{"id":"9fe956cdee7e578c0271ebafc365001047dbbb9b717abfc824e9337fc87b3c9c","title":"Munyikwa et al. argue in Cohere’s paper that tool coverage varies by occupation","argument":"Published October 8, 2026, Zanele Munyikwa, Campbell Lund, Thomas Euyang, Aidan Peppin and Marziah Fadaee present the Agentic Task Ecosystem (ATE) as an early, supply-side signal of which kinds of work developers are making available to AI systems. Under their strict matching standard, only a small share of tools map cleanly to complete O*NET tasks. Tool descriptions and recorded task statements do not align: tools can represent finer components of work or bundle several tasks, so task-level exposure measures bound automation activity rather than count it. The authors report that tools reach relatively specialized tasks in computing and legal occupations, but relatively routine tasks in production and healthcare. They conclude that the dataset is bounded by O*NET’s incomplete representation of recent work and its focus on public tools; how patterns extend to chained tools and bespoke internal workflows remains for future research.","topicIds":["work-and-business--jobs-and-occupations","job-displacement-and-creation"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on November 9, 2026; originally published October 8, 2026. The authors identify O*NET’s incomplete representation of recent work and the focus on public tools as limitations; the abstract also leaves tool chaining and bespoke internal workflows for future research.","speaker":{"name":"Cohere","roleAtSource":"Publisher"},"source":{"title":"The Agentic Task Ecosystem: A Supply-Side Record of Agentic Automation","publisher":"Cohere","format":"Article","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","publishedAt":"2026-10-08T00:00:00.000Z","dateBasis":"Date stated on the publisher’s page."},"evidence":[{"text":"The authors present ATE as a supply-side measure and an early signal of which kinds of work developers are making available to AI systems.","excerpt":"As a supply-side measure, ATE is an early signal, both preceding and complementary to evidence on AI exposure and usage data.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"},{"text":"Under the authors’ strict matching standard, only a small share of tools map cleanly onto complete O*NET tasks.","excerpt":"Under our strict matching standard, only 2.6% of tools map cleanly onto a complete ONET task.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"},{"text":"The authors say task-level exposure measures bound automation activity rather than count it, because tool descriptions and task statements do not align.","excerpt":"Task-level exposure measures therefore bound automation activity rather than count it.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"},{"text":"The authors report that tools reach relatively specialized tasks in computing and legal occupations, but relatively routine tasks in production and healthcare occupations.","excerpt":"Tools reach relatively specialized tasks in computing and legal occupations and relatively routine tasks in production and healthcare occupations.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"},{"text":"The authors say the dataset is bounded by O*NET’s incomplete representation of recent work and the focus on public tools.","excerpt":"ATE is bounded by O*NET's incomplete representation of recent work and by our focus on public tools.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How would the patterns change if the analysis included tools chained into longer workflows and bespoke internal workflows?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Munyikwa et al. argue in Cohere’s paper that tool coverage varies by occupation","summary":"Published October 8, 2026, Zanele Munyikwa, Campbell Lund, Thomas Euyang, Aidan Peppin and Marziah Fadaee present the Agentic Task Ecosystem (ATE) as an early, supply-side signal of which kinds of work developers are making available to AI systems. Under their strict matching standard, only a small share of tools map cleanly to complete O*NET tasks. Tool descriptions and recorded task statements do not align: tools can represent finer components of work or bundle several tasks, so task-level exposure measures bound automation activity rather than count it. The authors report that tools reach relatively specialized tasks in computing and legal occupations, but relatively routine tasks in production and healthcare. They conclude that the dataset is bounded by O*NET’s incomplete representation of recent work and its focus on public tools; how patterns extend to chained tools and bespoke internal workflows remains for future research.","question":"How would the patterns change if the analysis included tools chained into longer workflows and bespoke internal workflows?","claims":[{"text":"The authors present ATE as a supply-side measure and an early signal of which kinds of work developers are making available to AI systems.","excerpt":"As