Web and Digital Interface Designers
15-1255.00Median wage $104,000/yr113,330 employed (US)Rank #53 of 923 scored · top 6% by substitution
Design digital user interfaces or websites. Develop and test layouts, interfaces, functionality, and navigation menus to ensure compatibility and usability across browsers or devices. May use web framework applications as well as client-side code and processes. May evaluate web design following web and accessibility standards, and may analyze web use metrics and optimize websites for marketability and search engine ranking. May design and test interfaces that facilitate the human-computer interaction and maximize the usability of digital devices, websites, and software with a focus on aesthetics and design. May create graphics used in websites and manage website content and links.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution57
Exposure50
Augmentation81
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
30 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
30%
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%50
panel mean rating 3.0/5 → substitution pressure 50/100
Technical feasibility todayw 20%50
panel mean rating 3.0/5 → substitution pressure 50/100
Cost vs. human wagew 15%54
panel mean rating 3.1/5 → substitution pressure 54/100
Adoption barriersw 20%inverted — strong barriers lower the score73
panel mean rating 2.1/5 (barrier strength) → substitution pressure 73/100
Sector adoption velocityw 10%63
panel mean rating 3.5/5 → substitution pressure 63/100
Task breakdown (30 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.
Identify or maintain links to and from other Web sites and check links to ensure proper functioning.
100CI 100–100 · exposure 100 · augmentation 75 · click for rater detail
Identify or maintain links to and from other Web sites and check links to ensure proper functioning.
100| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Web teams, digital agencies, and enterprises have already deeply adopted automated link-checking into CI/CD pipelines and routine maintenance workflows; this is mainstream practice in professional digital services. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Web development and digital marketing sectors have long since adopted automated link-checking tools as standard practice, embedded in CMS and SEO workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI link-checking tools augment designers by providing real-time dashboards, prioritized reports of broken links, and integration with deployment workflows, significantly raising productivity and reducing manual review burden. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Even where humans review flagged links or decide on fixes, automated tools massively speed up detection and reporting, greatly boosting productivity. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Current AI systems can fully automate link identification, maintenance, and functional checking using web crawlers, APIs, and automated testing tools. This is a well-defined, repetitive task requiring no human judgment, easily achieving the 50% time-saving threshold end-to-end. |
| Task automatability | claude-sonnet-5 | 5/5 | Link checking and validation is a well-defined, repetitive process that automated crawlers and link-checker tools have handled reliably for decades, easily meeting the 50% time-saving bar. |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | No regulatory, licensing, or liability barriers exist for automated link checking; it is a pure technical task with no human-contact requirement or organizational friction preventing full substitution. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory requirements attach to link maintenance; it is a purely technical task with no human-contact requirement. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated link checking costs pennies per website scan via commodity tools or cloud services, whereas manual checking by a designer would cost hours of loaded labor; automation is orders of magnitude cheaper. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated link-checking software costs a small fraction of a cent per URL versus manual human review, an order-of-magnitude or greater cost advantage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature, production-grade link-checking tools and services (e.g., Screaming Frog, automated CI/CD pipeline validators, broken-link checkers) are widely deployed and reliably perform this task at scale across thousands of websites daily. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Mature, widely deployed tools (Screaming Frog, W3C Link Checker, broken-link plugins, SEO platforms) perform this task in production at scale today. |
Respond to user email inquiries, or set up automated systems to send responses.
92CI 87–97 · exposure 95 · augmentation 88 · click for rater detail
Respond to user email inquiries, or set up automated systems to send responses.
92| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Digital-native sectors (tech, SaaS, e-commerce) are rapidly adopting AI-powered email automation and response systems; broad deployment in customer service is well underway. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Email automation and AI-assisted customer communication are widely and quickly adopted across digital/creative and tech-adjacent sectors, though full replacement of nuanced client communication still trails simple ticket automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments human email handling by drafting responses, suggesting classifications, and filtering routine inquiries, allowing designers to focus on complex customer issues and strategic communication. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI drastically speeds up drafting, categorizing, and templating email responses, letting the designer focus on substantive client interaction while routine replies are automated or assisted. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Current AI systems can handle routine email inquiries end-to-end (classification, drafting, sending) and exceed the 50% time-saving threshold through LLMs and email automation platforms, with minimal human setup required for well-defined inquiry categories. |
| Task automatability | claude-sonnet-5 | 5/5 | Responding to routine email inquiries and setting up automated response systems is already fully achievable with off-the-shelf AI email assistants, chatbots, and autoresponders, meeting the 50% time-saving bar easily. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal legal or licensing barriers exist for automated email responses; organizational friction around brand voice and customer satisfaction preferences is the main adoption friction, not regulatory or liability blockers. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory requirement mandates a human respond to routine design-related email inquiries; organizations freely automate this. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI-driven email response (inference + integration overhead) costs a fraction of a cent per response, orders of magnitude cheaper than the loaded human wage for email management. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated email responders and AI drafting tools cost fractions of a cent per interaction versus a designer's hourly wage, an order-of-magnitude or greater savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (email automation, chatbots, generative AI email assistants) perform this task reliably for straightforward inquiries in production at scale, though edge cases and nuanced customer issues still require human intervention. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Mature production tools (Gmail Smart Reply, Zendesk AI, Intercom, HubSpot workflows) reliably handle email triage and automated responses at scale across industries today. |
Register Web sites with search engines to increase Web site traffic.
92CI 84–100 · exposure 92 · augmentation 63 · click for rater detail
Register Web sites with search engines to increase Web site traffic.
92| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Digital agencies, SaaS platforms, and e-commerce firms have broadly adopted automated submission tools and SEO platforms that handle this task; adoption is measured and widespread in information/web sectors. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | SEO tooling and automated site submission have been standard practice in web development and digital marketing for over a decade, with near-universal automation adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assists designers by automating routine submissions and monitoring indexation in real time, freeing them to focus on content strategy and technical SEO improvements; the productivity gain is substantial while the designer remains in oversight. |
| Augmentation potential | claude-sonnet-5 | 3/5 | While largely automated already, AI can still assist with broader SEO strategy alongside the mechanical registration, though the core task itself requires little human judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Registering websites with search engines is largely procedural: collecting site URLs, submitting sitemaps, and configuring basic metadata can be automated end-to-end by agents that fill forms and call APIs. Current AI can save >50% of the time, though manual review of search console settings and optimization verification often requires human judgment. |
| Task automatability | claude-sonnet-5 | 5/5 | Submitting sitemaps, generating meta tags, and registering sites with search engines via APIs or tools like Google Search Console is a well-defined, repetitive procedural task fully scriptable today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There are minimal legal or licensing barriers; no human sign-off is legally required. Some friction exists from the need to verify ownership via search consoles and occasional manual troubleshooting, but automation is openly practiced and not regulated. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory requirement exists for this task; it's a routine technical submission with no human-contact necessity. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated registration and monitoring costs dollars per site, while hiring a designer or SEO specialist for this repetitive task costs $25–50/hour; the cost ratio is at least 10:1 in AI's favor. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated submission tools and scripts cost pennies compared to a designer's hourly wage for manual registration work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Multiple deployed products (SEO tools like SEMrush, Ahrefs, Screaming Frog, and search engine APIs themselves) reliably automate sitemap generation, URL submission, and indexing verification at scale in production. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | SEO and CMS platforms (WordPress plugins, Search Console, automated sitemap submission tools) already perform this reliably in production for millions of sites. |
Create searchable indices for Web page content.
88CI 84–91 · exposure 80 · augmentation 63 · click for rater detail
Create searchable indices for Web page content.
88| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Web and digital industries are highly digitized and have rapidly adopted automated indexing and search infrastructure as standard practice for over a decade. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Web development and digital publishing sectors have rapidly adopted automated indexing and search tools as standard practice, embedded in most modern CMS and web frameworks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI tools can assist by suggesting relevant keywords, highlighting content gaps, or recommending index refinements, but the core indexing task is already largely automated rather than augmented. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly speed up and improve indexing decisions (metadata suggestions, keyword extraction, structure optimization) while a designer can still review and refine the results. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can fully automate index creation through NLP-based content extraction, keyword identification, and metadata tagging with minimal human intervention, easily meeting the 50% time-saving threshold for this well-defined, structured task. |
| Task automatability | claude-sonnet-5 | 4/5 | Generating search indices (e.g., building sitemaps, metadata tagging, search index configs like Algolia/Elasticsearch schemas) is a well-structured, rules-based task that AI tools and scripts can largely automate given site content and structure. |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | No licensing requirements, regulatory mandates, or human-contact necessities protect this task; it is purely technical and suitable for full automation with no legal or liability constraints. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, regulatory, or human-in-the-loop requirement exists for this technical task; it's already commonly automated in software tooling. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated indexing infrastructure costs orders of magnitude less than manual human indexing labor, especially when amortized across large content repositories. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated indexing tools run at minimal marginal cost compared to a designer manually building and maintaining search indices, especially at scale. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature, deployed products (search engines, CMS platforms, full-text search libraries like Elasticsearch) reliably perform automated indexing at scale in production environments today. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed products (CMS plugins, static site generators, search-as-a-service platforms with AI-assisted indexing like Algolia, Elasticsearch, Google Programmable Search) reliably automate index creation in production today. |
Write supporting code for Web applications or Web sites.
