs, integration, training on organization-specific knowledge, and required human oversight for escalations drive the total cost closer to a junior interviewer's wage for comparable output quality. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-driven phone/chat systems cost a fraction of a human interviewer's wage for routine inquiries, though integration with EHR/billing systems and oversight add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Chatbots and IVR systems handle basic call routing and simple FAQs in production, but they falter on nuanced medical or financial questions and lack the contextual understanding and empathy that patients expect; error rates remain material in real-world deployments. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Deployed healthcare call-center AI and IVR/chatbot systems handle basic patient inquiries in production, but error rates and escalation to humans remain common for medical questions requiring nuance or compliance sensitivity. |
Ensure payment for services by verifying benefits with the person's insurance provider or working out financing options.
41CI 32–50 · exposure 42 · augmentation 75 · importance 4.4/5 · click for rater detail
Ensure payment for services by verifying benefits with the person's insurance provider or working out financing options.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Healthcare and financial services are digitizing but adoption of full end-to-end verification automation remains inconsistent; many organizations pilot tools but retain human review due to error-cost sensitivity and regulatory caution. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Healthcare administrative functions are adopting automation for eligibility verification at a moderate pace, with RCM automation growing but human-assisted financing counseling still common; overall sector adoption is middling compared to fully digitized industries. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered insurance verification lookups and financing calculators meaningfully assist interviewers by reducing manual research time and flagging coverage gaps, allowing them to focus on negotiation and customer communication while staying in the loop. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered eligibility verification tools significantly speed up the benefits-checking portion of this task, letting interviewers focus more time on financing discussions with patients, meaningfully raising productivity even though full automation isn't achieved. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Verifying benefits with insurance providers can be partially automated through API integration and eligibility checks, but requires human judgment to interpret coverage details, handle exceptions, and negotiate financing options—tasks that are complex and context-dependent today. |
| Task automatability | claude-sonnet-5 | 3/5 | Verifying insurance benefits often involves structured data lookups (eligibility APIs, portals) that AI/automation can handle, but working out financing options requires judgment, negotiation, and exception handling that current AI cannot fully replicate end-to-end., so only partial time savings are achievable. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Insurance verification involves access to secure patient data, HIPAA compliance, and relationships with payers that carry legal liability if benefits are misrepresented; organizations typically require human sign-off on coverage determinations and financing agreements. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement for this specific task, but payer contracts, HIPAA compliance, and patient trust/preference for human interaction on financial matters create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Automated verification systems exist but require ongoing integration maintenance, human oversight for exceptions, and customer service follow-up, making total cost-per-task still competitive with or slightly higher than a human representative performing the task. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated eligibility checks are cheap per transaction, but the financing negotiation portion still requires human labor, oversight, and exception handling, keeping the blended cost roughly comparable to a human doing the full task today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Some platforms offer automated insurance verification tools and basic eligibility lookup, but these systems have material gaps in coverage complexity, require human review for accuracy, and cannot reliably handle all financing negotiation scenarios across different insurers. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Insurance eligibility verification tools and RCM software are deployed in production at many healthcare organizations, but financing arrangement conversations still typically require human staff, and error rates in automated verification remain non-trivial due to payer data inconsistencies. |
Identify and report problems in obtaining valid data.
40CI 35–45 · exposure 25 · augmentation 63 · importance 4.4/5 · click for rater detail
Identify and report problems in obtaining valid data.
40| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Survey research organizations, market research firms, and large-scale data collection operations have begun deploying automated quality checks, but many smaller and traditional interview-based settings still rely on manual review; adoption is spreading but not yet dominant. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Survey and research organizations are slow adopters of AI for qualitative data quality assessment, with pilots more common than production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered data quality dashboards and anomaly detection genuinely assist interviewers by highlighting suspicious patterns, missing values, and outliers in real time, allowing them to probe or re-collect data more efficiently while they retain judgment over which flags are actionable. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist by flagging statistical anomalies, missing values, or inconsistent responses, helping interviewers focus attention, though it doesn't replace their judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can flag some data quality issues (missing fields, inconsistent formats, statistical outliers) but struggles with contextual validity—determining whether a respondent's answer is genuinely problematic vs. legitimate variation requires nuanced judgment that AI systems lack reliability in today's deployments. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires judgment to notice inconsistencies, respondent behavior cues, and contextual data quality issues that current AI can only partially flag; full end-to-end automation with equal quality is not yet achievable.dynamic understanding.-> rated low. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Most interview workflows (surveys, research, HR) do not require licensed personnel to validate data; organizational friction around trust in automated flagging is moderate, and there are few regulatory mandates that a human must personally identify data problems. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational trust in human judgment for identifying subtle validity issues creates moderate friction to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated data quality monitoring is extremely cheap at scale (minimal inference cost per record) compared to paying interviewers to manually review and flag problems; integration costs are modest for existing data pipelines. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Automated data-quality checks are cheap, but the human judgment needed to interpret and report contextual problems still requires substantial oversight, keeping costs comparable rather than dramatically lower. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Data validation tools exist (pattern matching, schema checkers) but identifying *problems in obtaining* valid data—distinguishing data collection methodology issues, respondent reliability, and validity threats—requires contextual reasoning that current products handle only in narrow, pre-specified domains. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some survey/data QA tools flag anomalies or missing data statistically, but identifying and reporting nuanced data validity problems in interview contexts is not reliably done by deployed products. |
Identify and resolve inconsistencies in interviewees' responses by means of appropriate questioning or explanation.
34CI 30–39 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail
Identify and resolve inconsistencies in interviewees' responses by means of appropriate questioning or explanation.
34| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Interviewing roles remain concentrated in traditional customer-facing sectors (insurance, surveys, HR) with slower digital transformation. While some call centers use AI analytics to assist human interviewers, autonomous or near-autonomous inconsistency-resolution is not yet widely deployed in production. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Interviewing roles span market research, government surveys, and social services—sectors with mixed and generally slower AI adoption for live interactive judgment tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by highlighting potential contradictions and suggesting follow-up questions for the human interviewer to ask, raising their efficiency and consistency. However, the human must remain the primary decision-maker and conversationalist, limiting transformative impact. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can flag potential inconsistencies in transcripts or suggest follow-up questions, assisting interviewers, though it doesn't independently manage the interactive resolution process. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Identifying inconsistencies in responses requires contextual understanding and nuanced judgment, but current AI systems struggle with the conversational back-and-forth needed to resolve them authentically. While AI could flag potential inconsistencies in transcribed text, the dynamic questioning and explanatory dialogue required makes full end-to-end automation with 50% time savings unrealistic today. |
| Task automatability | claude-sonnet-5 | 2/5 | Detecting and probing inconsistencies in live human responses requires real-time judgment, follow-up questioning, and contextual understanding that current AI handles unreliably outside narrow scripted domains.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Many interviewing roles are governed by compliance and quality standards that require documented human judgment, and clients often prefer human interviewers for trust and legal defensibility. However, these are not absolute legal barriers, leaving room for augmentation or limited automation in some settings. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No strict licensing requirement typically governs this task, but organizations often prefer human judgment for sensitive interviews (e.g., surveys, research, hiring) creating moderate adoption friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration, training, and continuous human oversight to ensure quality and handle exceptions would be substantial relative to the task. The cost of AI infrastructure plus necessary human review and correction likely exceeds a human interviewer's hourly rate for this specialized work. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated conversational agents are cheap to run per interaction, but added complexity for inconsistency detection and quality assurance narrows the cost advantage compared to a human interviewer. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably perform this task autonomously in production. While conversational AI and NLP can detect textual contradictions, resolving them through appropriate questioning requires adaptive rapport-building and real-time decision-making that current systems do not handle reliably at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Chatbot-based survey and screening tools exist but they struggle with nuanced follow-up probing and inconsistency resolution compared to trained human interviewers, especially in unstructured contexts. |
Meet with supervisor daily to submit completed assignments and discuss progress.
