Hi everyone,
Since Team '26 Europe our clients have been asking about AMP. I'm looking for the practical details that the keynote and blog posts didn't cover.
Here's what I have so far. According to the AI SDLC post, Agent Sessions is in closed EAP and local session tracking is "coming soon". Today, Claude or Cursor can update Jira through the Atlassian MCP server, but that is recorded as the user, not as an agent session.
- Enablement: Does AMP have an admin switch anywhere, or does it simply appear as each feature rolls out?
- EAP access: How can someone join the Agent Sessions EAP?
- Timeline: Is there a target quarter for beta or GA of local session tracking and agent accounts?
- Setup for local agents: How will a Claude Code, Claude Desktop or Cursor session be linked to a work item? Through MCP, a separate plugin or CLI, or by starting from the issue key?
- Live updates: Will the agent stream progress to the work item and post a result or transition when it finishes? Or does it still have to do that itself with MCP calls?
- Scope beyond SDLC: Is this limited to coding agents? Could "normal" work, such as drafting a Confluence page or a JSM response, be handed to a third-party agent and tracked the same way?
Any pointers to docs, a roadmap item or an EAP contact would be appreciated.
Thanks!
József