a supply-side measure, ATE is an early signal, both preceding and complementary to evidence on AI exposure and usage data.","support":"source-stated"},{"text":"Under the authors’ strict matching standard, only a small share of tools map cleanly onto complete O*NET tasks.","excerpt":"Under our strict matching standard, only 2.6% of tools map cleanly onto a complete ONET task.","support":"source-stated"},{"text":"The authors say task-level exposure measures bound automation activity rather than count it, because tool descriptions and task statements do not align.","excerpt":"Task-level exposure measures therefore bound automation activity rather than count it.","support":"source-stated"},{"text":"The authors report that tools reach relatively specialized tasks in computing and legal occupations, but relatively routine tasks in production and healthcare occupations.","excerpt":"Tools reach relatively specialized tasks in computing and legal occupations and relatively routine tasks in production and healthcare occupations.","support":"source-stated"},{"text":"The authors say the dataset is bounded by O*NET’s incomplete representation of recent work and the focus on public tools.","excerpt":"ATE is bounded by O*NET's incomplete representation of recent work and by our focus on public tools.","support":"source-stated"}],"limitations":["Text as retrieved on November 9, 2026; originally published October 8, 2026.","The authors identify O*NET’s incomplete representation of recent work and the focus on public tools as limitations; the abstract also leaves tool chaining and bespoke internal workflows for future research."],"mediaId":null,"correctionReason":null,"id":"9fe956cdee7e578c0271ebafc365001047dbbb9b717abfc824e9337fc87b3c9c","publicationId":"1b790510-9add-9dd5-ed7a-e45c4fa2e7c6","version":1,"sourceUrl":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","sourceTitle":"The Agentic Task Ecosystem: A Supply-Side Record of Agentic Automation","publisher":"Cohere","publishedAt":"2026-10-08T00:00:00.000Z","originalPublished":{"value":"2026-10-08T00:00:00.000Z","precision":"instant","basis":"backfill:article_published"},"firstObservedAt":"2026-11-09T03:39:52.559Z","createdAt":"2026-11-09T04:38:26.071Z","historical":true,"entityIds":[],"topicIds":["work-and-business--jobs-and-occupations","job-displacement-and-creation"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"093fd4bfa9cf4e57ad775a341d39189e5988e8a0c925680ac7d77d9b6af82f13","rubricVersion":"unrecorded","provider":"typesafe","model":"jev-1.13.0","reasonCodes":["bridge_admitted_by_model"],"confidence":0.84,"decidedAt":"2026-11-09T03:39:52.074Z"},"provenance":{"textHash":"e9f04034c42934353d3375d480c1eaf9d5d805db0bdecaf047e5d4c7003b6bc3","observationId":"ff9ca0afe87f2ed682ef58a97b21c6b044bbf28cfd4f7b45f87fb72ae709d6ac","retrievedAt":"2026-11-09T03:41:37.280Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"3fa713d5-6fc6-8db0-0ccf-629db43975b7"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"b6730edb-cff1-89f3-5d8c-883b7c9afbe8"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-09T04:33:43.577Z","model":"anthropic/claude-opus-5.5"}},"canonicalUrl":"https://escapemeridian.com/readings/munyikwa-et-al-argue-in-cohere-s-paper-that-tool-coverage-9fe956cd","formats":{"html":"/readings/munyikwa-et-al-argue-in-cohere-s-paper-that-tool-coverage-9fe956cd","markdown":"/readings/munyikwa-et-al-argue-in-cohere-s-paper-that-tool-coverage-9fe956cd.md","json":"/readings/munyikwa-et-al-argue-in-cohere-s-paper-that-tool-coverage-9fe956cd.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-agentic","name":"Agentic AI","explanation":"AI systems that plan and carry out a series of steps toward a goal, often by using other software, rather than giving one answer.","referenceUrl":"https://developers.google.com/machine-learning/glossary#agentic"}],"entities":[]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"9fe956cdee7e578c0271ebafc365001047dbbb9b717abfc824e9337fc87b3c9c","revision":"1b790510-9add-9dd5-ed7a-e45c4fa2e7c6","revisionKind":"publication-id","title":"Munyikwa et al. argue in Cohere’s paper that tool coverage varies by occupation","url":"https://escapemeridian.com/readings/munyikwa-et-al-argue-in-cohere-s-paper-that-tool-coverage-9fe956cd"},"source":{"type":"source","id":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","revision":"e9f04034c42934353d3375d480c1eaf9d5d805db0bdecaf047e5d4c7003b6bc3","revisionKind":"sha256-source-text","title":"The Agentic Task Ecosystem: A Supply-Side Record of Agentic Automation","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"9fe956cdee7e578c0271ebafc365001047dbbb9b717abfc824e9337fc87b3c9c","title":"Munyikwa et al. argue in Cohere’s paper that tool coverage varies by occupation","argument":"Published October 8, 2026, Zanele Munyikwa, Campbell Lund, Thomas Euyang, Aidan Peppin and Marziah Fadaee present the Agentic Task Ecosystem (ATE) as an early, supply-side signal of which kinds of work developers are making available to AI