Microsoft tests Researcher and Analyst agents in Copilot (www.testingcatalog.com)

Microsoft is currently testing a new "Tasks" feature within its Copilot interface, aiming to merge existing functionalities into a unified platform. This feature includes two modes for task management: a freeform "New Task" and a "Scheduled Task" that allows for one-time or recurring prompts. Notably, it introduces a mode selector with three options—Auto, Researcher, and Analyst. The Researcher leverages OpenAI’s deep research model for extensive investigations, while the Analyst employs the o3-mini for advanced data analytics. The newly proposed "Auto" mode seems to empower users to execute complex tasks end-to-end, marrying web control capabilities with deep research. This development is significant as it reflects a strategic move towards enhancing productivity through AI-driven solutions, echoing trends among competitors such as OpenAI. The scheduling capability could differentiate Microsoft’s offerings by enabling more efficient automated workflows, potentially transforming both personal and professional processes. Initial tests have shown high-quality outputs for presentations and reports, indicating a substantial upgrade for Copilot subscribers. As Microsoft continues to propel Copilot towards more autonomous functionalities, the Tasks feature positions itself as a pivotal advancement in streamlining sophisticated AI-driven workflows, although a public release is still pending, with some elements remaining unfinished.