Computer User Support Specialists
15-1232.00Median wage $61,860/yr717,190 employed (US)Rank #82 of 923 scored · top 9% by substitution
Provide technical assistance to computer users. Answer questions or resolve computer problems for clients in person, via telephone, or electronically. May provide assistance concerning the use of computer hardware and software, including printing, installation, word processing, electronic mail, and operating systems.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution50
Exposure43
Augmentation76
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
13%
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%44
panel mean rating 2.8/5 → substitution pressure 44/100
Technical feasibility todayw 20%41
panel mean rating 2.7/5 → substitution pressure 41/100
Cost vs. human wagew 15%49
panel mean rating 3.0/5 → substitution pressure 49/100
Adoption barriersw 20%inverted — strong barriers lower the score66
panel mean rating 2.4/5 (barrier strength) → substitution pressure 66/100
Sector adoption velocityw 10%54
panel mean rating 3.2/5 → substitution pressure 54/100
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.
Maintain records of daily data communication transactions, problems and remedial actions taken, or installation activities.
79CI 75–84 · exposure 75 · augmentation 88 · importance 3.5/5 · click for rater detail
Maintain records of daily data communication transactions, problems and remedial actions taken, or installation activities.
79| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | IT support and ticketing automation are mature market segments with rapid AI adoption in larger organizations and cloud-native environments. Automated logging and incident capture are increasingly standard in commercial helpdesk platforms. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | IT support and helpdesk functions are within the fast-adopting information/professional services sector where AI-assisted ticketing and auto-documentation are already widely deployed. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assists human support staff by auto-populating tickets, suggesting problem categories, and summarizing interactions, significantly reducing manual documentation burden while the specialist remains responsible for accuracy and oversight. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools substantially reduce the burden of writing up incident records by auto-drafting summaries and categorizing issues, letting specialists focus on resolution while review stays with the human. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can capture, structure, and log transaction data, problems, and remedial actions from support interactions with minimal human setup. However, some context-dependent judgment about severity or categorization may still require human review, preventing a full 5 rating. |
| Task automatability | claude-sonnet-5 | 4/5 | Logging structured data about transactions, problems, and fixes is largely templated documentation work that AI/automation (ticketing system integrations, auto-summarization from chat logs) can handle with high fidelity today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few hard barriers exist; record-keeping is internal operational work with no licensing requirement or strict regulatory mandate that a human must perform it. Organizational inertia and preference for human oversight represent modest friction. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There is no licensing, legal, or human-contact requirement for record-keeping tasks; organizations are already comfortable delegating documentation to software. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI logging and record-keeping costs (automated transcript processing, ticket generation) are a small fraction of the hourly wage of a support specialist who would manually maintain these records, yielding at least an order-of-magnitude savings per task. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated logging and AI-generated summaries cost a fraction of a cent per ticket compared to the minutes a human spends manually documenting, yielding a large cost advantage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (ticketing systems with AI logging, LLM-based ticket summarization, automated incident capture) already perform parts of this task reliably in production for many IT organizations. Integration with existing support tools is mature, though some edge cases still require human correction. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Many production ITSM tools (ServiceNow, Zendesk, Freshservice) already auto-generate and populate incident records and summaries from support interactions, though some manual review/correction still occurs. |
Read technical manuals, confer with users, or conduct computer diagnostics to investigate and resolve problems or to provide technical assistance and support.
74CI 57–90 · exposure 70 · augmentation 75 · importance 3.9/5 · click for rater detail
Read technical manuals, confer with users, or conduct computer diagnostics to investigate and resolve problems or to provide technical assistance and support.
74| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Tech-sector companies and digital-native firms have rapidly deployed AI chatbots and support agents in production; major software vendors, cloud providers, and SaaS platforms are already using LLM-based support at scale, reflecting fast, deep adoption in information and services. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | IT and tech support is a digitized, tech-forward sector with fast adoption of AI chat and diagnostic assistants across many companies already in production. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants meaningfully augment human support specialists by drafting responses, suggesting diagnostics, and retrieving manual content; specialists remain in the loop to handle edge cases and judgment calls, boosting their throughput and accuracy significantly. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly speeds up manual searches, symptom-based troubleshooting suggestions, and knowledge retrieval, meaningfully boosting specialist productivity while they retain final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Current AI agents can read technical manuals, parse diagnostic data, and provide troubleshooting guidance end-to-end. LLMs excel at technical documentation retrieval, logical troubleshooting trees, and diagnostic interpretation; combined with tool-use, these can resolve or escalate issues with substantial time savings (often 70%+) versus human support specialists. |
| Task automatability | claude-sonnet-5 | 3/5 | AI chatbots and diagnostic tools can handle routine troubleshooting and manual lookup, but complex or novel hardware/software issues and nuanced user communication still require human judgment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal or regulatory barriers exist to automating technical support; customer preference for human contact and desire to escalate complex cases pose some friction, but self-service and AI-first tiers are widely accepted and already normalized in practice. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for general IT support, though some organizational trust and liability concerns around sensitive systems create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference cost per resolved ticket is orders of magnitude cheaper than loaded wages of a support specialist (typically $40–70k annually), especially when factoring in training and overhead; even with integration and oversight, the ratio easily favors automation. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-driven tier-1 support can be cheaper per interaction, but integration, oversight, and escalation costs keep overall savings moderate rather than order-of-magnitude for the full task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products like ChatGPT, Claude, and specialized helpdesk AI (e.g., Zendesk, ServiceNow with AI plugins) already perform frontline technical support and diagnostics reliably in production for routine and moderately complex issues; edge cases and novel problems still require human escalation. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Deployed helpdesk AI (e.g., chatbots, automated ticketing triage) exists and works for common issues, but escalations to humans for complex diagnostics remain frequent, indicating narrow reliable scope. |
Answer user inquiries regarding computer software or hardware operation to resolve problems.