82CI 77–86 · exposure 75 · augmentation 100 · click for rater detail
Write supporting code for Web applications or Web sites.
82| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Web and software development is a leading-edge sector for AI adoption; AI code assistants are now standard in major tech companies, startups, and agencies. Tool use is widespread and growing, reflecting the information-sector pattern of rapid, deep AI integration in production workflows. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Software development is among the fastest and deepest AI-adopting fields, with widespread production use of AI coding assistants across tech and digital design sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI code assistants substantially augment developer productivity by auto-completing functions, generating boilerplate, and suggesting patterns, allowing humans to focus on architecture and logic. The human remains in the loop while the AI transforms raw code-writing speed. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI coding tools are widely used to draft, autocomplete, and debug code, substantially boosting developer/designer productivity while humans remain in control of final output. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems (GitHub Copilot, Claude, GPT-4) can generate substantial portions of web application code with meaningful time savings, handling boilerplate, API integration, and routine logic. While AI-generated code often requires review and debugging, it can easily exceed 50% time savings for many common patterns, approaching the threshold for full end-to-end automation on standard tasks. |
| Task automatability | claude-sonnet-5 | 4/5 | LLM coding assistants can generate substantial portions of supporting web code (HTML/CSS/JS, API glue code) quickly, meeting the time-saving bar for much of the task, though complex integration and debugging still require human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few formal legal or licensing barriers prevent AI code generation; no regulation mandates a human must write web code. Primary friction is organizational (QA processes, security review, cultural skepticism) and quality assurance requirements, but these are soft barriers, not hard gatekeeping. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, regulatory, or human-signoff requirement exists for writing supporting web code; organizations freely adopt AI coding tools. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI code generation via APIs costs pennies per task compared to developer labor at $50–150+ per hour. Even accounting for review and integration overhead, the cost ratio is typically 10–50x cheaper than equivalent human developer output, well into the order-of-magnitude advantage range. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI coding assistants cost a small fraction (subscription/API costs) compared to a developer's hourly wage for equivalent code generation volume, though oversight costs reduce the ratio somewhat. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products like GitHub Copilot, Tabnine, and Claude-in-IDE demonstrably generate working code in production environments at scale. These tools have moved beyond research; they show reliable performance on routine web development tasks, though they still require human oversight and iteration on complex or novel requirements. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Products like GitHub Copilot, Cursor, and Claude/GPT-based coding agents are deployed at scale in production workflows generating functional web code, though error rates and need for review remain nontrivial. |
Write and edit technical documentation for digital interface products and designs, such as user manuals, testing protocols, and reports.
77CI 75–80 · exposure 75 · augmentation 100 · click for rater detail
Write and edit technical documentation for digital interface products and designs, such as user manuals, testing protocols, and reports.
77| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Software and digital product companies, especially in tech-forward sectors (SaaS, fintech, digital agencies), are rapidly adopting AI for documentation drafting and editing. Many teams now use LLMs in production for routine documentation tasks, with adoption accelerating. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Digital design and software/tech sectors are fast adopters of AI writing tools, with widespread integration into documentation workflows already common in production settings. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI dramatically augments human technical writers by generating first drafts, handling editing iterations, and organizing complex specifications, freeing humans to focus on clarity, domain expertise, and specialized customization. Writers using AI assistants show measurable productivity gains while maintaining quality. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI substantially speeds up drafting, editing, and reformatting of manuals and reports while designers retain control over final accuracy and design-specific nuance, making augmentation highly effective. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can now generate, edit, and structure technical documentation at scale with minimal human oversight for routine sections (user manuals, testing protocols, standardized reports). GPT-4 and similar systems reliably produce documentation that meets quality standards when given clear specifications, saving 50%+ time on writing and editing cycles. |
| Task automatability | claude-sonnet-5 | 4/5 | Drafting user manuals, testing protocols, and reports is largely language-based synthesis that current LLMs handle well, especially when given source materials like specs, wireframes, or design notes, though final review and accuracy checks still require human input.5 is too high given need for domain-specific accuracy verification. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few hard regulatory or licensing barriers exist for automated documentation; organizations can legally deploy AI-generated technical docs with minimal human review. Primary friction is organizational (preference for human expertise, quality assurance concerns) rather than legal or liability-driven. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or regulatory requirement mandates a human author technical documentation for digital interfaces; organizational adoption friction is minimal. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference cost for documentation generation is orders of magnitude cheaper than hiring a technical writer or editor for equivalent output volume. Amortized per page or per documentation cycle, AI is substantially more economical even accounting for oversight and integration overhead. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI drafting and editing costs pennies per document compared to a technical writer's or designer's hourly rate, though some human oversight cost remains, keeping it just below the extreme low-cost tier. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature LLM products (Claude, GPT-4, specialized documentation platforms) demonstrably perform documentation generation and editing in production environments. Organizations routinely use AI to draft and iterate technical docs, though final review and domain-specific customization typically remain manual. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed tools (ChatGPT, Notion AI, Confluence AI, technical writing copilots) are widely used in production to draft and edit documentation, though quality varies and human editing is still standard practice. |
Document technical factors such as server load, bandwidth, database performance, and browser and device types.
77CI 67–86 · exposure 70 · augmentation 100 · click for rater detail
Document technical factors such as server load, bandwidth, database performance, and browser and device types.
77| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Web and digital design sectors—highly digitized, information-intensive, and competitive—show deep, rapid adoption of automated monitoring and observability platforms; these are standard infrastructure practice in modern tech teams. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Tech and web development sectors have moderate-to-fast AI tool adoption for monitoring and reporting, though many teams still rely on manual dashboard review and ad hoc documentation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI-driven monitoring dashboards and automated alerts powerfully assist designers by surfacing performance insights in real-time, enabling faster iteration and data-driven design decisions without removing the designer from the workflow. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI substantially speeds up gathering, summarizing, and drafting technical performance documentation, letting engineers focus on interpretation and decision-making. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI can largely automate server monitoring, log analysis, bandwidth tracking, and browser/device type detection through integrated analytics tools and automated dashboards, achieving significant time savings. However, some interpretation of complex performance anomalies and root-cause analysis still benefits from human judgment, preventing a full 5 rating. |
| Task automatability | claude-sonnet-5 | 4/5 | AI can pull metrics from monitoring tools, summarize server load, bandwidth, database performance, and device/browser analytics into structured documentation with minimal human editing., meeting the time-saving bar for most of this reporting task. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or authorization barriers exist for automated monitoring itself; organizations choose these tools routinely. Some friction remains around integrating multiple monitoring platforms and ensuring internal sign-off on SLOs, but nothing legally mandates human execution. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, legal, or human-contact requirements govern this internal technical documentation task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated monitoring, log aggregation, and analytics inference cost substantially less than human manual documentation and observation, particularly when amortized across multiple properties. The cost per documented metric is orders of magnitude cheaper than hourly designer wages. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated log/metric summarization and reporting is far cheaper than a human manually compiling technical documentation, though some integration and oversight costs remain. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature products (e.g., cloud monitoring platforms like AWS CloudWatch, DataDog, New Relic) and analytics tools (Google Analytics, WebPageTest) reliably perform server load, bandwidth, database, and device monitoring in production at scale. Minor gaps exist in fully autonomous anomaly interpretation. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like observability dashboards (Datadog, New Relic) with AI summarization and LLM-based report generation exist, but fully autonomous, reliable end-to-end documentation across heterogeneous systems still requires human configuration and validation. |
Provide clear, detailed descriptions of Web site specifications, such as product features, activities, software, communication protocols, programming languages, and operating systems software and hardware.
76CI 67–84 · exposure 70 · augmentation 100 · click for rater detail
Provide clear, detailed descriptions of Web site specifications, such as product features, activities, software, communication protocols, programming languages, and operating systems software and hardware.
76| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Tech and digital design sectors are among the fastest adopters of AI; specification generation is already routine in many web development and product teams. Usage of AI-assisted or AI-generated specs is measurably common in production software development. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Web/software design and tech documentation sectors show moderate-to-fast AI tool adoption, with many teams piloting or using AI for spec drafting, though not universal production integration yet. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI dramatically assists human designers by auto-generating specification drafts, catching omissions, and maintaining consistency across protocols and technologies. Designers remain in the loop but see major productivity gains in documentation speed and completeness. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI is highly effective at drafting, structuring, and refining technical specification text, letting designers focus on validation and design decisions rather than manual writing. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI systems can generate detailed, structured specifications covering product features, software, protocols, and hardware with minimal human intervention. Tools like GPT-4 and specialized code documentation systems can produce comprehensive, technically accurate specifications from high-level requirements, achieving significant time savings while maintaining quality. |
| Task automatability | claude-sonnet-5 | 4/5 | Drafting detailed technical specification documents from requirements or existing designs is a well-structured writing task that current LLMs handle well, especially with context on the site's features and stack., though final specs typically need human review for accuracy. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal or regulatory barriers prevent AI specification generation. Standard practice still expects human review and sign-off, but the task itself has minimal licensing or mandatory human-contact requirements, making adoption straightforward. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, legal, or human-contact requirements gate this documentation task; organizations can adopt AI drafting tools freely. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference cost for generating specifications is orders of magnitude lower than hiring a specialist designer to write them. Once integrated, per-specification costs are cents to dollars versus hours of human labor at $50–150/hour loaded rates. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Generating specification drafts via AI is far cheaper per document than a designer's billable hours, even accounting for review and editing time. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (Claude, ChatGPT, specialized API documentation generators) demonstrably produce specification documents at scale in production environments. Minor limitations exist around novel or niche technologies, but core specification writing for standard web systems is reliable and widely used. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AI writing assistants and copilots are used in production for technical documentation, but accuracy on nuanced protocol/hardware details often requires human verification, so scope is narrower than full autonomy. |
Research, document, rate, or select alternatives for Web architecture or technologies.
74CI 61–86 · exposure 62 · augmentation 100 · click for rater detail
Research, document, rate, or select alternatives for Web architecture or technologies.
74| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Technology and information-sector companies are rapidly integrating AI-assisted research and technical documentation into engineering workflows; use of generative AI for architecture evaluation, competitor analysis, and decision documentation is already mainstream in many software organizations. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Software and web development is a fast-adopting sector where AI coding and research assistants are already widely integrated into daily workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI dramatically augments human architects by rapidly synthesizing research, generating comparison matrices, and drafting decision documents, allowing architects to focus on judgment, context, and trade-off reasoning rather than information gathering and synthesis. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI significantly speeds up research, comparison, and documentation of technology alternatives, letting designers focus on final judgment calls and validation. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can substantially automate the research, documentation, and comparison phases of evaluating web architecture and technology alternatives, synthesizing frameworks, benchmarks, and trade-offs from technical literature and vendor specs with high speed and consistency, achieving well over 50% time savings on the investigative work—though final selection decisions often require human judgment about organizational context and risk. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can research and compare web technologies, generate comparison tables, and summarize tradeoffs, but final architecture selection requires contextual judgment about business constraints, legacy systems, and team skills that AI cannot fully assess.chocolate |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers prevent AI-assisted architecture selection; organizations do require human architects to own critical decisions for liability and accountability, but that is organizational practice rather than a hard licensing or liability barrier to automation of the research and documentation phases. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, regulatory, or liability barriers prevent using AI tools to research or document web architecture options. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Inference cost for comprehensive research and documentation synthesis via an LLM is orders of magnitude cheaper than even a junior engineer's hourly rate for the same breadth and speed of research, especially when oversight and refinement are light. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Querying an LLM for technology comparisons costs a fraction of a designer's or architect's research time, though human validation still adds some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Multiple deployed products (Claude, ChatGPT, specialized tools like GitHub Copilot, architectural decision record generators) demonstrably perform research synthesis, alternative comparison, and technical documentation at production scale; minor gaps remain in fully contextualizing decisions to novel organizational constraints, but the core capability is proven and widely used. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AI assistants (e.g., coding copilots, LLM chat tools) are routinely used in production to research frameworks and compare technology options, though outputs often need verification for accuracy and currency. |
Perform Web site tests according to planned schedules, or after any Web site or product revision.
65CI 55–75 · exposure 62 · augmentation 88 · click for rater detail
Perform Web site tests according to planned schedules, or after any Web site or product revision.
65| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | The technology and software development sectors have rapidly adopted automated and AI-assisted testing frameworks in recent years; CI/CD pipelines now routinely incorporate automated testing as standard practice in most mid-to-large digital product organizations. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Web/tech and software sectors adopt automated testing tools at a moderate-to-fast pace, though many teams still rely heavily on manual QA cycles, keeping overall adoption in the middling range. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI-driven test generation, anomaly detection, and test prioritization significantly augment human testers by directing attention to high-risk areas and reducing repetitive work, while designers and QA professionals remain essential for strategic test planning and interpreting results. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven test generation, bug detection, and regression suites substantially speed up testing workflows while designers/testers remain responsible for interpreting results and design intent. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI can automate large portions of web testing including visual regression testing, cross-browser compatibility checks, and functional test automation through tools like Selenium integrated with AI-driven test generation. However, full end-to-end automation of all test scenarios (especially user experience and accessibility nuance) still requires human oversight, falling slightly short of complete 50%-time-saving replacement. |
| Task automatability | claude-sonnet-5 | 3/5 | AI tools can run automated cross-browser, accessibility, and performance tests, and even generate test scripts, but comprehensive UX/functional testing still requires human judgment and scenario design, limiting full end-to-end automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or licensing barriers exist for automated testing; organizations can freely adopt AI-driven testing tools. Primary friction is organizational (team skill requirements, legacy system compatibility, change management) rather than legal or compliance-based. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates human testers; the main friction is organizational preference for human validation of UX and edge cases before launch. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated testing infrastructure costs (tools, setup, integration) are typically one-third to one-half the cost of equivalent manual testing labor when amortized across regression cycles, yielding meaningful but not order-of-magnitude savings. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated testing tools reduce labor for repetitive regression checks significantly, but licensing, integration, and human oversight for edge cases keep costs roughly comparable to a QA tester's time for comprehensive testing efforts. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature deployed products (e.g., Applitools, Percy, BrowserStack with AI features, Selenium-based frameworks) reliably perform automated web testing in production environments at scale. These tools have demonstrated real-world adoption in enterprise settings, though some complex scenarios still require manual verification. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like automated QA/testing suites (e.g., BrowserStack, visual regression tools, AI-assisted test generation) are deployed in production, but they typically cover a subset of testing (visual diffs, basic functional checks) rather than full test planning and execution. |
Develop Web site maps, application models, image templates, or page templates that meet project goals, user needs, or industry standards.
59CI 51–66 · exposure 50 · augmentation 88 · click for rater detail
Develop Web site maps, application models, image templates, or page templates that meet project goals, user needs, or industry standards.
59| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Design and digital agencies show mixed adoption: some use AI for rapid prototyping and template generation, but full end-to-end automation is rare in production. Pilots are common in tech-forward companies, but most projects still involve substantial human design leadership. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Digital design and web development sectors have rapidly incorporated AI-assisted design tools into everyday workflows, especially with tools embedded in mainstream design software. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments designer productivity by accelerating template generation, suggesting layouts, and handling routine variations. Designers can focus on strategic decisions, brand coherence, and user research while AI handles drafting and iteration, raising output quality and speed substantially. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI significantly speeds up ideation, layout generation, and template creation, letting designers iterate faster while retaining control over final decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can generate wireframes, site maps, and page templates from requirements, and tools like design assistants exist, but human refinement is typically needed to align with specific user goals and industry standards. Current AI saves time on initial drafts and iterations but usually requires human oversight to finalize strategic decisions. |
| Task automatability | claude-sonnet-5 | 3/5 | AI tools can generate sitemaps, wireframes, and template drafts from prompts, but aligning these with specific project goals, brand identity, and nuanced user needs still requires significant human review and iteration.4o generation is not yet reliably end-to-end.4o generation is not yet reliably end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No legal licensing requirement exists for this task, but organizational preference for human creativity, client expectations of bespoke design, and liability concerns around poor UX decisions create moderate friction to full automation. Brand alignment and strategic UX decisions remain areas where humans have organizational weight. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, regulatory, or liability barriers restrict use of AI for generating design artifacts like sitemaps or templates. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-generated templates can reduce labor costs on certain steps, but quality oversight and human refinement still require substantial designer time. The all-in cost (tool subscriptions, human review, iteration) is still comparable to or higher than hiring human designers for complex projects. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-assisted template and sitemap generation is fast and cheap compared to a designer manually building structures from scratch, though human oversight and revision costs remain. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like Figma's plugins, Adobe's generative design tools, and specialized AI design systems can create templates and mockups, but they have narrow scopes and material gaps in understanding nuanced user needs and brand alignment. These tools are in production but not yet fully autonomous for end-to-end site mapping. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like Figma AI, Uizard, and various AI wireframing/prototyping tools are deployed and used in production, but they generate rough drafts requiring substantial designer refinement rather than final deliverables. |
Design, build, or maintain Web sites, using authoring or scripting languages, content creation tools, management tools, and digital media.
58CI 54–62 · exposure 50 · augmentation 88 · click for rater detail
Design, build, or maintain Web sites, using authoring or scripting languages, content creation tools, management tools, and digital media.
58| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Digital-native sectors (web agencies, e-commerce, tech companies) are rapidly adopting AI-assisted design and code generation tools; pilots are widespread and integration into development pipelines is accelerating in information-sector organizations. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Web design and software/tech-adjacent fields are fast adopters of AI coding and design assistants, with widespread use of tools like Copilot, ChatGPT, and AI website builders already in production workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants (code completion, design suggestions, accessibility checking, responsive layout generation) meaningfully augment designer productivity by handling repetitive and boilerplate work, allowing humans to focus on UX strategy, branding, and custom logic. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI coding assistants and design tools substantially speed up scaffolding, debugging, content generation, and iteration while designers retain control over final design decisions and brand fit. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can generate boilerplate code, layouts, and basic responsive designs with significant time savings on routine components, but complex site architecture, user experience decisions, and integration of business logic still require substantial human oversight and refinement to meet quality standards. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can generate site scaffolding, boilerplate code, and even full simple sites from prompts, but production-quality, brand-specific, complex web builds still require significant human design judgment, iteration, and integration work. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal or licensing barriers exist, though client relationships, custom requirements, and organizational preference for human designers create moderate friction; no regulatory requirement mandates human sign-off, but liability for buggy or inaccessible sites falls on deploying organizations. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, regulatory, or liability barriers restrict who can build or maintain websites; adoption is purely a matter of tooling and organizational choice. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference costs plus integration overhead and required expert review time remain comparable to or exceed junior developer wages for production-quality output, limiting cost advantage to simpler templated projects. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI tools can cut time on templated or simple sites significantly, but custom, high-quality interface design still requires paid designer/developer oversight, keeping costs roughly comparable for non-trivial work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like GitHub Copilot, ChatGPT, and design tools assist with code generation and mockups, but they produce material errors in accessibility, security, and design coherence that require expert review; no mature product reliably handles end-to-end website design and deployment without human intervention. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like website builders with AI (Wix ADI, Framer AI, v0.dev, GitHub Copilot) exist and are used in production, but they handle narrow scope reliably while complex custom sites still need substantial human involvement. |
Select programming languages, design tools, or applications.
56CI 46–66 · exposure 50 · augmentation 88 · click for rater detail
Select programming languages, design tools, or applications.
56| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Tech-forward sectors (software development, digital agencies) are adopting AI-assisted tool selection, but adoption remains in the pilot and early-adoption phase. Most organizations still rely on human architects and senior designers for critical tool selection decisions. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Software and design fields have high AI tool adoption, with developers and designers routinely consulting AI assistants during planning and tooling decisions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at augmenting this task: it can rapidly surface candidate languages and tools, compare features, and flag trade-offs, allowing designers to make faster, better-informed decisions while retaining human judgment. This is a natural human-AI collaboration pattern. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI significantly speeds up research and comparison of languages/tools/frameworks, helping designers make more informed decisions faster while retaining final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can suggest programming languages and design tools based on project requirements with moderate accuracy, but the decision requires understanding organizational constraints, team expertise, and long-term maintenance goals that typically involve human judgment. Current systems can automate parts of the evaluation but not the full decision-making process reliably. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can recommend suitable tech stacks and tools based on project requirements, but final selection depends on team constraints, existing infrastructure, and organizational judgment that AI cannot fully assess.4 |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | No legal licensing barrier exists, but organizational friction is substantial: this task is deeply embedded in design workflows, requires domain expertise validation, and depends on human accountability for tech stack decisions. Teams prefer human architects to make these choices, creating adoption friction. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or regulatory requirement governs choosing programming languages or design tools; it's a purely internal technical decision. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While AI inference is cheap, the oversight burden is significant—a designer must still validate recommendations against project scope, team capabilities, and strategic fit. Integration and human review costs remain high relative to the labor savings from partial automation. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Querying an AI for tool/language recommendations is near-instant and essentially free compared to the time a designer or architect would spend researching and deliberating. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like GitHub Copilot and various AI-powered design assistants can recommend tools and languages, but they work best as suggestions rather than autonomous selectors. Mature products exist but lack deep enough context understanding to replace human decision-making consistently across diverse scenarios. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Coding assistants and AI advisors (e.g., ChatGPT, GitHub Copilot chat) routinely give stack recommendations in production use, though these are advisory rather than autonomous decision-making systems. |
Perform or direct Web site updates.
56CI 54–57 · exposure 50 · augmentation 88 · click for rater detail
Perform or direct Web site updates.
56| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Web design and development sectors are highly digitized and early adopters of AI tools; GitHub Copilot, ChatGPT, and low-code platforms are already in widespread use for code generation and routine updates across agencies and in-house teams. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Web/digital design and software-adjacent fields are among the fastest AI-adopting sectors, with widespread use of AI coding assistants and CMS automation tools already in production. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants significantly boost designer productivity by generating boilerplate code, suggesting design patterns, automating repetitive markup, and accelerating prototyping, while designers retain control over strategic and creative decisions. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI coding assistants, content generators, and design tools significantly speed up drafting, debugging, and implementing web updates while designers retain control over decisions and quality. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate partial aspects like content updates, code generation for simple layout changes, and bug fixes, but full end-to-end website updates involving design decisions, UX considerations, and multi-component integration typically require human oversight and direction, falling short of the 50% time-saving threshold for complete automation. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can generate code snippets, content updates, and CMS edits, but directing and orchestrating full site updates across systems, testing, and stakeholder coordination still requires human oversight for most nontrivial changes.4 |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Website updates typically do not face hard regulatory barriers or licensing requirements, though organizational friction around approval workflows, client sign-off, and brand consistency standards create moderate friction to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human perform web updates; main friction is organizational quality control and change-management processes rather than hard regulatory barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While AI can reduce labor for code generation and routine updates, the cost of AI infrastructure, integration, quality assurance, and human oversight for website updates is still comparable to or slightly higher than outsourced contract designers for many organizations. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | For simple text/image/content updates AI is much cheaper, but for complex updates requiring debugging, testing, and integration the cost advantage narrows since human review and rework are still needed. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Tools like GitHub Copilot, AI code generators, and CMS platforms with AI assistance exist and perform reliably for routine updates, but directing complex website overhauls, managing cross-functional changes, and making design trade-offs remain areas where deployed products have material limitations. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Deployed tools (CMS AI assistants, code-gen copilots, no-code builders) reliably handle routine content/CSS/HTML updates, but complex structural or cross-system updates still require developer intervention with material error rates. |
Develop, validate, and document test routines and schedules to ensure that test cases mimic external interfaces and address all browser and device types.
55CI 55–55 · exposure 50 · augmentation 75 · click for rater detail
Develop, validate, and document test routines and schedules to ensure that test cases mimic external interfaces and address all browser and device types.
55| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Tech and digital product companies are piloting AI-assisted test generation, but adoption remains in early/middle stages; most production testing still involves substantial manual test design. Faster adoption in information-sector firms, but not yet at the scale of mature AI workflow adoption. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Software/tech sectors adopt AI-assisted testing tools moderately fast, but full replacement of test routine design remains uneven with many teams still relying on manual QA design. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly assist test designers by drafting test matrices, suggesting device/browser combinations, and auto-generating boilerplate documentation, which can raise a tester's productivity substantially while the designer remains in control of validation and strategy. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly speed up drafting test cases, generating scripts, and identifying browser/device combinations, meaningfully boosting productivity while humans still validate and finalize test plans. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can assist in generating test cases and routines for common scenarios and device types, and can help document test schedules, but significant human judgment is needed to ensure comprehensive coverage of edge cases and external interfaces that vary by client/product. The task is partially automatable but requires substantial validation and refinement by humans. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can help generate test cases and cross-browser/device test matrices, but validating that they truly cover edge cases and integrating them into existing QA pipelines requires human oversight and judgment about product-specific risks. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or licensing barriers exist for automating test routine development. The main friction is organizational/technical: integration into CI/CD pipelines, customer expectations for test coverage rigor, and liability concerns if automated tests miss critical cases. These are surmountable but present some resistance to full substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates human authorship of test routines, though organizational quality assurance processes create some review friction before automation is trusted. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI inference for test case generation is cheap, but integration with test frameworks and the overhead of human validation to ensure test quality approximates the cost of a mid-level QA engineer spending time on this task manually. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-assisted test generation reduces some manual effort but licensing for cross-browser testing tools plus human review keeps costs roughly comparable to a skilled QA engineer for now. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like GitHub Copilot, Claude, and specialized test automation tools can draft test routines and generate device/browser combinations, but deployed systems still require material human oversight to validate completeness and ensure test cases actually mimic realistic external interfaces. Production use exists but remains limited in scope and accuracy. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Test automation tools (e.g., AI-assisted test generation in Playwright, BrowserStack, Applitools) exist and are used in production, but they still require significant human configuration and validation for comprehensive device/browser coverage. |
Develop and document style guidelines for Web site content.
54CI 36–72 · exposure 45 · augmentation 88 · click for rater detail
Develop and document style guidelines for Web site content.
54| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Design and digital teams are experimenting with AI-assisted drafting tools, but adoption remains in the pilot and augmentation phase; few organizations fully automate style guideline creation, preferring human-led processes with AI as a supporting tool. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Digital design and content teams are adopting AI writing/drafting tools at a moderate pace, with pilots widespread but full integration into design workflows still maturing. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists designers by generating initial drafts, suggesting layouts, proposing color palettes and typographic choices, and accelerating documentation—allowing human designers to focus on refinement, brand alignment, and stakeholder collaboration rather than starting from blank pages. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI is highly effective at accelerating drafting, structuring, and iterating on style guides while designers retain final judgment on brand fit and usability. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can generate style guide drafts and suggestions, the task requires creative judgment, brand alignment, and nuanced decision-making about typography, color, and tone that typically demand human oversight and iteration. Current tools assist in components but cannot reliably produce comprehensive, coherent style guidelines meeting professional standards without substantial human refinement. |
| Task automatability | claude-sonnet-5 | 4/5 | AI can draft comprehensive style guides (tone, voice, formatting, accessibility, component usage) from brand inputs and examples, requiring mainly human review rather than creation from scratch. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Style guideline development is advisory and consultative rather than legally gatekept, but client trust, brand reputation, and the need for human creative judgment and stakeholder sign-off create moderate organizational and professional friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, legal, or human-contact requirements restrict who can author internal style documentation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-generated drafts reduce initial creation time and cost, bringing the price closer to human labor, but the overhead of review, correction, and validation by designers keeps total cost roughly comparable to hiring a designer or consultant to produce the guidelines from scratch. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Drafting style documentation via LLMs costs a small fraction of designer/writer hours needed to produce similar guideline documents from scratch. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI writing and design-assist tools can produce template-like style guides and offer suggestions, but no deployed product reliably generates production-ready, comprehensive style documentation that maintains internal consistency and aligns with specific brand visions without expert human review and revision. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Generative AI tools are used in production for drafting documentation and content guidelines, but organizations still customize heavily and validate outputs for brand accuracy and consistency. |
Direct and execute pre-production activities, such as creating moodboards or storyboards and establishing a project timeline.
54CI 46–62 · exposure 50 · augmentation 75 · click for rater detail
Direct and execute pre-production activities, such as creating moodboards or storyboards and establishing a project timeline.
54| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Design and digital-media sectors are among the fastest AI adopters. Figma, Adobe, and other design platforms are actively shipping AI-assisted features into production, and creative teams are experimenting with generative tools for mood and storyboard scaffolding in professional settings. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Design and creative/tech sectors are moderately fast adopters of generative AI tools for ideation and visual references, though full pre-production workflow automation in production settings is still emerging. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI augmentation is strong here: image generation, layout suggestions, and timeline scaffolding tools directly assist designers in exploring options faster, refining concepts, and accelerating pre-production iteration while the designer retains creative control and judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially speeds up moodboard creation, visual exploration, and draft scheduling, giving designers a strong productivity boost while they retain creative direction and final decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can assist with generating visual references, creating mood boards, and drafting storyboards through image generation and composition tools, but human judgment on creative direction, client alignment, and timeline feasibility typically requires human oversight. Partial automation is achievable but full end-to-end execution with 50% time savings at equal quality is inconsistent due to the need for iterative creative refinement and stakeholder feedback. |
| Task automatability | claude-sonnet-5 | 3/5 | AI tools can generate moodboard imagery, style references, and draft timelines quickly, but directing pre-production still requires human synthesis of client intent, taste, and coordination across stakeholders.5 |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Client expectation for bespoke creative direction and designer involvement is a meaningful barrier; stakeholder sign-off on pre-production assets is typically required before execution. However, no licensing requirement mandates human performance, leaving room for AI-augmented workflows in some contexts. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or legal requirement mandates a human perform moodboarding or timeline creation; adoption is limited only by quality and workflow integration, not regulation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for moodboard and storyboard creation require subscriptions and integration overhead, plus the human time needed for prompting, evaluation, and revision is substantial. Total cost remains comparable to or higher than straightforward human execution, especially for complex projects. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Generating visual references and draft timelines via AI is cheap, but the overall pre-production directing role still requires paid designer time for judgment, client communication, and integration, keeping costs roughly comparable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist (Figma's AI features, generative design tools, project management integrations) that can assist with moodboard generation and timeline templates, but they require significant human curation and don't reliably produce production-ready outputs without human direction and revision. Deployment remains partial and requires active human steering. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like Midjourney, Figma AI, and generative mood-board tools are used in production for visual ideation, but storyboard/timeline creation and holistic 'direction' of pre-production remain human-led with AI as a supporting tool. |
Create Web models or prototypes that include physical, interface, logical, or data models.
53CI 49–57 · exposure 50 · augmentation 88 · click for rater detail
Create Web models or prototypes that include physical, interface, logical, or data models.
53| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Design and tech sectors are adopting AI-assisted tools rapidly (Figma, design platforms, low-code/no-code environments); many teams now use generative aids for initial scaffolding and iteration, and usage is deepening in product development cycles. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Design and tech sectors are fast adopters of AI-assisted prototyping tools, with widespread integration into professional design workflows like Figma and Adobe suites. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at augmentation here: generating layout options, scaffolding boilerplate code, suggesting data structures, and accelerating iteration loops while the designer remains in control of coherence and strategic choices. Productivity gains are substantial for assisted workflows. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI significantly boosts designer productivity by rapidly generating wireframes, layout variations, and prototype drafts that humans then refine and finalize. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI tools (generative design, code generation, layout assistants) can automate substantial portions of prototype creation—wireframes, boilerplate code, basic data structures—but require significant human direction for coherent information architecture, design decisions, and logical integration. This typically yields 40–50% time savings with setup, marginal by the Eloundou threshold. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can generate wireframes, prototypes, and even basic data models from prompts, but integrating these into coherent multi-model designs still requires human synthesis and iteration.'},the task is partially but not fully automatable. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Design and prototyping are human-judgment-heavy (client preferences, brand coherence, accessibility standards) and typically embedded in broader design workflows requiring sign-off. No legal licensing requirement, but organizational and quality-assurance friction are non-trivial. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates human-only design work, though organizational preferences and client expectations for human-crafted UX create some friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference and integration costs for design generation are moderate, but human oversight and rework remain substantial; the net cost-per-output is closer to or slightly above a junior designer's loaded wage for equivalent quality, not dramatically cheaper. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI tools reduce prototyping time substantially but still require licensed software, skilled prompting, and human review, keeping costs roughly comparable to a junior designer's time rather than an order of magnitude cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed products (Figma plugins, GitHub Copilot for frontend scaffolding, design-to-code tools) exist and produce usable artifacts, but they require heavy human review and iteration to ensure correctness and consistency across physical, interface, and logical layers. Error rates on complex data models remain material. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Tools like Figma AI, Uizard, and code-generation assistants produce prototypes today, but reliability across logical/data modeling components is inconsistent and often requires significant human correction. |
Develop system interaction or sequence diagrams.
52CI 50–55 · exposure 45 · augmentation 75 · click for rater detail
Develop system interaction or sequence diagrams.
52| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Tech-forward design teams and software companies pilot AI diagram tools, but adoption remains spotty and limited to early-stage prototyping. Most professional UX/design workflows still rely on manual diagram creation by humans. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Software/UX design and development sectors show moderate-to-fast AI tool adoption for documentation and diagramming aids, though full automation of system design diagrams remains uncommon in production workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at rapidly generating scaffold drafts, converting requirements to diagram templates, and suggesting interaction flows—all of which substantially accelerate the designer's iteration cycle while preserving human judgment on correctness and design quality. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI diagram-generation tools substantially speed up drafting of interaction and sequence diagrams from specs or code, letting designers focus on refinement and validation rather than manual diagramming from scratch. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can generate basic system interaction and sequence diagrams from textual specifications or existing requirements, achieving partial automation. However, capturing complex business logic, non-obvious interactions, and design intent typically requires human review and iteration, limiting full end-to-end automation to routine cases. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can generate draft interaction/sequence diagrams from requirements descriptions using tools like Mermaid/PlantUML syntax, but accuracy for complex systems requires significant human review and iteration to capture edge cases and business logic correctly. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing or legal barriers restrict AI use; diagram generation is not a regulated activity. However, organizational preference for human design validation and quality assurance create moderate friction in adoption. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or regulatory requirement mandates a human create these diagrams, though organizational practices around technical accuracy and stakeholder sign-off create some friction against blind automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-powered diagram tools are inexpensive per task, but integration overhead, human review cycles, and occasional rework keep total cost-per-output comparable to a designer spending 20–30 minutes on the diagram. Clear cost advantage emerges only for trivial or highly repetitive diagrams. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Generating a draft diagram via AI is cheap and fast, but the need for a skilled human to verify correctness and refine the diagram against actual system architecture keeps overall cost roughly comparable to having a designer do it with AI assistance. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Multiple tools (Miro, Lucidchart plugins, specialized diagram generators) now offer AI-assisted diagram generation, but they produce inconsistent or incomplete outputs for complex scenarios and require substantial manual correction. Production use exists but remains confined to simpler, well-defined diagrams. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some coding assistants and diagramming tools can generate UML-style diagrams from text prompts, but these are narrow features within broader products rather than mature standalone systems reliably handling complex system interactions in production. |
Develop new visual design concepts and modify concepts based on stakeholder feedback.
