Technical Writers
27-3042.00Median wage $90,390/yr45,500 employed (US)Rank #66 of 923 scored · top 7% by substitution
Write technical materials, such as equipment manuals, appendices, or operating and maintenance instructions. May assist in layout work.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution55
Exposure50
Augmentation75
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
15 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
20%
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%52
panel mean rating 3.1/5 → substitution pressure 52/100
Technical feasibility todayw 20%48
panel mean rating 2.9/5 → substitution pressure 48/100
Cost vs. human wagew 15%52
panel mean rating 3.1/5 → substitution pressure 52/100
Adoption barriersw 20%inverted — strong barriers lower the score70
panel mean rating 2.2/5 (barrier strength) → substitution pressure 70/100
Sector adoption velocityw 10%51
panel mean rating 3.0/5 → substitution pressure 51/100
Task breakdown (15 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.
Arrange for typing, duplication, and distribution of material.
100CI 100–100 · exposure 100 · augmentation 63 · importance 3.8/5 · click for rater detail
Arrange for typing, duplication, and distribution of material.
100| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | This task has already been largely automated across information work, finance, and professional services. Email, cloud storage, and document management are standard infrastructure; displacement of manual typing and distribution labor is complete in digitized sectors. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Office and administrative automation for document handling and distribution is already deeply and widely adopted across virtually all industries using standard software tools. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | While full automation dominates, AI can assist humans in organizing, labeling, or categorizing material before distribution, and can suggest distribution lists or optimal formats—meaningful but not transformative assistance. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and automation tools significantly speed up arranging typing, formatting, and distributing materials, though a human may still coordinate final decisions on distribution lists or formatting nuances. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | This task is almost entirely automatable today. Typing can be replaced by text generation or copy-paste, duplication is handled by cloud storage and file replication, and distribution is automated via email systems, document management platforms, or scheduled batch processes. These steps easily meet the ≥50% time-saving bar. |
| Task automatability | claude-sonnet-5 | 5/5 | Typing, duplication, and distribution logistics are almost entirely digital workflow tasks (document generation, file sharing, automated distribution lists) that off-the-shelf software and AI tools already handle with minimal human input. |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | No licensing, authorization, or legal requirement mandates human involvement in typing, copying, or distributing documents. Organizational adoption is straightforward with minimal regulatory friction. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There are no licensing, liability, or regulatory requirements tied to typing or distributing documents; this is purely administrative and freely automatable. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Modern cloud and automation tools (email, cloud storage, batch distribution) cost pennies per document, making them orders of magnitude cheaper than a human managing typing, copying, and hand-distribution tasks. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated typing, file duplication, and distribution via software costs a fraction of a cent per action compared to any human labor time spent on these clerical tasks. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Production systems for typing, duplication, and distribution are mature and widely deployed: email systems, cloud storage (Google Drive, OneDrive, SharePoint), document management systems, and print-on-demand services handle these workflows at scale in real organizations daily. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Document management systems, cloud sharing platforms, and automated distribution/email tools are mature, widely deployed products that reliably perform these functions at scale today. |
Maintain records and files of work and revisions.
90CI 80–100 · exposure 87 · augmentation 75 · importance 4.2/5 · click for rater detail
Maintain records and files of work and revisions.
90| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Technical writing teams and software organizations have nearly universal adoption of version control and document management systems; these are standard practice in professional environments, particularly in tech, publishing, and corporate sectors where technical writers work. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Technical writing and documentation workflows are highly digitized, with version control and automated file management already standard practice in most professional and tech-adjacent sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI systems assist technical writers by automatically organizing files, suggesting metadata, detecting changes, and recommending archive strategies, while the human retains control over what to keep, discard, or organize—significantly raising productivity in administrative overhead management. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can auto-generate changelogs, tag revisions, and summarize edit histories, significantly boosting a writer's efficiency in maintaining organized records while they retain oversight. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Maintaining records and files of work and revisions is highly automatable. Version control systems, document management platforms, and file organization workflows can be fully automated or AI-assisted with minimal human intervention, easily achieving >50% time savings at equal or better quality through automatic versioning, metadata tagging, and change tracking. |
| Task automatability | claude-sonnet-5 | 4/5 | Version control systems, document management platforms, and AI-assisted metadata tagging can automate most record and revision tracking with minimal human oversight, easily saving over half the time versus manual logging. |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | There are no legal, regulatory, or organizational barriers preventing automation of file maintenance and revision tracking. No licensing requirement or human authorization is mandated for this purely administrative task, and organizations actively adopt these systems. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, regulatory, or human-contact requirements restrict automating administrative record-keeping tasks. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | The cost of automated file management and version control systems (often included in broader enterprise software or available open-source) is orders of magnitude cheaper than paying a technical writer to manually organize, store, and track revisions of their own work. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated version control and file management systems cost a small fraction of the labor cost of manual record-keeping, though some integration and setup overhead remains. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed products (Git, SharePoint, Confluence, cloud storage with versioning, document management systems) reliably perform this task at scale in production environments across thousands of organizations today, with robust capabilities for tracking changes, maintaining histories, and organizing files. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Mature tools like Git, Confluence, SharePoint, and CMS platforms already handle revision history and file tracking reliably in production across most organizations. |
Edit, standardize, or make changes to material prepared by other writers or establishment personnel.