68CI 61–75 · exposure 62 · augmentation 88 · importance 3.8/5 · click for rater detail
Answer user inquiries regarding computer software or hardware operation to resolve problems.
68| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | IT support is a digitized, information-centric function with rapid AI adoption. Thousands of organizations deploy AI chatbots, ticketing automation, and knowledge-base integration in production; this is among the fastest-moving sectors for AI displacement. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | IT support functions in tech-forward and corporate sectors have rapidly adopted AI chatbots and virtual assistants, reflecting fast adoption typical of information/professional services. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists human support specialists by instantly surfacing knowledge base articles, suggesting solutions, drafting responses, and handling routine queries before human review. This transforms throughput and reduces cognitive load while keeping humans in complex decision loops. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools like knowledge-base search, automated diagnostics, and suggested resolutions significantly speed up human support specialists' work even when full automation isn't used. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can handle routine troubleshooting questions (password resets, common error codes, basic connectivity issues) and answer predictable software/hardware inquiries, but complex, context-dependent problems requiring deep system diagnostics, user environment investigation, or multi-step reasoning still require human intervention. Current systems achieve partial automation but not the 50% time savings at equal quality threshold end-to-end. |
| Task automatability | claude-sonnet-5 | 4/5 | Most user inquiries follow common patterns (password resets, software errors, driver issues) that current LLM-based chatbots and troubleshooting agents can resolve or triage with substantial time savings, though complex hardware diagnostics still need humans. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers prevent AI support deployment. Customer satisfaction and service quality expectations create some friction, and some organizations prefer human escalation paths, but nothing legally mandates human support for software/hardware inquiries. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement and low liability risk for routine troubleshooting, though some organizations retain human-contact preferences for sensitive systems or VIP users. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-powered support (cloud-based chatbots, LLM APIs) costs significantly less than human support specialists per inquiry handled, especially at scale. The loaded cost of a support specialist ($40–60k+ annually) vastly exceeds the per-query inference and integration cost of modern systems. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated chat/voice support and knowledge-base-driven resolution cost a fraction of a human agent's loaded wage per interaction, though oversight and escalation infrastructure add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Chatbot support systems and AI-powered help desk tools are deployed in production at many organizations, but they handle a narrow scope of common queries and frequently escalate or provide incorrect answers for edge cases. Material error rates and coverage limitations prevent mature, fully reliable performance. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | AI-powered helpdesk chatbots, virtual IT assistants, and automated ticketing/triage systems are already deployed at scale in many enterprises, though escalation to human agents remains common for edge cases. |
Enter commands and observe system functioning to verify correct operations and detect errors.
68CI 55–81 · exposure 62 · augmentation 75 · importance 3.6/5 · click for rater detail
Enter commands and observe system functioning to verify correct operations and detect errors.
68| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Information and technology sectors have already deeply adopted automated monitoring, log aggregation, and synthetic monitoring agents; displacement is measurable and widespread in cloud and enterprise infrastructure roles. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | IT operations and support functions are moderately fast adopters of monitoring/AIOps tools, though full end-to-end automation of verification tasks remains uneven across firms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-driven alerting, anomaly detection, and command suggestions significantly amplify human specialist productivity by filtering noise, highlighting anomalies, and automating routine checks, though humans interpret context and prioritize remediation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven log analysis, anomaly detection, and command suggestion tools significantly speed up a support specialist's ability to verify system function and spot errors. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI and automated monitoring systems can execute commands, log outputs, and compare system states against expected baselines to detect many operational errors and anomalies with high consistency. However, some edge cases and novel failure modes may still require human judgment, preventing a full 5. |
| Task automatability | claude-sonnet-5 | 3/5 | AI agents can execute diagnostic commands and interpret common error patterns for well-known systems, but novel or complex environments still require human judgment and hands-on troubleshooting. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or licensing barriers exist for deploying automated monitoring; organizations mainly face adoption friction from legacy system integration and preference to maintain human oversight, but nothing legally prevents automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational risk aversion around system changes and the need for accountability in production environments creates moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated monitoring and command execution systems operate continuously at marginal cost per task and are orders of magnitude cheaper than paying a human specialist to manually execute commands and watch system logs 24/7. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated monitoring reduces some labor costs, but integration, tuning, and human oversight for edge cases keep total cost roughly comparable to a support specialist's time for many organizations. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature products like automated monitoring tools, log analysis systems, and synthetic monitoring platforms (e.g., Datadog, New Relic, Splunk) reliably perform command execution and error detection in production environments at scale, though human interpretation of novel alerts remains common. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AIOps and monitoring tools (e.g., automated log analysis, scripted diagnostics) are deployed in production, but they typically flag anomalies for human review rather than fully replacing verification and error detection. |
Develop training materials and procedures, or train users in the proper use of hardware or software.
64CI 55–72 · exposure 58 · augmentation 88 · importance 3.3/5 · click for rater detail
Develop training materials and procedures, or train users in the proper use of hardware or software.
64| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Medium adoption: tech companies and large enterprises are piloting AI-driven training platforms and content generation, but many smaller IT support teams still rely on manual procedures; pilots outnumber full production rollouts. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | IT support and corporate training functions show moderate AI adoption for content generation, with pilots common but full replacement of trainers still rare. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists support specialists by drafting procedures, generating training slides, and fielding basic user questions, freeing humans to focus on complex troubleshooting and relationship-building while maintaining quality oversight. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI substantially speeds up drafting manuals, FAQs, and training scripts, letting support specialists focus on delivery and tailoring, a clear productivity boost. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can generate substantial training materials (documentation, slides, video scripts) and execute basic training delivery at scale, achieving >50% time savings on content creation and initial delivery, though human judgment on pedagogical approach and individualized coaching remains valuable. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft training materials, manuals, and step-by-step procedures effectively, but delivering interactive training and adapting to user questions still requires human involvement for full task completion., so only part of the task meets the 50% bar. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal or licensing barriers exist; training delivery is not a regulated profession. Main friction comes from organizational preference for human rapport and real-time Q&A, plus the need to verify training effectiveness, but these are not hard blockers. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There are no licensing or regulatory requirements for creating training materials or training users on software/hardware, so no hard barriers exist. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-generated training materials and automated delivery (chatbots, self-paced video) cost far less per user than human instructor time; operational costs drop dramatically when scaled, though oversight and customization add overhead. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Drafting content with AI is cheap, but the full task including live user training, revisions, and organizational context integration keeps overall costs comparable to human-led efforts. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist that generate training content and deliver basic tutorials (LLMs, learning platforms with AI), but they show material limitations in adapting to diverse user skill levels and handling nuanced software interactions; deployment is growing but not yet mature across organizations. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Generative AI tools are commonly used in production to draft documentation and training content, but live training delivery and troubleshooting-specific customization still rely on human trainers. |
Read trade magazines and technical manuals, or attend conferences and seminars to maintain knowledge of hardware and software.
62CI 50–75 · exposure 50 · augmentation 88 · importance 3.0/5 · click for rater detail
Read trade magazines and technical manuals, or attend conferences and seminars to maintain knowledge of hardware and software.
62| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Tech companies and large IT departments are beginning to deploy AI-assisted tools for knowledge management and content curation, but adoption is still in the pilot and early production phase rather than mainstream or deep. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | IT support roles are in a fast-adopting sector (professional/technical services) where AI research and summarization tools are already commonly integrated into workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered summarization, alerting, and keyword extraction significantly enhance a specialist's ability to consume and filter relevant technical material faster, raising their effective reading velocity and allowing focus on interpretation and application rather than raw content intake. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools substantially enhance a specialist's ability to stay current by surfacing, summarizing, and organizing technical content, greatly increasing efficiency while the human retains judgment on application. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can help summarize technical content and extract key information from trade publications, but the task requires sustained judgment about relevance to evolving job needs and strategic selection of learning priorities that remains fundamentally human. This covers only a fraction of the full task of 'maintaining knowledge' in a meaningful professional sense. |
| Task automatability | claude-sonnet-5 | 4/5 | AI can summarize trade publications, technical manuals, and conference content, and can synthesize updates on hardware/software far faster than manual reading, meeting the time-saving threshold for the research/knowledge-acquisition portion of this task. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There are minimal legal or licensing barriers to automating content consumption and summarization. Organizational culture and the preference that specialists develop their own contextual judgment provide light friction, but nothing prevents substitution. |
| Adoption barriers | claude-sonnet-5 | 1/5 | This is a self-directed learning/professional development task with no licensing, liability, or human-contact requirements blocking AI assistance. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated filtering and summarization of trade content is very cheap compared to paying a human specialist to read extensively; however, some human curation time remains necessary, making it not quite a full order of magnitude cheaper. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI summarization and knowledge-aggregation tools cost a small fraction of the analyst's time compared to manually reading magazines or attending events. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI-powered content summarization and filtering tools exist and are deployed, but they cannot replace the full workflow of reading, contextualizing, and integrating new technical knowledge into professional practice. Products work on narrow content streams but lack the judgment to identify what matters for a specific role. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like AI research assistants and summarization tools reliably digest technical documents today, but attending conferences/seminars and synthesizing nuanced professional context is not yet fully replicated by deployed systems. |
Prepare evaluations of software or hardware, and recommend improvements or upgrades.