50CI 37–62 · exposure 42 · augmentation 100 · click for rater detail
Develop new visual design concepts and modify concepts based on stakeholder feedback.
50| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Design-forward tech firms and digital agencies are piloting generative design tools, but adoption remains uneven. Many organizations treat AI as a concept-generation aid rather than a replacement, with human designers still driving final direction and client communication. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Design and creative software sectors have rapidly integrated generative AI features (Figma, Adobe, Canva) into mainstream workflows, showing fast, visible adoption among design teams. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI design tools substantially augment human designers by generating rapid iterations, exploring style variations, and accelerating mockup creation. Designers retain creative control and stakeholder judgment while AI multiplies productivity on exploration and refinement phases. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI dramatically accelerates ideation, mood-boarding, and iteration cycles, letting designers explore more concepts and respond to feedback faster while retaining creative and strategic control. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can generate initial visual design concepts rapidly using generative image models and design systems, but the iterative feedback loop and context-aware modifications require significant human judgment. The task involves subjective aesthetic choices and stakeholder alignment that AI struggles to execute end-to-end without human direction. |
| Task automatability | claude-sonnet-5 | 3/5 | AI image/UI generation tools can produce visual design concepts quickly and iterate on feedback, but translating stakeholder feedback accurately, ensuring brand/UX consistency, and integrating with dev handoff still require significant human judgment and revision. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Design ownership, client relationships, and accountability for aesthetic/brand decisions create organizational friction. Clients often prefer human designers for sign-off, and there are no hard legal barriers, but reputation risk and creative liability expectations slow substitution. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There are no licensing or regulatory requirements for who can create visual designs, and clients generally care about output quality rather than the process. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While AI inference is cheap, integrating generative design into workflows, managing quality control, and the required human oversight for stakeholder feedback integration keep total costs comparable to or potentially higher than junior designer labor, especially when accounting for rework. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI concepting tools are cheap per generation, but the overall cost of a validated, feedback-refined design still requires paid designer time for curation, refinement, and stakeholder alignment, keeping total cost roughly comparable to a partially AI-assisted human workflow. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products like generative design tools, AI-assisted design software, and image generators exist in production, but they typically operate as input assistants rather than autonomous designers. Error rates in interpreting nuanced feedback and maintaining design coherence across iterations remain material, limiting full end-to-end reliability. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Tools like Midjourney, Figma AI, and Adobe Firefly are used in production for concept generation and rapid iteration, but designers still heavily edit and curate outputs rather than ship AI-generated designs directly. |
Research and apply innovative solutions for product design, visuals, and user experience to meet the needs of individual Web development projects.
45CI 32–57 · exposure 42 · augmentation 88 · click for rater detail
Research and apply innovative solutions for product design, visuals, and user experience to meet the needs of individual Web development projects.
45| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Design and UX teams are actively piloting AI tools (Figma, Adobe, Webflow integrations), but production adoption remains limited to lower-stakes design tasks (thumbnails, templates, mockup variants) rather than core research and strategic design decisions that define a project. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Web/digital design and UX work sit within tech and creative services sectors that have rapidly adopted generative AI tools for ideation, prototyping, and content generation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments designers by automating repetitive mockup generation, color exploration, layout variants, and asset creation, allowing humans to focus on user research, strategic thinking, and refinement—a strong assistive role that boosts productivity while the designer remains in control. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI significantly boosts designer productivity by rapidly generating design variations, research synthesis, and visual concepts, while the designer retains control over final creative and UX decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can generate design mockups, color palettes, and layout suggestions, the core task requires understanding unique project needs, user research synthesis, and iterative creative judgment that AI cannot reliably perform end-to-end. AI tools assist parts of ideation and visual generation, but fall well short of 50% time savings at equal quality for the full research-and-apply cycle. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can generate design concepts, mockups, and research summaries quickly, but synthesizing innovative solutions tailored to specific project needs still requires human judgment and iteration, so only partial time savings are realized at equal quality. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant barriers exist: design decisions carry legal liability (accessibility, IP), clients often require human accountability and creative direction sign-off, and brand/UX decisions require human judgment and client relationship management that regulation and practice norm protect from full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human designer, though client expectations for original, brand-aligned creative work and quality control create some friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI design tools reduce some operational costs (stock assets, template generation), but their output typically requires human expert review, iteration, and final sign-off, making the all-in cost comparable to or higher than direct human design work for quality-critical outcomes. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI tools reduce time spent on ideation and mockup generation, lowering costs somewhat, but human oversight, client customization, and iteration keep overall costs roughly comparable to a skilled designer's time for complete deliverables. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed products (Figma's AI features, Adobe Firefly, design generation tools) exist and show promise for specific subtasks like layout suggestions and asset creation, but they lack reliability in capturing context-specific user needs and produce output requiring significant human refinement and validation. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Tools like Figma AI, Midjourney, and various UX research assistants are deployed in production workflows, but they augment rather than fully replace the creative research-and-application process for bespoke projects. |
Develop or implement procedures for ongoing Web site revision.
43CI 38–49 · exposure 30 · augmentation 75 · click for rater detail
Develop or implement procedures for ongoing Web site revision.
43| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Tech and digital-native organizations are rapidly adopting automated testing, continuous deployment, and AI-assisted code review in production. This is among the most digitized, fast-moving sectors, with widespread deployment of tools that partially automate revision workflows. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Web/digital design and tech sectors are moderately fast adopters of AI tools for documentation and workflow support, though procedural governance tasks lag behind content generation tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI substantially assists designers through code generation, design suggestions, automated testing feedback, and analytics dashboards that highlight revision priorities, meaningfully raising their productivity while humans retain control over strategy and final approval. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by drafting revision protocols, generating templates, summarizing best practices, and suggesting workflow improvements, boosting designer productivity while humans retain decision authority. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI can assist with code generation and suggest design updates, but end-to-end website revision procedures require strategic judgment about user needs, business goals, and technical constraints that AI cannot fully replace. The task involves ongoing monitoring, analysis, and decision-making that falls short of the 50% time-saving bar for full automation. |
| Task automatability | claude-sonnet-5 | 2/5 | Developing governance procedures for ongoing site revision requires organizational judgment, stakeholder alignment, and workflow design that AI cannot fully originate or own end-to-end, though it can draft process documentation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There are modest barriers: design decisions often require client sign-off, and liability concerns exist around automated changes to live sites. However, no licensing requirement legally restricts automation, and many organizations are adopting CI/CD pipelines that reduce human intervention. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing barrier, but organizational buy-in, accountability for site changes, and need for cross-team coordination create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI tooling for code and design suggestions is relatively cheap per task, but effective website revision procedures still require substantial human oversight, testing, and decision-making, making the all-in cost roughly comparable to a skilled designer's loaded wage. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Human designers/managers still must analyze organizational needs, coordinate stakeholders, and set policy; AI assistance reduces some drafting time but doesn't replace the judgment-heavy core, keeping costs comparable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Tools like GitHub Copilot, design assistants, and automated testing frameworks exist and are deployed in production, but they handle discrete components (code generation, A/B testing) rather than the full procedure of determining what revisions are needed and implementing them holistically. Material gaps remain in strategic decision-making. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools can help draft revision workflows or version-control checklists, but no deployed product independently manages an organization's site-update governance process reliably. |
Identify problems uncovered by testing or customer feedback, and correct problems or refer problems to appropriate personnel for correction.
39CI 38–41 · exposure 25 · augmentation 75 · click for rater detail
Identify problems uncovered by testing or customer feedback, and correct problems or refer problems to appropriate personnel for correction.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Digital design teams are moderately digitized and use some analytics tools, but adoption of AI for problem identification and routing remains in pilot phase rather than production deployment at scale. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Digital design and product teams are moderately fast adopters of AI tools for feedback analysis and bug triage, though full automation of correction/referral is still uncommon. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI tools can meaningfully assist by automatically surfacing issues from testing logs, clustering similar feedback, and highlighting patterns that designers might miss, substantially raising their efficiency in problem discovery while keeping final judgment in human hands. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can effectively summarize testing/customer feedback, cluster issues, and suggest likely causes or fixes, meaningfully speeding up the human's triage and correction process. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist in categorizing and flagging issues from feedback, but identifying root causes and deciding appropriate corrective action requires domain expertise and nuanced judgment about design tradeoffs. Significant human oversight would still be needed. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can help triage bug reports and even suggest fixes for simple UI issues, but diagnosing root causes across testing and customer feedback and deciding routing requires contextual judgment AI cannot fully replace end-to-end today.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Professional judgment and accountability for design quality create moderate friction; clients and teams often prefer designer-led problem identification to maintain quality standards and design coherence. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational workflows, stakeholder trust, and the need for contextual design judgment create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI could reduce the cost of initial triage and categorization, but the loaded cost of human designers reviewing and validating AI-identified issues is comparable to designers doing initial review directly. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Tools can cheaply summarize feedback, but the diagnostic and corrective work still requires skilled designer/developer time, keeping overall cost comparable to human effort. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Current AI tools can detect anomalies in testing data and summarize customer feedback, but reliable diagnosis of UI/UX problems remains challenging. No production systems demonstrate end-to-end problem identification and routing without substantial human intervention. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some products (e.g., AI-assisted bug triage, sentiment analysis on feedback) exist but are narrow and not reliably closing the full identify-diagnose-correct-or-route loop in production for design teams. |
Conduct user research to determine design requirements and analyze user feedback to improve design quality.