74CI 64–84 · exposure 67 · augmentation 88 · importance 4.0/5 · click for rater detail
Edit, standardize, or make changes to material prepared by other writers or establishment personnel.
74| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Technical writing is digitized and information-sector native, but adoption remains in the pilot-to-early-production phase. Some firms integrate AI editing into workflows; many still maintain human-centric processes due to quality concerns and organizational inertia. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Publishing, tech, and documentation-heavy industries have rapidly integrated AI writing/editing assistants into standard workflows over the past two years. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI editing tools significantly accelerate human editors by automating routine corrections, suggesting rewordings, and enforcing style consistency, allowing writers to focus on substantive improvements and clarity. This assistive function is widely deployed and measurably raises editor productivity. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI editing tools are widely used to assist technical writers by flagging inconsistencies, suggesting rewrites, and enforcing style guides, substantially boosting editing throughput while humans retain final control. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automatically detect and correct grammar, style, and formatting errors with high accuracy, handling roughly half the editorial workload. However, substantive edits requiring contextual judgment, brand voice alignment, and domain expertise still require human oversight, preventing full automatability at equal quality. |
| Task automatability | claude-sonnet-5 | 4/5 | LLMs can edit for style, grammar, consistency, and terminology adherence across documents with substantial time savings, though final judgment calls on technical accuracy still benefit from human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No regulatory or licensing requirement mandates human editorial oversight; adoption depends mainly on organizational confidence and quality assurance protocols rather than legal barriers. Humans may be preferred for final sign-off on critical content, but this is friction rather than a hard requirement. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There is no licensing or legal requirement for a human to perform copyediting or style standardization; organizations can freely adopt automated tools. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference costs for editing are negligible ($0.01–0.10 per document), while human editors command $25–50/hour loaded wages. Even accounting for oversight and integration, AI is 10–100× cheaper per equivalent edit pass. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | AI editing tools cost a small fraction of a technical writer's hourly wage and can process large volumes of text near-instantly, making the cost differential very large. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Production-grade AI editing tools (e.g., Grammarly, Claude, GPT-4 with structured prompts) reliably standardize content and apply style guides at scale. Minor gaps exist around nuanced tone and organization, but mature products perform core editing tasks reliably in enterprise environments. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed tools (Grammarly, Acrolinx, ChatGPT-based editing workflows, style-guide enforcement plugins) reliably perform copyediting and standardization tasks in production content pipelines today. |
Develop or maintain online help documentation.
65CI 55–75 · exposure 58 · augmentation 88 · importance 3.9/5 · click for rater detail
Develop or maintain online help documentation.
65| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Tech companies and software vendors are increasingly experimenting with AI-assisted documentation generation, but adoption remains pilot-heavy rather than deeply embedded; financial services and regulated sectors move more slowly. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Tech writing sits within software/IT industries with fast AI tool adoption, and AI-assisted authoring tools are already common in tech companies' documentation pipelines. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at rapidly drafting outline, examples, and boilerplate sections, allowing technical writers to focus on complex technical accuracy, architecture explanation, and customer-specific customization, substantially raising human productivity. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI substantially speeds up drafting, reformatting, updating, and summarizing technical content while writers retain control over accuracy and final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can draft, organize, and generate large portions of online help content from code or specifications, potentially saving 40–60% of writing time, but human review for accuracy, tone consistency, and domain validation remains essential and time-consuming. |
| Task automatability | claude-sonnet-5 | 4/5 | LLMs can draft, structure, and update online help content from source materials (release notes, code comments, product specs) with substantial time savings, though final QA and accuracy checks still need human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Documentation is not legally regulated or gated behind licensing, and customer preference for human-authored technical accuracy is not a hard barrier; organizational friction around AI-generated content and company style guides pose mild friction but no formal barriers. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or human-contact requirements block AI-assisted documentation; organizations widely permit AI tools in this workflow. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI inference and editing overhead is modest compared to human hourly rates, but integration with version control, testing pipelines, and quality assurance processes adds overhead; costs are roughly comparable when human oversight is factored in. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Generating and updating drafts via AI is far cheaper per unit of content than a full writer-hour, though human review still adds cost, keeping it just under the top tier. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | LLM-based documentation tools (e.g., GitHub Copilot, specialized doc generators) exist in production, but they still require significant human oversight for correctness, completeness, and integration with existing doc systems, limiting reliable end-to-end automation. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like doc-generation copilots, Confluence/Notion AI, and API-doc generators are deployed in production, but they still require editing for accuracy, tone, and completeness, so scope remains narrow to moderate. |
Select photographs, drawings, sketches, diagrams, and charts to illustrate material.