51CI 46–55 · exposure 50 · augmentation 75 · importance 3.4/5 · click for rater detail
Prepare evaluations of software or hardware, and recommend improvements or upgrades.
51| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | IT organizations are piloting automated diagnostic and evaluation tools, but adoption of AI-driven recommendations remains inconsistent; most remain human-in-the-loop workflows rather than autonomous decision systems, reflecting the cost of errors in production environments. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | IT support functions are adopting AI tools for research and reporting at a moderate pace, with pilots more common than full production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments support specialists by automating log analysis, generating performance baselines, and suggesting upgrade candidates, substantially accelerating the evaluation process while the specialist retains judgment on business priorities and risk. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly speeds up research, comparison drafting, and report writing, making it a strong productivity aid even though final recommendations require human judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can assist with evaluations by analyzing system performance data, generating test reports, and identifying common issues, but the task typically requires context-specific judgment about organizational needs, cost-benefit tradeoffs, and integration risks that currently demand human expertise. This represents roughly half the work automated with significant setup. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can gather specs, compare products, and draft evaluation reports, but synthesizing organizational context, testing compatibility, and making final judgment calls still requires human oversight, so only part of the task meets the 50% threshold. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Weak-to-moderate barriers: recommendations must typically be reviewed and approved by IT managers or architects before implementation, and incorrect upgrades can cause operational failures that create liability concerns, but no legal licensing requirement mandates human sign-off. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this task, though organizational trust and accountability for hardware/software decisions creates some friction against pure automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted evaluation tools reduce per-task analysis time and integration costs remain material; while inference is cheap, the need for human validation, custom configuration for each environment, and the cost of misrecommendations keeps total cost-per-recommendation closer to or slightly below a specialist's loaded wage rather than substantially cheaper. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI can cut research and drafting time significantly, but human review, testing, and validation still add substantial cost, keeping the ratio only moderately favorable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist (asset management platforms, automated testing tools, diagnostic systems) that can generate evaluation reports and flag performance issues, but they operate within narrow scopes and require human validation of recommendations; no mature end-to-end solution reliably recommends upgrades independent of human oversight. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like Copilot and ChatGPT-based tools can assist in drafting comparisons and researching specs, but no deployed product autonomously performs full hardware/software evaluations reliably at scale. |
Conduct office automation feasibility studies, including workflow analysis, space design, or cost comparison analysis.
50CI 32–67 · exposure 45 · augmentation 75 · importance 3.2/5 · click for rater detail
Conduct office automation feasibility studies, including workflow analysis, space design, or cost comparison analysis.
50| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | IT departments and professional services firms increasingly use automation for preliminary analysis and data synthesis, but comprehensive feasibility studies are still often conducted manually or with only partial tool support. Pilot and tool adoption are common; full replacement is less common. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | IT support and business analysis functions are adopting AI tools for reporting and analysis at a moderate pace, though feasibility studies specifically remain a niche, less-digitized task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assists specialists significantly in data gathering, cost calculation, scenario modeling, and report generation. Humans remain essential for validating assumptions, managing stakeholder requirements, and making final trade-off decisions, making this a strong augmentation scenario. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly speed up drafting workflow diagrams, cost comparisons, and report writing, giving specialists strong productivity gains while they retain oversight and judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can automate large portions of feasibility studies through workflow analysis (process mining, task extraction), cost comparison (data aggregation, calculation), and preliminary space design recommendations. Significant setup required, but most data processing and analytical steps can be completed with 50%+ time savings using current tools like LLMs and data analysis platforms. |
| Task automatability | claude-sonnet-5 | 2/5 | This task requires site-specific data gathering, stakeholder interviews, and judgment-based synthesis that AI cannot fully perform end-to-end, though it can assist with parts like cost modeling or drafting reports. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Feasibility studies often inform significant organizational or capital decisions, creating organizational friction and demand for human accountability. No strict licensing barrier exists, but client expectations and internal governance often require a qualified human to own the analysis and recommendations. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational trust and the need for physical space assessment and stakeholder buy-in create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-based analysis tools have low per-task inference costs, and bulk workflow/cost data processing is orders of magnitude cheaper than hiring analysts for the same work. Human oversight and validation remain necessary, but the marginal cost of analysis is substantially below loaded specialist wages. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI can cheaply generate draft analyses or cost comparisons, the human effort needed for data collection, verification, and stakeholder engagement keeps overall costs comparable to human-led work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist (workflow analysis software, cost estimation tools, even some space-planning AI) but they typically require human validation of assumptions, cost data, and design constraints. Most deployed systems excel at data collection and initial analysis rather than end-to-end autonomous feasibility study generation. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously conducts full feasibility studies including workflow analysis and space design; AI is used piecemeal for spreadsheets or writing, not the integrated study itself. |
Inspect equipment and read order sheets to prepare for delivery to users.