15CI 0–30 · exposure 13 · augmentation 38 · importance 3.5/5 · click for rater detail
Meet with supervisor daily to submit completed assignments and discuss progress.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Organizations continue to rely on direct supervisor-employee meetings as a core management practice; there is no meaningful adoption of AI replacements for this supervisory touchpoint in any sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Interviewer roles (e.g., survey/research interviewers) are in sectors with modest AI adoption for supervisory workflows; daily meetings remain a standard human practice with little displacement. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could prepare a draft status report or summarize work completed prior to the meeting, offering minor assistance, but the supervisory meeting itself remains human-driven with limited scope for AI to augment the core interaction. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help prepare progress summaries, track completed assignments, and draft talking points, meaningfully aiding preparation for the meeting even though the meeting itself stays human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task fundamentally requires synchronous human-to-human communication with a supervisor to discuss progress and submit work. AI cannot autonomously initiate or conduct a meaningful supervisory meeting that involves bidirectional dialogue and judgment about work progress. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a synchronous interpersonal check-in requiring judgment about progress and context; AI could support parts (status summaries) but cannot conduct the actual meeting/discussion end-to-end.deleteCharAt.io_score.text.reduce.acted |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Organizational hierarchy and management responsibility legally and functionally require a human supervisor to conduct performance discussions and sign off on work assignments. This is a hard structural barrier. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No legal/licensing barrier, but organizational norms and management practice strongly favor human-to-human supervisory check-ins for accountability and relationship building. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task's core value is the interpersonal supervisory relationship and feedback loop, which has no AI analogue that would reduce cost below the human wage for that interaction. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Human-to-human daily check-ins are low-cost already; AI substitution would require automating an interpersonal interaction, offering little cost advantage and possibly added overhead. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably conducts supervisor meetings or replaces the human interaction component. While AI can draft status reports, it cannot meet with and discuss with a supervisor in a way that substitutes for the task as stated. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously meets with a supervisor to discuss progress; at best AI tools generate status reports or summaries that a human still presents. |
Supervise or train other staff members.
11CI 5–16 · exposure 9 · augmentation 50 · importance 3.9/5 · click for rater detail
Supervise or train other staff members.
11| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Supervision and training remain deeply human-centric functions in all sectors; adoption of AI for autonomous staff management is negligible. Cultural and legal norms strongly favor human accountability in people-management roles. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | While AI tools are being piloted for training content and onboarding, actual supervisory functions in this occupation see limited AI-driven transformation so far. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist supervisors by drafting feedback summaries, flagging performance trends, or scheduling training—useful for productivity but the human supervisor retains full decision-making responsibility. Augmentation is meaningful but limited to administrative support rather than transformative. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help generate training materials, quizzes, or feedback summaries, moderately aiding supervisors and trainers without replacing their core function. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Supervision and training of staff require real-time judgment, personalized coaching, performance feedback, and adaptive communication that demand human presence and contextual understanding. Current AI cannot reliably replicate the interpersonal dynamics, accountability, and mentoring required for this task at equal quality. |
| Task automatability | claude-sonnet-5 | 1/5 | Supervising and training staff involves interpersonal leadership, motivation, and contextual judgment that current AI cannot perform end-to-end; AI can support materials but not replace the supervisory relationship. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Legal liability, employment law, fiduciary duty to employees, and contractual obligations mean that a licensed, accountable human must sign off on performance evaluations, hiring decisions, and disciplinary actions. Organizations face both legal risk and employee relations concerns that strongly protect this function from automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Supervisory authority typically requires organizational accountability, HR/legal responsibility, and human judgment in personnel matters, creating strong structural barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI today cannot replace a human supervisor; any AI system would still require human oversight, vetting of decisions, and final accountability. The cost of integration, monitoring, and human review would exceed the cost of direct human supervision, making the AI solution more expensive. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Because AI cannot fully replace the supervisory function, human labor remains necessary, so cost comparisons favor humans except for content-generation subcomponents. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can draft training materials and suggest performance feedback templates, no deployed product reliably supervises or trains staff in production. Systems lack the situational judgment, empathy, and accountability that real supervision demands, and any attempt at autonomous staff management would face severe organizational and legal friction. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Products exist for generating training content or coaching scripts, but no deployed system autonomously supervises staff or manages performance in production. |
Related occupations — Office & Administrative Support
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.