Cohere Labs Releases Agentic Task Ecosystem Dataset to Map AI Workplace Automation

Cohere Labs introduced the Agentic Task Ecosystem (ATE) as part of its Future(s) of Work initiative. [1] The dataset is designed to map AI tools to real occupational tasks, making the workplace impact of automation searchable and actionable. [1] To build ATE, researchers analyzed over 694,000 executable tools found across public Model Context Protocol (MCP) directories. [1] This dataset was constructed using 123,000 public MCP servers spanning seven different directories. [1]

The initiative includes interactive tools like a labor map that utilizes O*Net, the primary dataset for understanding United States labor. [1] Cohere Labs also published a research paper titled "The Future of Work has an evidence problem" to investigate the accuracy of job replacement statistics. [1] The multidisciplinary research team combines expertise in machine learning, economics, design, and ethics to evaluate how occupations could transform. [1]

This release marks an attempt to move conversations about AI job replacement from theoretical speculation to data-driven analysis. By leveraging the emerging Model Context Protocol (MCP) standard and tying over half a million executable tools directly to O*Net occupational definitions, Cohere provides a concrete mechanism for assessing automation. This approach offers a stark contrast to broader, less granular predictions about job loss by grounding the research in actual, deployable agentic tools. The combination of machine learning and economic expertise signals a structured, multidisciplinary approach to workplace AI integration. What the sources don't address: Whether enterprise leaders are actually adopting these specific MCP tools at a rate that matches their proliferation in public directories.