47CI 26–67 · exposure 41 · augmentation 63 · importance 3.2/5 · click for rater detail
Inspect equipment and read order sheets to prepare for delivery to users.
47| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large IT asset management and logistics providers have begun piloting automated inspection, but uptake remains uneven. Most small and mid-sized support teams still rely on manual inspection; adoption is emerging but not yet mainstream in the broader IT support industry. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | IT support functions are moderately digitized but this specific physical inspection/logistics task lags behind ticketing and software-based support automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can prepare a detailed inspection report and flag anomalies, leaving a human to make final judgment calls and handle exceptions. This significantly accelerates the review process while keeping human oversight in place for unusual or high-risk orders. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help by cross-referencing order sheets, flagging discrepancies, and generating checklists, aiding the human who still must physically inspect equipment. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI systems can automate most of this task: computer vision can inspect equipment for visible defects, OCR can read order sheets, and inventory systems can verify item presence. The integration of these components into a streamlined workflow can easily achieve >50% time savings, though final human sign-off may still be prudent in some contexts. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical inspection of equipment cannot be done by AI today; reading order sheets and matching data is automatable but is only a small slice of the overall task.subsequently the end-to-end task requires physical presence.rating reflects that limitation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing or legal requirement mandates human inspection of IT equipment deliveries. Organizational adoption may be slowed by risk-aversion and user preference for human verification, but these are soft barriers, not hard ones. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing barrier, but the physical nature of inspecting and handling equipment creates a structural barrier to remote AI substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference for vision and OCR is inexpensive ($0.01–0.10 per inspection), while human inspection costs $15–30 per order. Even accounting for integration and occasional human oversight, the AI cost is substantially lower—typically 10–20% of human labor cost. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the physical handling and inspection component, so there is no viable AI-only cost comparison; a human must still perform the core task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed computer vision and OCR products exist and work reliably for standardized orders and equipment, but real-world scenarios involve variable lighting, damaged labels, and complex multi-item orders where error rates remain material. Some logistics firms use automated inspection, but it is not yet standard practice at scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical equipment inspection and delivery preparation; this remains a manual, in-person task. |
Oversee the daily performance of computer systems.
42CI 40–45 · exposure 34 · augmentation 88 · importance 4.0/5 · click for rater detail
Oversee the daily performance of computer systems.
42| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | IT organizations widely deploy automated monitoring and alerting systems today. However, humans remain in the loop for decision-making and complex troubleshooting, representing augmentation rather than replacement. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | IT operations and support functions in tech-forward sectors have rapidly adopted AI-driven monitoring and alerting tools (AIOps), reflecting fast adoption within IT/professional services contexts. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI-driven monitoring systems significantly augment support specialists by automating alert generation, log analysis, and trend detection, allowing humans to focus on complex problem-solving and strategic system improvements. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered dashboards, anomaly detection, and predictive alerts significantly enhance a specialist's ability to monitor system performance, surfacing issues faster than manual review while the human remains responsible for interpretation and action. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Monitoring system metrics and logs can be partially automated with alerting tools, but interpreting anomalies, prioritizing issues, and deciding on interventions requires human judgment. Current AI cannot reliably oversee daily performance end-to-end without significant human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | AI monitoring tools can flag anomalies and metrics but the ongoing oversight, triage, and judgment calls about system health still require human interpretation and decision-making, limiting full automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Organizations often prefer human oversight for critical systems due to liability and error-cost asymmetry. Regulatory and compliance requirements in some sectors mandate human accountability, though no blanket legal barrier prevents automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this task, but organizational trust and accountability for system uptime create some resistance to fully removing human oversight, especially in regulated or mission-critical environments. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Automated monitoring tools reduce labor but require infrastructure setup, integration, and continuous human oversight. The combined cost of monitoring software, maintenance, and required staff review approaches or exceeds the cost of a dedicated support specialist. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Monitoring software licensing plus integration and human review costs are substantial; while automation reduces manual checking, the overall oversight function still requires paid staff time, keeping cost savings moderate rather than order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Monitoring and alerting products exist (e.g., Datadog, New Relic, generic SIEM tools), but they flag issues rather than independently oversee systems. Reliable oversight still requires human review and decision-making in production environments. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Deployed monitoring/observability platforms (Datadog, Splunk, AI-assisted anomaly detection) are widely used in production, but they augment rather than replace the human overseer, and often produce false positives requiring human review. |
Refer major hardware or software problems or defective products to vendors or technicians for service.