62CI 52–72 · exposure 58 · augmentation 88 · importance 4.0/5 · click for rater detail
Select photographs, drawings, sketches, diagrams, and charts to illustrate material.
62| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Technical writing and publishing sectors show modest, uneven adoption of AI image-curation tools; most firms still rely on manual selection and stock-image services, and integration into content workflows remains limited compared to sectors like software or finance. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Technical writing and documentation teams increasingly use AI tools for visuals and drafting, but full-scale production adoption specifically for illustration selection is still uneven. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI image search, tagging, and layout suggestions meaningfully boost the speed at which human technical writers can curate and organize illustrations, allowing them to focus on relevance and context rather than manual browsing. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI substantially speeds up brainstorming, sourcing, and generating candidate visuals, letting writers focus on final curation and accuracy checks. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can generate or select candidate illustrations from existing libraries and suggest layouts, cutting manual curation time substantially; however, subject-matter alignment, aesthetic fit, and context-specific quality checks require human review, so full end-to-end automation with 50%+ time savings is achievable only with significant editorial setup. |
| Task automatability | claude-sonnet-5 | 4/5 | AI can review content and suggest or generate relevant images/diagrams and identify appropriate existing visuals with human review, saving significant time though final selection judgment often still needed. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal or regulatory barriers exist to automating image selection itself; organizational friction (stakeholder preference for human editorial judgment, brand consistency requirements) provides modest friction but does not prevent substitution. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or regulatory requirement mandates human selection of illustrative material; adoption is limited only by quality preferences, not hard barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-powered image search and curation tools are inexpensive per task, but oversight and human selection add meaningful cost; the combined cost is roughly comparable to paying a junior technical writer or illustrator for the same work. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-assisted image search/generation is far cheaper per unit than a human manually sourcing or commissioning illustrations, though some oversight cost remains. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Image-search systems and AI-assisted design tools exist in production (e.g., stock-photo APIs, design templates, generative image suggestions), but they still require meaningful human judgment to filter for accuracy and appropriateness, limiting fully autonomous performance at scale. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like image generation and multimodal search tools exist and are used in documentation workflows, but reliable, context-aware selection of the exact right technical illustration still has notable error rates. |
Organize material and complete writing assignment according to set standards regarding order, clarity, conciseness, style, and terminology.
61CI 46–75 · exposure 62 · augmentation 88 · importance 4.7/5 · click for rater detail
Organize material and complete writing assignment according to set standards regarding order, clarity, conciseness, style, and terminology.
61| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Technical writing teams are adopting AI assistants for drafting and editing in pilots and some production workflows, but adoption is middling; conservative industries and heavily regulated sectors move slowly, while software-heavy organizations adopt faster. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Technical writing sits within software/professional services and publishing sectors, which show fast, deep AI tool adoption for drafting and content generation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI provides strong productivity gains by automating outlining, style standardization, terminology checking, and first-draft generation, allowing human writers to focus on domain expertise, clarity refinement, and validation—a pattern already common in practice. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI substantially boosts productivity for structuring, style conformance, and terminology consistency while humans retain responsibility for accuracy and final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can assist with organizing content structure, enforcing style/terminology consistency, and drafting sections, but the end-to-end task requires human judgment on clarity priorities, domain expertise validation, and adjustments to audience needs that current systems cannot reliably perform without substantial oversight. |
| Task automatability | claude-sonnet-5 | 4/5 | LLMs can draft, organize, and edit technical content to house style with significant time savings, though final review for accuracy and domain nuance still requires human oversight for equal quality. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Organizations often require human sign-off on technical documentation for liability and accuracy reasons, and customer-facing or regulated documentation may demand human authorship, though internal or lower-stakes writing faces fewer formal barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for technical writing; main friction is organizational quality control and accuracy verification, especially in regulated industries (medical, legal, aerospace documentation). |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While inference is cheap, the integration overhead, custom style training, and mandatory human review to ensure technical accuracy and adherence to domain standards makes the all-in cost comparable to or higher than hiring technical writers for complex documents. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI drafting and editing costs are a small fraction of a technical writer's loaded wage per page produced, even accounting for review time. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like AI writing assistants and style-checking tools exist in production (Grammarly, specialized technical writing platforms), but they typically handle formatting and grammar rather than the full organizational and standards-compliance aspects reliably, and still require material human review. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed tools (e.g., AI writing assistants, documentation copilots integrated into docs platforms) reliably generate and restructure technical content in production workflows today, though not fully unsupervised. |
Assist in laying out material for publication.
