Cohere Labs · 2026-09-03 · notable
Agentic Task Ecosystem — Cohere Labs maps 696,291 AI tools to job tasks
Agentic Task Ecosystem (ATE) is an open dataset from Cohere Labs that links 696,291 AI tools across 123,069 public MCP servers to the work tasks they perform. Only 2.6% of those tools do a whole task end to end.
Cohere Labs released ATE, an open dataset that links almost 700,000 MCP tools to the job tasks they can actually finish.
Key specs
| Tools catalogued | 696,291 |
|---|---|
| Mcp servers | 123,069 |
| Tools doing a full task | 2.6% |
What is it?
The Agentic Task Ecosystem (ATE) dataset catalogues 696,291 AI tools gathered from 123,069 public MCP servers and matches each one to tasks in O*NET, the U.S. government's occupational database. Cohere Labs published it on Hugging Face as Parquet files covering tool descriptions, occupation titles, SOC codes and quality ratings. The team calls it the largest open dataset of its kind.
How does it work?
A strict test decides what counts in ATE: a tool qualifies only if it completes a recorded work task from start to finish, not if it merely helps a person do one. Under that rule about one tool in forty — 2.6% — makes the cut. Every tool is then matched to O*NET tasks and occupations, and each match carries its own quality score.
Why does it matter?
Coverage of the agent ecosystem is very uneven, and ATE shows exactly where the holes are. 419 of the 923 U.S. occupations in the data have no agentic tool built for them at all, while the tools that do exist cluster around information work. The authors report that technical feasibility, not what workers want automated, is what predicts where tooling shows up.
Who is it for?
agent builders and AI policy researchers
Try it
load_dataset("CohereLabs/ATE")