39CI 36–41 · exposure 25 · augmentation 75 · importance 3.4/5 · click for rater detail
Refer major hardware or software problems or defective products to vendors or technicians for service.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | IT support functions show moderate AI adoption in ticketing and initial triage, but deep production deployment of escalation and vendor-referral automation remains in the pilot phase across most organizations rather than standard practice. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | IT support functions are adopting AI chatbots and ticket routing tools at a moderate pace, with pilots and partial deployments common in tech-forward organizations but not universal. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly assist by suggesting severity classifications, identifying pattern matches to known vendor issues, and recommending appropriate service channels, enabling support specialists to make faster, better-informed referral decisions while retaining final judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by pre-classifying issues, drafting vendor communications, and suggesting escalation paths, improving specialist efficiency while they retain final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI can help triage and classify problems into categories, but the task requires nuanced judgment about problem severity, vendor capabilities, and technical context that AI systems handle inconsistently. End-to-end automation would require reliable problem diagnosis and vendor matching at 50%+ time savings, which is not yet demonstrated at scale. |
| Task automatability | claude-sonnet-5 | 2/5 | This is largely a triage/routing decision requiring judgment about severity and vendor relationships; AI can help classify tickets but the actual referral and vendor coordination still needs human decision-making and accountability. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Support organizations have incentives to automate routing, but customer preferences for human judgment on escalation, vendor contractual relationships, and liability concerns around misrouting critical issues create meaningful friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational trust, vendor relationship management, and liability for misrouted critical issues create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-assisted triage and routing tools cost roughly comparable to a support specialist's time spent on problem assessment and vendor selection, though savings depend heavily on integration depth and the quality of vendor databases maintained. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-assisted ticket triage tools are relatively cheap to run, but human oversight is still needed for accurate escalation decisions, keeping costs roughly comparable rather than dramatically lower. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some ticketing systems and support platforms use AI for initial classification and routing, but these have material error rates in correctly identifying major vs. minor issues and selecting appropriate vendors. No mature, production-scale systems reliably perform this task without significant human oversight. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Ticketing systems with AI-based categorization and routing exist, but reliable end-to-end determination of 'major' vs minor issues and escalation to the right vendor/technician is not yet a mature, widely deployed autonomous product. |
Install and perform minor repairs to hardware, software, or peripheral equipment, following design or installation specifications.
36CI 30–41 · exposure 25 · augmentation 75 · importance 3.8/5 · click for rater detail
Install and perform minor repairs to hardware, software, or peripheral equipment, following design or installation specifications.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | IT organizations use remote deployment and patch management widely, but physical hardware repair and on-site diagnostics remain largely human-driven; adoption of AI-driven repair diagnosis is nascent and pilot-focused rather than production-dominant. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | IT support functions are adopting AI chatbots for tier-1 software issues, but hardware installation/repair remains physically bound and has seen little robotic automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can assist significantly with remote diagnostics, knowledge bases, step-by-step guided repair procedures, and automated software deployment, allowing technicians to work faster and resolve more complex issues. These tools are already in use in production IT environments. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI diagnostic assistants and knowledge bases significantly speed up troubleshooting, parts identification, and step-by-step repair guidance for human technicians. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Installation of hardware and peripherals can be partially automated through scripted deployment systems, but minor repairs requiring diagnosis, physical manipulation, and contextual judgment remain difficult for current AI. Remote software installation is feasible, but diagnosing and fixing hardware issues end-to-end falls short of the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical installation and hardware repair require manual dexterity and on-site presence that current AI cannot perform; software configuration guidance can be automated but the full task including physical hands-on work cannot. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some regulatory and liability concerns exist around installation without oversight, especially for critical systems, and organizations often maintain human verification requirements. Customer preference for human interaction on repairs adds organizational friction but not hard legal barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement typically exists, but physical presence, liability for hardware damage, and equipment access create real friction against pure AI substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Remote software installation and routine deployments can be cost-effective through automation, but the need for technician dispatch to physical sites and oversight of repairs keeps overall cost-per-incident comparable to human labor. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply generate troubleshooting steps, but a human technician is still required to physically execute installation and repair, so overall cost savings versus a full human technician are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Automated deployment tools exist for software and some configurations, but diagnosing and repairing hardware faults with visual inspection or hands-on troubleshooting requires human intervention. No deployed AI system reliably performs the full repair workflow independently. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI chatbots and diagnostic tools exist for software troubleshooting guidance, but no deployed product physically installs or repairs hardware/peripherals end-to-end. |
Set up equipment for employee use, performing or ensuring proper installation of cables, operating systems, or appropriate software.