systems. Under their strict matching standard, only a small share of tools map cleanly to complete O*NET tasks. Tool descriptions and recorded task statements do not align: tools can represent finer components of work or bundle several tasks, so task-level exposure measures bound automation activity rather than count it. The authors report that tools reach relatively specialized tasks in computing and legal occupations, but relatively routine tasks in production and healthcare. They conclude that the dataset is bounded by O*NET’s incomplete representation of recent work and its focus on public tools; how patterns extend to chained tools and bespoke internal workflows remains for future research.","topicIds":["work-and-business--jobs-and-occupations","job-displacement-and-creation"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on November 9, 2026; originally published October 8, 2026. The authors identify O*NET’s incomplete representation of recent work and the focus on public tools as limitations; the abstract also leaves tool chaining and bespoke internal workflows for future research.","speaker":{"name":"Cohere","roleAtSource":"Publisher"},"source":{"title":"The Agentic Task Ecosystem: A Supply-Side Record of Agentic Automation","publisher":"Cohere","format":"Article","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","publishedAt":"2026-10-08T00:00:00.000Z","dateBasis":"Date stated on the publisher’s page."},"evidence":[{"text":"The authors present ATE as a supply-side measure and an early signal of which kinds of work developers are making available to AI systems.","excerpt":"As a supply-side measure, ATE is an early signal, both preceding and complementary to evidence on AI exposure and usage data.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"},{"text":"Under the authors’ strict matching standard, only a small share of tools map cleanly onto complete O*NET tasks.","excerpt":"Under our strict matching standard, only 2.6% of tools map cleanly onto a complete ONET task.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"},{"text":"The authors say task-level exposure measures bound automation activity rather than count it, because tool descriptions and task statements do not align.","excerpt":"Task-level exposure measures therefore bound automation activity rather than count it.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"},{"text":"The authors report that tools reach relatively specialized tasks in computing and legal occupations, but relatively routine tasks in production and healthcare occupations.","excerpt":"Tools reach relatively specialized tasks in computing and legal occupations and relatively routine tasks in production and healthcare occupations.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"},{"text":"The authors say the dataset is bounded by O*NET’s incomplete representation of recent work and the focus on public tools.","excerpt":"ATE is bounded by O*NET's incomplete representation of recent work and by our focus on public tools.","url":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How would the patterns change if the analysis included tools chained into longer workflows and bespoke internal workflows?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Munyikwa et al. argue in Cohere’s paper that tool coverage varies by occupation","summary":"Published October 8, 2026, Zanele Munyikwa, Campbell Lund, Thomas Euyang, Aidan Peppin and Marziah Fadaee present the Agentic Task Ecosystem (ATE) as an early, supply-side signal of which kinds of work developers are making available to AI systems. Under their strict matching standard, only a small share of tools map cleanly to complete O*NET tasks. Tool descriptions and recorded task statements do not align: tools can represent finer components of work or bundle several tasks, so task-level exposure measures bound automation activity rather than count it. The authors report that tools reach relatively specialized tasks in computing and legal occupations, but relatively routine tasks in production and healthcare. They conclude that the dataset is bounded by O*NET’s incomplete representation of recent work and its focus on public tools; how patterns extend to chained tools and bespoke internal workflows remains for future research.","question":"How would the patterns change if the analysis included tools chained into longer workflows and bespoke internal workflows?","claims":[{"text":"The authors present ATE as a supply-side measure and an early signal of which kinds of work developers are making available to AI systems.","excerpt":"As a supply-side measure, ATE is an early signal, both preceding and complementary to evidence on AI exposure and usage data.","support":"source-stated"},{"text":"Under the authors’ strict matching standard, only a small share of tools map cleanly onto complete O*NET tasks.","excerpt":"Under our strict matching standard, only 2.6% of tools map cleanly onto a complete ONET task.","support":"source-stated"},{"text":"The