59CI 46–72 · exposure 58 · augmentation 88 · importance 3.9/5 · click for rater detail
Assist in laying out material for publication.
59| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Information and publishing sectors have adopted AI-assisted layout tools, but adoption remains mixed. Pilots are common, but full autonomous layout automation is less prevalent than in other technical writing domains, reflecting the creative and quality-assurance demands of publication. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Publishing and technical writing sectors show moderate AI tool adoption for layout and design assistance, with pilots and partial integration common but full automation less prevalent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists technical writers by automating routine formatting, applying templates, suggesting layouts, and generating style-consistent drafts, which allows writers to focus on content and design refinement. This augmentation is substantial even when human judgment remains essential for final publication decisions. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI-powered layout and design tools substantially enhance a technical writer's productivity by automating repetitive formatting tasks while the writer retains control over final design decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate parts of layout tasks (template application, basic formatting, style consistency checks) but typically requires human oversight for design decisions, custom layouts, and final quality control. Current systems handle routine formatting at roughly 50% time savings with moderate setup. |
| Task automatability | claude-sonnet-5 | 4/5 | Layout tasks involving formatting, templating, and document design are largely automatable using desktop publishing tools with AI features and AI-assisted layout generation, though final judgment on visual/publication fit still requires human review.imestamp |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Layout quality and brand consistency often require organizational sign-off and human judgment. Customer expectations and liability concerns around publication errors create moderate friction; no hard legal barrier exists, but companies often prefer human accountability for final publication-ready output. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There are no licensing, regulatory, or legal barriers preventing AI tools from assisting with or performing publication layout tasks. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI layout tools reduce costs for routine formatting, but integration, oversight, and human refinement of output still add substantial expense. For non-standardized layouts requiring design judgment, the all-in cost remains comparable to or higher than hiring a technical writer. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-assisted layout tools significantly reduce the time and labor cost compared to manual layout work, though some integration and review costs remain. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like Markdown converters, automated layout tools, and AI-assisted design systems exist and perform basic layout tasks, but they operate with limitations in complex or custom layouts and require human review. Production use is common in standardized document contexts but not fully autonomous. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like Adobe InDesign with AI features, Canva, and AI document generators can assist with layout, but reliable end-to-end automated publication layout at production quality still requires human oversight and correction. |
Review published materials and recommend revisions or changes in scope, format, content, and methods of reproduction and binding.
48CI 41–55 · exposure 42 · augmentation 75 · importance 3.6/5 · click for rater detail
Review published materials and recommend revisions or changes in scope, format, content, and methods of reproduction and binding.
48| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Publishing and documentation organizations have begun adopting AI-assisted review tools, but adoption remains mixed; many still rely on human reviewers as primary gatekeepers. Pilot projects are common, but production-only autonomous recommendation is rare. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Technical writing and publishing sectors have moderate AI tool adoption for drafting/editing, though production and binding decisions remain a slower-adopting, more traditional process. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at highlighting typos, formatting inconsistencies, and structural gaps, significantly accelerating the review process when a human decides which changes to implement. The human remains the decision-maker on scope and binding, but AI's assistive value is substantial. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially aids the review process by flagging content and clarity issues, suggesting revisions, and speeding up editorial iteration, even though final judgment on format/binding stays human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can identify surface-level revision suggestions (grammar, clarity) and flag formatting inconsistencies, but evaluating scope changes, reproduction methods, and binding decisions requires domain expertise and stakeholder context that AI struggles to synthesize reliably. The task rarely meets the 50% time-saving threshold end-to-end. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can review documents for clarity, consistency, and content issues and suggest revisions effectively, but judgments about scope, reproduction methods, and binding require contextual/business knowledge and physical production awareness that current systems handle poorly.:contentReference[oaicite:0]{index=0} |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | No legal requirement mandates human review, but organizational quality standards and liability concerns around publishing errors create moderate friction. Publishers often retain final human sign-off on revisions, especially for scope and binding decisions affecting production. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational workflows and reliance on human editorial judgment for scope/format decisions create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI inference for document review is inexpensive, but integration, oversight, and human review of recommendations add overhead. The all-in cost approaches that of a junior technical reviewer, making the ratio roughly comparable rather than dramatically cheaper. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-assisted review is cheap for text-level feedback, but the physical production and format aspects still require human coordination with printers/vendors, keeping overall cost comparable rather than dramatically cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like GPT-based document analyzers and grammar checkers exist and can flag issues reliably, but they often miss nuanced scope decisions or binding/reproduction trade-offs that require human technical judgment. Deployments are typically assistive, not autonomous. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like grammar/style checkers and LLM-based editing assistants are deployed widely for content review, but recommending changes to format, reproduction, and binding methods is rarely handled by mature production tools. |
Review manufacturer's and trade catalogs, drawings and other data relative to operation, maintenance, and service of equipment.
43CI 34–52 · exposure 45 · augmentation 75 · importance 3.7/5 · click for rater detail
Review manufacturer's and trade catalogs, drawings and other data relative to operation, maintenance, and service of equipment.
43| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While technical and professional services sectors are adopting AI, the specialized, regulation-heavy nature of equipment documentation review and the embedded need for qualified human sign-off mean adoption remains slow; most firms still rely on human technical writers as primary reviewers rather than AI-led automation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Technical writing and documentation functions in manufacturing/engineering sectors show slower AI adoption compared to pure information-sector roles, with pilots more common than full production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI tools for document summarization, automated cross-referencing, drawing annotation, and data extraction can substantially augment a human technical writer's productivity in locating and organizing manufacturer data, enabling faster review cycles while the human retains judgment and authority over accuracy. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can significantly speed up review and summarization of catalogs and technical data, helping writers extract relevant information faster while they retain judgment over final content and accuracy. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can assist with document scanning, metadata extraction, and cross-referencing catalogs and drawings, achieving substantial time savings on information retrieval and organization. However, contextual judgment about relevance, completeness assessment, and linking technical data to specific equipment variants requires human expertise, preventing full end-to-end automation at consistent quality. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can ingest and summarize catalogs, drawings metadata, and technical documents quickly, but interpreting engineering drawings and cross-referencing physical equipment specs still requires human verification for accuracy and completeness. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Technical documentation review for equipment maintenance and service often falls under regulated domains (aerospace, medical, automotive) where liability, certification, and regulatory sign-off typically require a licensed or qualified human reviewer to take responsibility for accuracy and completeness. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but liability for errors in equipment maintenance documentation creates moderate incentive for human review and sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Document processing and drawing analysis via AI incurs moderate inference and integration costs, but the human oversight needed to validate extracted information, resolve ambiguities, and ensure completeness keeps total cost near human wage levels, especially for safety-critical equipment documentation. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI can cut review time substantially for text-heavy catalogs, but the need for human verification of technical accuracy narrows the cost advantage compared to fully automated tasks. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed products exist for document processing and technical illustration analysis (OCR, technical drawing parsing, metadata extraction), but error rates on complex or aged catalogs remain material, and integration into real technical-writing workflows requires manual quality checks and human curation. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Document summarization tools exist, but reliably extracting and synthesizing technical specs from diverse drawings and trade catalogs (often non-standardized PDFs/images) remains error-prone in production use today. |
Study drawings, specifications, mockups, and product samples to integrate and delineate technology, operating procedure, and production sequence and detail.