35CI 32–38 · exposure 25 · augmentation 75 · importance 3.9/5 · click for rater detail
Set up equipment for employee use, performing or ensuring proper installation of cables, operating systems, or appropriate software.
35| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large organizations are piloting AI-assisted support (chatbots, automated documentation) but actual deployment of hardware setup automation remains limited to remote software provisioning in mature sectors like finance and tech; on-site physical installation lags significantly. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | IT support functions are adopting automation tools (imaging, zero-touch deployment) at a moderate pace, but the physical hardware setup portion lags behind due to its non-digital nature. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can substantially assist by generating step-by-step installation guides, diagnosing compatibility issues, automating software deployment, and troubleshooting—allowing technicians to work faster and more reliably while remaining responsible for physical execution and verification. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven deployment scripts, automated OS imaging, and software provisioning tools significantly speed up the non-physical portions of this task, letting technicians focus on physical installation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could potentially guide or document setup workflows, the task requires physical cable installation and hardware-specific OS/software configuration that demands hands-on intervention. Current AI systems lack the embodied robotics capability to perform cable management and equipment assembly end-to-end, though they could assist with step-by-step instruction. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical cable installation and hands-on hardware setup cannot be performed by current AI systems, though software/OS configuration can be scripted or automated with tools like imaging systems and remote deployment software.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Equipment setup typically requires verification and sign-off by IT support staff for liability and compliance; organizational IT governance often mandates human oversight. However, there are no absolute licensing barriers preventing greater AI-guided or semi-autonomous workflows. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational reliance on physical presence and hands-on troubleshooting creates practical friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Even with AI-assisted documentation and remote guidance, a technician's on-site labor remains necessary for physical installation. AI reduces time spent on knowledge retrieval but cannot eliminate the core labor cost, keeping total cost-per-task in the same range as traditional support. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Physical labor and on-site presence still require human wages; software automation tools reduce some cost but the physical component keeps overall cost comparable to or only modestly cheaper than a technician. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI products reliably perform full hardware setup, OS installation, and cable management as an autonomous service. Chatbots can provide guidance, but actual setup in heterogeneous enterprise environments with varied hardware remains primarily human-performed. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Deployed products (e.g., MDM, imaging tools, RPA scripts) automate software installation but physical setup and troubleshooting still require a human on-site; no product does the full task end-to-end. |
Modify and customize commercial programs for internal needs.
35CI 32–38 · exposure 25 · augmentation 75 · importance 2.9/5 · click for rater detail
Modify and customize commercial programs for internal needs.
35| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | IT and software development organizations are actively adopting AI coding assistants in pilot and production settings, but customization of commercial software remains a domain where human developers retain primary responsibility. Adoption is growing but uneven, with many organizations still relying on traditional vendor support or consulting. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | IT support functions are in sectors with moderate-to-fast AI tool adoption (coding copilots, ticketing automation), though customization work specifically remains a pilot-stage use case. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI code assistants significantly augment developer productivity for software customization by generating boilerplate, suggesting modifications, and identifying bugs. Developers using these tools can customize programs faster and more reliably, while retaining control over architectural and business-logic decisions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI coding assistants and configuration tools meaningfully speed up scripting, debugging, and documentation for customization tasks, keeping the specialist in the loop for judgment and testing. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Modifying and customizing commercial programs requires domain knowledge of the specific software, understanding of internal business logic, and often interaction with legacy systems. While AI can assist with code generation and debugging, end-to-end customization—including requirements gathering, architectural decisions, and testing—still demands significant human judgment and context that current AI systems cannot reliably handle independently. |
| Task automatability | claude-sonnet-5 | 2/5 | Customizing commercial software requires understanding specific business context, existing configurations, and testing changes safely, which current AI can assist with but not fully execute end-to-end without significant human oversight and iteration. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Commercial software modifications may involve licensing restrictions, vendor support implications, and liability concerns if customizations break system functionality. However, there are no hard legal barriers preventing AI-assisted customization, though organizational risk management and vendor agreements create moderate friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists, but vendor support agreements, software warranties, and organizational risk tolerance for unauthorized modifications create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI code generation tools cost relatively little per inference, but the customization task requires skilled developers to verify, test, and integrate changes. The total cost (AI + developer oversight) is comparable to or potentially exceeds hiring a developer directly, especially for complex or mission-critical modifications. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI can reduce time spent on some coding subtasks, the need for human validation, testing, and domain knowledge means overall costs remain comparable to or only modestly cheaper than a human specialist. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI code assistants (Copilot, Claude) can generate code snippets and help with specific modifications, but deployed products do not reliably customize complex commercial software for arbitrary internal needs without human oversight. The error rate and need for human verification remain high, and the task's contextual complexity exceeds what production systems handle autonomously. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI coding assistants can help write scripts, macros, or config changes, but reliably diagnosing needs and safely modifying commercial programs in production environments is not yet a mature, deployed product capability at scale. |
Confer with staff, users, and management to establish requirements for new systems or modifications.