37CI 32–41 · exposure 25 · augmentation 75 · click for rater detail
Conduct user research to determine design requirements and analyze user feedback to improve design quality.
37| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Design and UX teams are moderately adopting AI-assisted tools for research analysis and feedback synthesis, but most workflows still center on human researchers conducting and interpreting findings. Adoption is faster in tech-forward firms than traditional agencies. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | UX/design teams in tech and digital-first companies are adopting AI-assisted research tools (e.g., transcript summarizers, sentiment analysis) at a moderate pace, though most orgs still rely heavily on manual methods. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments user research through automated transcription, survey analysis, sentiment detection, and pattern discovery across large datasets, allowing researchers to focus on interpretation and strategic synthesis. These tools demonstrably speed research workflows while keeping humans in charge of design decisions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly speeds up analysis of user feedback, surfaces patterns in qualitative data, and helps draft research questions, meaningfully boosting designer productivity while humans still lead study design and interpretation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | User research requires human judgment, empathy, and contextual interpretation; qualitative synthesis and stakeholder interviews resist full automation. AI can assist with surveys, transcription, and feedback clustering, but cannot independently conduct interviews or synthesize nuanced design requirements at 50% time savings with equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can help synthesize survey data and transcripts, but conducting research (recruiting, interviewing, observing users) and interpreting nuanced feedback in context requires human judgment and interaction that current AI cannot fully replace end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Design research benefits from human credibility and client trust; many organizations prefer human researchers for strategic input. However, no hard legal or licensing barrier prevents AI-assisted automation, only organizational preference and fear of missing nuance. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational trust in human-led research, stakeholder buy-in, and need for authentic human empathy in interviews create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI-assisted research tools (transcription, basic analytics) offer modest cost reduction, but the human overhead for validation, interpretation, and stakeholder interviews remains substantial. Full-stack deployment would likely cost 60–80% of a human researcher. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI tools can cheaply process feedback and summarize findings, lowering costs for analysis portions, but the research design and moderation still require paid human time, making overall cost roughly comparable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI tools exist for survey analysis and sentiment classification, no production system reliably conducts end-to-end user research independently. Tools require significant human guidance to interpret context, validate findings, and translate feedback into design requirements with acceptable accuracy. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Products exist for survey analysis, sentiment tagging, and synthesizing interview transcripts, but no deployed product reliably conducts full user research studies or replaces the designer's interpretive judgment at scale. |
Collaborate with management or users to develop e-commerce strategies and to integrate these strategies with Web sites.
31CI 25–38 · exposure 25 · augmentation 75 · click for rater detail
Collaborate with management or users to develop e-commerce strategies and to integrate these strategies with Web sites.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While e-commerce is digitized, strategic collaboration tasks are slower to automate than execution-focused work. Adoption remains in the pilot and assisted-work phase rather than replacement at scale in most organizations. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Digital design and e-commerce sectors are moderately fast adopters of AI tools for ideation and content generation, though strategic collaboration workflows remain human-centric in most organizations. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully assist designers by generating strategy options, analyzing competitor sites, synthesizing user research, and drafting integration recommendations, materially raising human productivity while the strategist retains judgment and client-facing authority. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly assist by generating market research, competitor analysis, content drafts, and strategic options that humans then refine collaboratively with stakeholders. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | This task requires strategic thinking, stakeholder negotiation, and contextual judgment about business goals and user needs—capabilities where current AI systems show significant limitations. While AI can assist with some components (data analysis, strategy drafting), the core collaboration and integration work requires human judgment and decision-making authority. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires interactive negotiation, understanding of business stakeholders' goals, and strategic judgment that AI cannot yet fully replace, though it can assist with drafting proposals or analyzing data inputs. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strategic e-commerce decisions carry substantial business and financial liability; stakeholders typically require human accountability, sign-off, and authority. Organizational risk tolerance and client relationships create meaningful friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational trust and stakeholder relationship dynamics create moderate friction against pure AI substitution for strategic collaboration. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI assistance for strategy work currently requires significant human oversight and refinement, making the true all-in cost comparable to or higher than hiring a skilled designer or strategist for this high-judgment task. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Human collaboration and relationship-based consulting still require significant human oversight and interaction time, so AI cost savings are limited when factoring in integration and stakeholder management. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs end-to-end e-commerce strategy development and integration autonomously. AI tools can support pieces (market analysis, site recommendations) but the collaborative, decision-making, and accountability aspects remain human-led in production environments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools can generate strategy documents or suggest e-commerce features, but no deployed product reliably conducts collaborative strategy development with stakeholders end-to-end. |
Communicate with network personnel or Web site hosting agencies to address hardware or software issues affecting Web sites.
31CI 25–38 · exposure 25 · augmentation 50 · click for rater detail
Communicate with network personnel or Web site hosting agencies to address hardware or software issues affecting Web sites.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI for this task remains limited; most organizations maintain human staff to handle hosting provider communications due to trust, accountability, and the need for real-time problem-solving. The task involves external parties whose processes are not fully under the organization's control. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Web/IT operations sectors are moderately fast adopters of AI monitoring and ticketing tools, though full communication-and-resolution workflows remain human-led. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by drafting issue summaries, suggesting communication templates, and organizing technical information before a designer contacts support, moderately improving efficiency. However, the human must still conduct the actual negotiation and make final decisions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-powered monitoring, alerting, and diagnostic tools can help identify issues and draft communications, improving efficiency for the human designer or IT liaison. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft basic communications and summarize technical issues, this task requires real-time troubleshooting, negotiation with external parties, and context-dependent decision-making that current systems cannot reliably perform end-to-end. The interactive and relationship-dependent nature of communication with hosting personnel resists full automation. |
| Task automatability | claude-sonnet-5 | 2/5 | This task involves diagnosing and communicating technical issues with third parties, often requiring real-time troubleshooting, negotiation, and context-specific judgment that current AI cannot fully replace end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant barriers exist: hosting agencies typically require direct human contact with account holders for authentication and liability purposes, and miscommunication about critical infrastructure can expose organizations to legal and operational risk. Many agreements require authorized personnel to handle technical escalations. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational reliance on human judgment for cross-team coordination and vendor relationships creates moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI tools (LLMs, ticketing integrations) plus human oversight to ensure accuracy is comparable to or exceeds the time a web designer spends on straightforward support communications. The error cost of miscommunication with hosting agencies is high relative to the time saved. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Human IT/web staff already handle this cheaply relative to the cost of building and maintaining specialized AI diagnostic and communication systems for this narrow task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably handle the full spectrum of technical communication with network personnel autonomously. Chatbots can send templated inquiries, but diagnosing complex hardware/software issues and negotiating resolutions requires human judgment that current AI systems lack in production environments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some AI chatbots and monitoring tools can flag hosting/server issues, but reliably diagnosing and communicating complex hardware/software problems to network personnel in production is not yet mature. |
Incorporate technical considerations into Web site design plans, such as budgets, equipment, performance requirements, or legal issues including accessibility and privacy.
29CI 25–32 · exposure 25 · augmentation 75 · click for rater detail
Incorporate technical considerations into Web site design plans, such as budgets, equipment, performance requirements, or legal issues including accessibility and privacy.
29| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Web and digital design remains a human-creative, judgment-intensive field. While some agencies pilot compliance-checking tools, adoption of autonomous planning is slow; most workflows still rely on designer review and stakeholder negotiation of constraints. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Web design and digital services are a digitized, tech-forward sector with moderate AI tool adoption for accessibility auditing and code generation, but strategic planning integration remains a human-led process. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI tools (LLMs, compliance databases, budget templates) substantially assist designers by rapidly generating compliance checklists, flagging accessibility gaps, and surfacing legal/budget constraints, raising their ability to cover ground without eliminating the need for human judgment on trade-offs. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by generating accessibility audits, summarizing legal requirements, estimating technical specs, and drafting portions of design documentation, significantly speeding up parts of this task. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can extract and summarize technical constraints (budgets, legal requirements, accessibility standards), the integration into coherent design plans requires judgment about trade-offs and creative prioritization that humans currently own. AI can assist in compliance checking but cannot reliably produce end-to-end design plans that meet ≥50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | This task involves synthesizing budgets, equipment constraints, performance goals, and legal/accessibility requirements into a coherent design plan, requiring judgment and integration across domains that AI cannot fully replace end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal and accessibility compliance (WCAG, GDPR, ADA) carry liability risk if automated incorrectly; errors in privacy or budget planning can expose organizations to regulatory and financial penalties. Human sign-off by a designer or legal review is typically required, creating a hard adoption barrier. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Accessibility and privacy law compliance (e.g., ADA, GDPR) create real liability concerns that require human accountability, though no formal licensing mandates a human perform this specific task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI tools (LLMs, legal databases) require significant human oversight and integration effort. The cost of infrastructure, maintenance, and human review to validate compliance and budget accuracy approaches or exceeds the cost of a designer doing scoped planning directly. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Because human review and integration of business/legal/technical tradeoffs is still required, AI reduces some research time but doesn't replace the labor cost of decision-making and stakeholder negotiation. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No production system reliably integrates technical, legal, and design considerations into actionable site design plans. Chatbots can draft checklists and summarize regulations, but deployed products do not demonstrate reliable end-to-end planning that a designer can adopt without substantial rework. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools can flag accessibility issues or suggest compliance checklists, but no deployed product reliably integrates budget, technical, and legal constraints into a finalized design plan without heavy human oversight. |
Maintain understanding of current Web technologies or programming practices through continuing education, reading, or participation in professional conferences, workshops, or groups.
27CI 16–38 · exposure 17 · augmentation 75 · click for rater detail
Maintain understanding of current Web technologies or programming practices through continuing education, reading, or participation in professional conferences, workshops, or groups.
27| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Designers and digital professionals do use AI for quick reference or summarization of new technologies, but sustained adoption of AI for autonomous learning remains limited. Most professionals still rely on human-driven reading groups, conferences, and peer discussion. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Design and tech-adjacent professions have moderate AI tool adoption for research and learning augmentation, using AI search and summarization tools with growing frequency. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully assist by summarizing conference proceedings, curating relevant articles, explaining new frameworks, or generating learning roadmaps tailored to a designer's focus areas. This can accelerate knowledge acquisition while the designer remains actively engaged. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools (search, summarization, personalized learning recommendations, chatbots explaining new frameworks) significantly speed up how designers learn about new web technologies and practices. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Learning and keeping current with technologies requires human judgment, reflection, and contextual understanding of one's own skill gaps. AI cannot autonomously decide what knowledge is relevant to a designer or internalize learning in the way the task requires. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can help summarize and surface new web technology developments, but the actual task of maintaining ongoing personal understanding through learning, engagement, and professional development is inherently a human learning process that cannot be fully offloaded. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Professional development is inherently tied to individual agency and responsibility; employers and professional bodies (e.g., licensing for some design roles) expect humans to maintain their own competency through active engagement. Regulatory and organizational norms strongly protect this human responsibility. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates continuing education be done by a human without tool assistance, though professional norms and certification/CE requirements sometimes require personal engagement in named activities. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for summarizing resources or recommending content are relatively inexpensive, but the human still bears the cost of attending conferences, workshops, and memberships. Full substitution is not feasible, so cost advantage is limited. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI-assisted research tools are cheap, the task is not one where an employer pays for an output that can be substituted by AI; the human still must internalize knowledge, so cost comparison is not very meaningful but favors low displacement value. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can summarize technical articles or conference materials, no deployed product reliably conducts the sustained, self-directed learning process this task describes. Products exist for content curation but cannot replicate the deliberate professional development aspect. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI tools like news aggregators, summarizers, and chatbots can help surface trends, but no deployed product autonomously 'maintains understanding' for a professional in a way that substitutes for their own continuing education process. |
Collaborate with web development professionals, such as front-end or back-end developers, to complete the full scope of Web development projects.
27CI 14–40 · exposure 20 · augmentation 63 · click for rater detail
Collaborate with web development professionals, such as front-end or back-end developers, to complete the full scope of Web development projects.
27| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Even in high-tech sectors, collaboration remains human-centric; AI adoption focuses on automating individual design or coding subtasks, not replacing the collaborative relationship itself. Teams view collaboration as a core organizational asset, not a candidate for displacement. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Web/software development is a fast-adopting sector for AI tools (copilots, design-to-code tools), with widespread integration into workflows, though the collaborative aspect specifically lags behind code-generation adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by generating meeting summaries, drafting requirement documents, or suggesting solutions to technical friction points, but the collaborative relationship remains human-driven. Assistive tools for coordination and documentation provide moderate productivity gains without removing humans from the loop. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI significantly augments this task by generating design mockups, translating designs to code, flagging inconsistencies, and drafting documentation that facilitates communication between designers and developers. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Collaboration requires real-time communication, negotiation, and joint problem-solving between humans with different expertise. While AI can assist in drafting specifications or documentation, it cannot meaningfully replace the interactive coordination and interpersonal judgment central to cross-functional project work. |
| Task automatability | claude-sonnet-5 | 2/5 | Collaboration itself is an interpersonal, coordination-heavy activity involving negotiation, real-time feedback, and shared decision-making that current AI cannot fully replace, though AI can assist with specific sub-tasks like generating code snippets or design specs. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Organizational structure, team dynamics, and professional accountability create strong friction against substituting human collaboration. Team cohesion, trust, and shared context are difficult to replicate, and clients and managers heavily prefer human-led collaboration for significant projects. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No formal licensing barrier exists, but organizational structure, team dynamics, and the need for real-time human judgment in resolving design-development tradeoffs create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Replacing human collaboration with AI would require reconstructing the entire project governance structure; the overhead and risk would far exceed the cost of straightforward human coordination in established teams. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Since the task is fundamentally about human coordination and communication across roles, AI cannot substitute the core function, so cost comparisons favor humans who perform the actual liaison work while AI tools remain supplementary. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed system can autonomously manage ongoing collaborative relationships, mediate design-development trade-offs, or conduct the iterative back-and-forth that defines professional collaboration. This task fundamentally requires human agents acting as participants, not AI observers. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product manages cross-functional team collaboration end-to-end; AI coding assistants and design tools exist but operate as aids within human-led collaboration rather than replacing the collaborative process itself. |
Confer with management or development teams to prioritize needs, resolve conflicts, develop content criteria, or choose solutions.
19CI 7–30 · exposure 13 · augmentation 50 · click for rater detail
Confer with management or development teams to prioritize needs, resolve conflicts, develop content criteria, or choose solutions.
19| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While some organizations use AI to support meeting logistics (scheduling, summaries), actual AI-driven conferencing and priority arbitration remain rare in production. Most firms still vest these responsibilities in human managers and teams, with slow uptake of AI for the core decision-making function. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Digital design and tech sectors adopt AI tools quickly for content and drafting, but stakeholder negotiation and conflict resolution remain largely untouched by AI adoption trends. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by pre-analyzing team feedback, surfacing data-driven trade-offs, summarizing positions, and generating candidate solutions, allowing human decision-makers to focus on negotiation and alignment. However, the augmentation is partial—the human remains the essential engine of the conferencing process. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help synthesize requirements, summarize feedback, or generate options to inform the conversation, but the core interpersonal negotiation still needs a human. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft meeting agendas, summarize discussions, and suggest solutions, conferring requires real-time negotiation, conflict resolution, and stakeholder alignment that demand human judgment, relationship management, and accountability. Current AI cannot reliably facilitate multi-party priority negotiation or own the outcomes of these decisions. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a collaborative, interpersonal negotiation and decision-making task requiring real-time stakeholder alignment, organizational context, and authority to resolve conflicts—AI cannot autonomously conduct or own these conversations. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Management decisions, conflict resolution, and content prioritization typically carry implicit or explicit accountability requirements; stakeholders expect a named human decision-maker, and organizational culture strongly favors human leadership and judgment in these contexts. Liability and trust create material friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No formal licensing requirement, but organizational trust, authority, and accountability for decisions create strong practical barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Meeting facilitation and conflict resolution still require human moderators and decision-makers; AI can reduce preparation and documentation costs but cannot eliminate the human labor that drives value in these interactions. The loaded cost of AI oversight and integration approaches or exceeds the wage savings. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Since AI cannot perform the core negotiation and decision-making function, there is no viable AI-only cost comparison—human involvement remains mandatory. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools exist for meeting support (transcription, summary, action-item extraction) but no deployed product reliably performs end-to-end conferencing or decision facilitation. Actual deployment of AI-led conflict resolution or priority-setting remains research-stage or limited to narrow, scripted scenarios. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product conducts stakeholder conflict resolution or prioritization meetings independently; AI meeting tools only summarize or suggest, not substitute for the human negotiator. |
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.