41CI 32–50 · exposure 33 · augmentation 75 · importance 3.8/5 · click for rater detail
Study drawings, specifications, mockups, and product samples to integrate and delineate technology, operating procedure, and production sequence and detail.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Information-sector organizations have begun piloting AI drafting tools, and some larger tech firms experiment with automated documentation generation, but widespread production deployment remains limited. Most technical writing still relies on human authorship with AI as a suggestion or draft aid. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Technical writing and documentation sectors show moderate AI tool adoption (e.g., AI-assisted drafting), but integration with physical/technical review processes lags in many industries like manufacturing and engineering. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at assisting technical writers by auto-generating initial drafts from specs, summarizing product samples, and flagging inconsistencies between drawings and sequences. Writers remain in the loop for judgment and accuracy, but AI substantially accelerates the information-gathering and initial-composition phases. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially assists by drafting procedural text, organizing specifications, and generating documentation structure from source materials, while humans validate accuracy against samples and drawings. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can extract and summarize information from drawings and specs, but integrating complex technical details into coherent sequences requires domain expertise, judgment about priority and audience, and validation against actual product behavior. Current systems struggle with the synthesis and error-checking needed for reliable documentation. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can process and summarize drawings, specs, and sample descriptions to draft procedural content, but integrating physical mockups and product samples often requires visual/physical inspection and judgment beyond current standard tooling.》Roughly half the synthesis work is automatable with setup.》Time savings are meaningful but not full end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | While documentation automation is not legally restricted, liability concerns (errors in operating procedures can cause injury or equipment damage) and organizational norms favoring subject-matter-expert sign-off create moderate friction. Customers and regulators often expect human accountability. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for technical writing, but organizational quality control and accuracy requirements for technical documentation create moderate friction against pure automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI vision and language inference costs are falling but still require significant human oversight, revision, and validation to meet documentation standards. For typical technical-writing workflows, the all-in cost (inference + integration + human review) remains comparable to or exceeds direct technical-writer labor. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI tools reduce time spent reviewing documents and drafting summaries, but human verification against physical samples and specs adds oversight cost, keeping total cost roughly comparable to a human-only workflow in many cases. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Tools like vision models can parse diagrams and Claude-style systems can draft documentation fragments, but no production system reliably transforms raw technical artifacts into complete, accurate, sequenced procedure docs without substantial human review and correction. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Multimodal AI products can parse technical drawings and specs with growing accuracy, but reliable production-grade interpretation of mockups and physical samples alongside spec integration is still narrow and error-prone in deployed tools. |
Analyze developments in specific field to determine need for revisions in previously published materials and development of new material.
40CI 30–50 · exposure 33 · augmentation 75 · importance 3.7/5 · click for rater detail
Analyze developments in specific field to determine need for revisions in previously published materials and development of new material.
40| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Technical writing is a knowledge-work domain with moderate digitization; adoption of AI for monitoring and change analysis is still in pilot phase at most organizations, with most teams relying on manual field monitoring and stakeholder feedback. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Technical writing and content-adjacent fields are adopting AI tools at a moderate pace, with pilots for content gap analysis and update tracking emerging but not yet standard practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can substantially enhance productivity by automatically scanning field sources, summarizing technical changes, flagging outdated sections, and drafting preliminary revision suggestions—leaving the technical writer to validate, prioritize, and finalize updates with much higher efficiency than manual monitoring alone. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly aids by monitoring sources, summarizing changes, and drafting comparison points, letting writers focus judgment on prioritization and final decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can monitor field developments through document scanning and summarization, determining whether revisions are truly needed requires nuanced judgment about technical accuracy, user impact, and strategic priorities—tasks where AI struggles without domain expertise and human oversight to avoid false positives or missed critical updates. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can help scan and summarize field developments and flag likely gaps versus existing docs, but determining true need for revision requires domain judgment, prioritization, and stakeholder context that current systems can't fully replace. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Technical writers typically work within organizations with editorial review processes and subject-matter expert sign-off requirements; these oversight workflows create some friction but do not legally mandate a human, allowing partial automation of data-gathering phases. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational trust and accuracy concerns mean humans typically still own the decision of what needs updating, creating moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools (monitoring, summarization APIs) add cost ($20–100/month per technical writer) on top of human analyst time; the human still performs the core decision-making work, so all-in cost remains dominated by labor rather than AI savings. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI can cheaply process large volumes of field literature/changelogs, but human review and validation of implications still adds substantial cost, making overall savings moderate rather than dramatic. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No production systems reliably perform independent monitoring-to-revision-decision analysis at scale; existing tools can flag documents for review or summarize changes, but humans must make final determinations about material updates, limiting end-to-end deployment. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some products (research assistants, doc-diffing tools, AI summarizers) exist to surface changes, but no mature deployed system reliably performs this end-to-end analytical judgment task in production. |
Draw sketches to illustrate specified materials or assembly sequence.
37CI 30–44 · exposure 30 · augmentation 63 · importance 3.7/5 · click for rater detail
Draw sketches to illustrate specified materials or assembly sequence.
37| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Technical writing and documentation are moderately digitized, but adoption of AI sketch generation is still in pilot phases across most sectors. Publishing, software, and engineering firms show some early adoption, but production deployment remains limited. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Technical writing and documentation sectors are adopting AI for text drafting faster than for precise technical illustration, which remains a niche, slower-adopting subtask. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI sketch generation significantly assists technical writers by rapidly producing draft visuals for assembly sequences and material illustrations, allowing writers to focus on refining accuracy and compliance rather than manual drawing from scratch. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help generate draft concepts, reference imagery, or rough layouts that a technical writer or illustrator refines into accurate final sketches. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI can generate diagram sketches from text descriptions, but technical accuracy, proper assembly sequencing, and compliance with technical standards typically require human review and iteration. End-to-end automation without significant human oversight falls short of the 50% time-saving bar. |
| Task automatability | claude-sonnet-5 | 2/5 | AI image generation can produce illustrative sketches but struggles with precise, technically accurate assembly diagrams matching exact specifications, limiting true end-to-end automation.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Technical documentation may require sign-off by qualified personnel and liability concerns around incorrect assembly sequences create organizational friction. However, no hard legal requirement prevents AI assistance or preliminary sketch generation from being deployed. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but accuracy and liability concerns in technical documentation (e.g., safety-critical assembly instructions) create moderate quality-control friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI generation cost is low per sketch, but integration time, prompting, review cycles, and human correction of inaccuracies make the all-in cost comparable to having a technical illustrator produce the work with fewer iterations. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools are cheap per generation but require significant human correction and technical validation, so total cost including oversight is not dramatically lower than a skilled technical illustrator. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI image generation tools (DALL-E, Midjourney) can produce sketches, and some CAD-adjacent systems exist, but they struggle with precise technical specifications, correct spatial relationships, and standards compliance. Products exist but with material error rates requiring human correction. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some AI image tools can generate rough concept sketches, but no mature deployed product reliably creates accurate technical assembly diagrams matching engineering specs in production workflows.' |
Interview production and engineering personnel and read journals and other material to become familiar with product technologies and production methods.
34CI 30–38 · exposure 25 · augmentation 63 · importance 4.0/5 · click for rater detail
Interview production and engineering personnel and read journals and other material to become familiar with product technologies and production methods.
34| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Technical writing remains a human-centric profession with slow AI adoption; most firms still rely on human writers to conduct interviews and research, and there is limited evidence of production-scale AI replacement for this interviewing task. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Technical writing and documentation functions in tech/engineering sectors are moderately adopting AI tools for drafting and research, though the interview-based knowledge-gathering step remains largely human-driven. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist by summarizing technical documents, pre-processing journal articles, and organizing research notes, which accelerates the research phase without replacing the human interviewer's critical role in extracting knowledge from personnel. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools substantially help writers prepare interview questions, summarize journal articles, transcribe and organize interview notes, and synthesize findings, meaningfully boosting productivity while humans retain the interview role. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI cannot reliably conduct genuine interviews with personnel—it lacks the interactive, contextual problem-solving and follow-up probing needed to extract nuanced technical knowledge. While AI can summarize journals and documents, replacing the interview component (which is central to the task) remains infeasible at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | The 'read journals and other material' portion can be significantly augmented by AI research/summarization tools, but the core act of interviewing engineering personnel to extract tacit, undocumented knowledge requires human interpersonal interaction and judgment AI cannot replace end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Organizations typically prefer human technical writers conduct interviews to maintain relationship-building, capture tacit knowledge, and ensure accountability for accuracy. Some regulatory or quality documentation standards may also implicitly require human judgment in the research phase. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No formal licensing barrier exists, but organizational trust, need for personnel rapport, and access to proprietary internal engineering knowledge create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | An AI system capable of conducting quality technical interviews and contextual research would require significant custom training, integration, and human oversight—likely approaching or exceeding the cost of a human technical writer performing the research directly. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply assist with document synthesis, but the interviewing component still requires a human writer's time and relationship-building, keeping overall cost comparable to human-only workflows. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs end-to-end interviewing of domain experts. AI can assist with document summarization, but conducting effective technical interviews at production quality requires human expertise and judgment that current systems cannot replicate. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI products can transcribe and summarize interviews or documents, but no deployed product autonomously conducts substantive technical interviews with engineers to extract nuanced product knowledge reliably. |
Confer with customer representatives, vendors, plant executives, or publisher to establish technical specifications and to determine subject material to be developed for publication.