31CI 23–39 · exposure 25 · augmentation 63 · importance 3.7/5 · click for rater detail
Confer with staff, users, and management to establish requirements for new systems or modifications.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Requirements gathering is a high-touch, relationship-driven activity where organizations strongly prefer human specialists. Adoption of AI for autonomous requirements elicitation is minimal; the task involves embedded organizational knowledge and stakeholder trust that resist automation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | placeholder |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by transcribing meetings, generating initial requirement summaries, identifying gaps in specifications, and organizing stakeholder inputs—helping a human specialist work more efficiently—but the human specialist remains the primary driver of the conferencing and negotiation process. |
| Augmentation potential | claude-sonnet-5 | 4/5 | placeholder |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | This task fundamentally requires understanding nuanced organizational context, interpersonal dynamics, and implicit stakeholder needs that AI cannot reliably elicit end-to-end. While AI can help draft requirements or analyze specifications, the core activity—conferring with multiple parties to establish what is actually needed—remains dependent on human judgment and relationship management. |
| Task automatability | claude-sonnet-5 | 2/5 | This task requires live interpersonal negotiation, reading organizational politics, and synthesizing ambiguous stakeholder needs, which current AI cannot conduct end-to-end without heavy human involvement.'},'feasibility':{'rating':2,'rationale':'No deployed product autonomously conducts requirements-gathering conversations with staff and management; AI meeting assistants only summarize or transcribe rather than conduct the conferring itself.'}, but rated as needed below.'},'placeholder':1}}}, |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant organizational and relationship barriers exist: stakeholders generally expect a human specialist to understand their context and concerns, liability falls on the organization if requirements are misunderstood, and management typically requires human accountability for system design decisions. |
| Adoption barriers | claude-sonnet-5 | 2/5 | placeholder |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI could assist with note-taking and synthesis, but the cost of building and maintaining an AI agent for requirements conferences plus human oversight would likely exceed the loaded wage of having a specialist conduct interviews and meetings directly. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | placeholder |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product performs this task autonomously in production. AI can assist with documentation and summarization, but actually conducting requirements-gathering conversations with staff, users, and management requires real-time negotiation and relationship skills that current systems cannot reliably execute without continuous human oversight. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | placeholder |
Hire, supervise, and direct workers engaged in special project work, problem-solving, monitoring, and installation of data communication equipment and software.
6CI 0–11 · exposure 8 · augmentation 50 · importance 2.9/5 · click for rater detail
Hire, supervise, and direct workers engaged in special project work, problem-solving, monitoring, and installation of data communication equipment and software.
6| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Hiring and worker supervision remain strictly human functions across all sectors; no meaningful automation or AI adoption displacement has occurred because legal accountability cannot be transferred to machines. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | While IT support functions see moderate AI tool adoption, the managerial/supervisory component of this task shows little to no AI displacement in practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist with resume screening, project tracking, performance metrics, and scheduling, providing useful support to human managers making final hiring and supervisory decisions, though these remain human-centric tasks. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with scheduling, tracking project status, or drafting job postings and performance summaries, but the core supervisory judgment remains human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | This task involves complex human resource decisions (hiring, supervision, direction) and judgment about worker capabilities and project fit. While AI could assist with candidate screening or project tracking, the core supervisory and hiring decisions require human accountability and cannot be automated end-to-end with 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 1/5 | Hiring, supervising, and directing employees involves interpersonal judgment, accountability, and legal responsibility that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Legal and regulatory requirements (employment law, discrimination compliance, liability for worker direction) and organizational necessity for a human manager create hard barriers. Human managers must legally sign off on hiring and employment decisions. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Hiring and supervisory authority carry legal, HR, and liability requirements that mandate human decision-makers and accountability. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Hiring and supervision require human managers with domain expertise and legal accountability; AI cannot replace this function cost-effectively because the liability and compliance costs of autonomous management decisions would exceed human manager wages. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this managerial function, so cost comparison favors the human entirely since AI cannot deliver the output. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No production system can reliably perform hiring, worker supervision, and project direction autonomously. These tasks require legal accountability, employment law compliance, and human judgment that current AI systems cannot operationally handle without extensive human oversight. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously hires or supervises workers; HR-adjacent AI tools only assist with screening or scheduling, not actual management authority. |
Related occupations — Computer & Mathematical
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.