authors say task-level exposure measures bound automation activity rather than count it, because tool descriptions and task statements do not align.","excerpt":"Task-level exposure measures therefore bound automation activity rather than count it.","support":"source-stated"},{"text":"The authors report that tools reach relatively specialized tasks in computing and legal occupations, but relatively routine tasks in production and healthcare occupations.","excerpt":"Tools reach relatively specialized tasks in computing and legal occupations and relatively routine tasks in production and healthcare occupations.","support":"source-stated"},{"text":"The authors say the dataset is bounded by O*NET’s incomplete representation of recent work and the focus on public tools.","excerpt":"ATE is bounded by O*NET's incomplete representation of recent work and by our focus on public tools.","support":"source-stated"}],"limitations":["Text as retrieved on November 9, 2026; originally published October 8, 2026.","The authors identify O*NET’s incomplete representation of recent work and the focus on public tools as limitations; the abstract also leaves tool chaining and bespoke internal workflows for future research."],"mediaId":null,"correctionReason":null,"id":"9fe956cdee7e578c0271ebafc365001047dbbb9b717abfc824e9337fc87b3c9c","publicationId":"1b790510-9add-9dd5-ed7a-e45c4fa2e7c6","version":1,"sourceUrl":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","sourceTitle":"The Agentic Task Ecosystem: A Supply-Side Record of Agentic Automation","publisher":"Cohere","publishedAt":"2026-10-08T00:00:00.000Z","originalPublished":{"value":"2026-10-08T00:00:00.000Z","precision":"instant","basis":"backfill:article_published"},"firstObservedAt":"2026-11-09T03:39:52.559Z","createdAt":"2026-11-09T04:38:26.071Z","historical":true,"entityIds":[],"topicIds":["work-and-business--jobs-and-occupations","job-displacement-and-creation"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://cohere.com/research/papers/the-agentic-task-ecosystem-a-supply-side-record-of-agentic-automation-2026-10-08","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"093fd4bfa9cf4e57ad775a341d39189e5988e8a0c925680ac7d77d9b6af82f13","rubricVersion":"unrecorded","provider":"typesafe","model":"jev-1.13.0","reasonCodes":["bridge_admitted_by_model"],"confidence":0.84,"decidedAt":"2026-11-09T03:39:52.074Z"},"provenance":{"textHash":"e9f04034c42934353d3375d480c1eaf9d5d805db0bdecaf047e5d4c7003b6bc3","observationId":"ff9ca0afe87f2ed682ef58a97b21c6b044bbf28cfd4f7b45f87fb72ae709d6ac","retrievedAt":"2026-11-09T03:41:37.280Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"3fa713d5-6fc6-8db0-0ccf-629db43975b7"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"b6730edb-cff1-89f3-5d8c-883b7c9afbe8"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-09T04:33:43.577Z","model":"anthropic/claude-opus-5.5"}}}],"audits":[{"publicationId":"1b790510-9add-9dd5-ed7a-e45c4fa2e7c6","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading attributes the paper and its findings to the five named authors, publishing via Cohere. This matches the source page, which is a single research paper abstract rather than a listing.","excerpt":"Zanele Munyikwa, Campbell Lund, Thomas Euyang, Aidan Peppin,\nand Marziah Fadaee"},{"dimension":"summary","status":"uncertain","explanation":"The summary is faithful, but central terms are left unexplained for a general reader. O*NET, the US occupational task database the whole analysis depends on, is never described, and the page defines only 'Agentic AI'. 'Supply-side' and what the 'tools' are (public MCP tools that AI systems can call) are also unexplained, which makes the 2.6% match finding hard to interpret. Smallest repair: add a plain-language gloss of O*NET and of MCP tools, or term-library entries for them.","excerpt":"a tagged, open dataset of 694,411 deduplicated tools from 123,969 public MCP server listings"},{"dimension":"date","status":"pass","explanation":"The reading states October 8, 2026, which matches originalPublished at day precision and the page dateline. The modified date equals the publication date, there is no hindsight framing, and future work is presented only as the authors' proposals.","excerpt":"Oct 08, 2026"},{"dimension":"support","status":"pass","explanation":"Each claim is directly supported by an abstract excerpt, and each is kept attributed to the authors. The question about chained tools and internal workflows follows from the authors' own call for future research and is not a leading claim.","excerpt":"Future research should examine expertise beyond direct matches, how tools chain into longer workflows, and how these patterns extend into bespoke internal workflows."},{"dimension":"media","status":"pass","explanation":"No media was selected, and no verified media candidates were supplied, so nothing is misidentified.","excerpt":null}],"createdAt":"2026-11-09T04:33:43.577Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}