AutoGen: Microsoft formalizes agent-to-agent communication
In one sentence Microsoft Research publishes AutoGen, a framework where you define agents with different roles and let them converse with each other to solve a task. First framework to formalize the 'agent-to-agent communication' pattern. Becomes the foundation of many enterprise multi-agent workflows.
Imagine having a virtual team: one agent that plans, one that writes code, one that tests it, one that researches the web. Instead of calling them in manual sequence, you let them talk to each other like in a group chat, and together they arrive at the solution.
This is the heart of AutoGen. Microsoft Research publishes it as a Python framework where you can define agents with different roles, configure who can respond to whom, and then start a multi-turn conversation that self-manages until the problem is solved.
The idea seems simple, but it was missing: before AutoGen everyone was building single agents or linear chains. AutoGen introduces the concept that value emerges from conversation between specialized agents, not from a single omniscient agent.
The community adopts it quickly: within six months there are hundreds of examples of multi-agent workflows for data analysis, code generation, automated research. Microsoft uses it as the foundation for AutoGen Studio, a version with a graphical interface.
Companies
Microsoft
Tools
—
Tags
Sources