31CI 30–32 · exposure 25 · augmentation 63 · importance 3.5/5 · click for rater detail
Confer with customer representatives, vendors, plant executives, or publisher to establish technical specifications and to determine subject material to be developed for publication.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Technical writing organizations use AI for documentation drafting and editing, but stakeholder conference automation remains rare and experimental. Most firms still prioritize human presence in specification meetings. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Technical writing and publishing sectors show moderate AI tool adoption (drafting, summarization) but the specific stakeholder-negotiation function remains largely unautomated in practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by transcribing meetings, flagging ambiguities, and suggesting specification frameworks, allowing the human technical writer to conduct the conversation more efficiently and document requirements faster. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by transcribing meetings, summarizing requirements, drafting specification documents, and preparing questions, significantly boosting the writer's efficiency while they remain in the loop. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft meeting summaries and extract requirements from conversations, the interactive negotiation, stakeholder consensus-building, and judgment calls about scope are inherently dialogic. Current systems cannot reliably manage the full back-and-forth needed to establish specifications without human oversight at each step. |
| Task automatability | claude-sonnet-5 | 2/5 | This task requires live interpersonal negotiation, relationship-building, and real-time clarification of ambiguous requirements across multiple stakeholders, which current AI cannot conduct autonomously end-to-end."},'rating reflects only partial support like meeting summarization."},' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Customer and vendor relationships typically favor human contact and sign-off on specifications; there is modest organizational friction around removing the human from stakeholder engagement, though no strict legal barrier prevents recording or AI assistance. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement, but strong organizational and relationship-based friction exists since stakeholders expect human judgment and accountability in scoping decisions. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for meeting support (transcription, summarization) are cheap, but the human technical writer must still attend and lead the conversation; integration cost and required human review make the all-in cost approach that of the human alone, with marginal AI savings. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Human labor is still required to conduct the actual conferring; AI tools add cost for transcription/summarization without replacing the core interpersonal task, so savings are marginal. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably conducts multi-stakeholder specification conferences autonomously. Some meeting transcription and note-taking tools exist, but they do not make autonomous decisions about technical scope or resolve conflicting requirements between stakeholders. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI meeting assistants can transcribe and summarize conversations, but no deployed product independently confers with stakeholders to establish specifications and negotiate scope. |
Observe production, developmental, and experimental activities to determine operating procedure and detail.
25CI 20–30 · exposure 20 · augmentation 50 · importance 3.8/5 · click for rater detail
Observe production, developmental, and experimental activities to determine operating procedure and detail.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing and experimental environments lag in AI adoption overall. While some digitization exists, replacing on-site human observation with autonomous AI is not a common practice in most sectors; adoption remains in early or pilot phases. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and R&D environments have historically slower AI adoption for physical observation tasks compared to office-based information work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by analyzing video feeds, flagging anomalies, or organizing logged data for a human technical writer to review and interpret. This augmentation—reducing manual review time and highlighting key details—is feasible and moderately useful, though the human observer remains essential for judgment. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help structure, summarize, and cross-reference observational notes or sensor logs into procedural documentation, aiding the writer once data is captured. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Observing complex manufacturing or experimental activities requires real-time spatial awareness, contextual judgment, and domain expertise. While AI can analyze videos or logs after the fact, autonomous observation of live production with sufficient detail and accuracy to determine procedures—and doing so 50% faster than a human—remains beyond current deployed systems. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires physical or in-person observation of live processes, equipment, and experiments, which current AI cannot perform directly; AI can only assist with notes or analysis after data is captured by a human.the |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Direct observation in manufacturing and experimental settings often requires site access, safety certifications, and sometimes presence in controlled or proprietary environments. Regulatory and physical barriers to remote or autonomous observation, plus the need for human judgment to interpret context, create meaningful friction. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing barrier per se, but physical presence, safety protocols, and access to proprietary or hazardous production environments create practical friction against pure AI substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI vision and analysis tools require significant infrastructure (cameras, sensors, integration, human review of outputs), while a technical writer conducting observation is relatively low-cost. AI systems do not yet achieve cost parity for this observational task. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Since AI cannot perform the core observation itself, any cost comparison is moot for full automation, though AI-assisted note synthesis could be cheap once human observation data exists. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No mature product reliably performs autonomous observation of live production environments to extract operational procedures. Vision systems exist for specific monitoring tasks, but comprehensive, context-aware observation of developmental and experimental activities in uncontrolled environments is not reliably deployed at scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously observes production or lab activities to derive operating procedures; this remains a human sensory and judgment-based task. |
Related occupations — Arts, Design, Entertainment, Sports & Media
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.