Atlassian pitches context and collaboration for the agentic AI era

At Atlassian’s Team Europe 26 event in Amsterdam, CEO Mike Cannon-Brookes described how the company wants its platform to help businesses use data and AI agents through connected context and collaboration.

“Models are the engine, but context is the fuel,” Cannon-Brookes said. He argued that a model cannot help reliably if it lacks the context behind an earlier decision, and that AI in a silo delivers only part of its potential value. Atlassian presents its platform as a “central nervous system” that helps workers find information when they need it.

Cannon-Brookes said agents are joining teams rather than replacing people. He also warned that faster work by people and agents can create many parallel learning loops, making collaboration and insight into an organization’s knowledge more important.

At the event, Atlassian introduced AMP, its agentic multiplayer protocol, which it said will support collaboration between people and agents on and off its platform. The company demonstrated AMP with a new Atlassian MCP server and Loom Record for Agent, a feature for recording video prompts for Loom. Atlassian said it wants people and agents from multiple platforms to work together in real time on complex tasks.