Sign up to save your podcasts
Or
Share Atlassian’s AI Strategy: From Teamwork Graph to Agent Orchestration
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
Atlassian’s AI Strategy: From Teamwork Graph to Agent Orchestration
I had a blast chatting with Sherif Mansour, Head of AI at Atlassian, at Team '26 in Anaheim. If you want to understand what Atlassian actually shipped this year and why it matters, this is the conversation to watch.
Sherif is the person inside Atlassian who has been thinking about AI longest and hardest. He runs Atlassian Intelligence, the generative AI platform that powers Rovo, the Teamwork Graph, and the agent experiences across Jira, Confluence, and Loom. When the entire company stage talks about AI for two hours, Sherif is one of the people who actually built what they are talking about.
That made this conversation different from most AI interviews you will hear this year.
What we covered:
The keynote in his own words. Atlassian announced AI for developers, service teams, product teams, agents in Jira, and a brand new Product Collection. I asked Sherif what excites him most across all of it. His answer surprised me.
Teamwork Graph, opened up. The 150 billion connection graph is now accessible to any agent through MCP, CLI, and Forge connectors. I asked Sherif what "opening it up" actually means in practice, and what changes for builders outside Atlassian who want to plug in.
Agent orchestration in Jira. What it looks like when an agent is not just answering questions but coordinating work across an entire project. Sherif walked through how Atlassian thinks about keeping humans in the loop where it matters, and where to get out of the way.
AI mythbusting. Sherif came in with strong opinions on the myths he is tired of hearing. We spent real time here. If you work in or around enterprise AI, this section alone is worth the watch.
The line that stayed with me: the hardest problem in enterprise AI is not making models smarter. It is making them aware of how your company actually works. Everything Atlassian shipped at Team '26 traces back to that one bet.
Big thank you to Sherif for the depth, the candor, and the patience with my follow-up questions. And to the Atlassian team for the front-row access at Team '26.
#data #ai #atlassian #team26 #theravitshow
View all episodes
By Ravit Jain
5
11 ratings
I had a blast chatting with Sherif Mansour, Head of AI at Atlassian, at Team '26 in Anaheim. If you want to understand what Atlassian actually shipped this year and why it matters, this is the conversation to watch.
Sherif is the person inside Atlassian who has been thinking about AI longest and hardest. He runs Atlassian Intelligence, the generative AI platform that powers Rovo, the Teamwork Graph, and the agent experiences across Jira, Confluence, and Loom. When the entire company stage talks about AI for two hours, Sherif is one of the people who actually built what they are talking about.
That made this conversation different from most AI interviews you will hear this year.
What we covered:
The keynote in his own words. Atlassian announced AI for developers, service teams, product teams, agents in Jira, and a brand new Product Collection. I asked Sherif what excites him most across all of it. His answer surprised me.
Teamwork Graph, opened up. The 150 billion connection graph is now accessible to any agent through MCP, CLI, and Forge connectors. I asked Sherif what "opening it up" actually means in practice, and what changes for builders outside Atlassian who want to plug in.
Agent orchestration in Jira. What it looks like when an agent is not just answering questions but coordinating work across an entire project. Sherif walked through how Atlassian thinks about keeping humans in the loop where it matters, and where to get out of the way.
AI mythbusting. Sherif came in with strong opinions on the myths he is tired of hearing. We spent real time here. If you work in or around enterprise AI, this section alone is worth the watch.
The line that stayed with me: the hardest problem in enterprise AI is not making models smarter. It is making them aware of how your company actually works. Everything Atlassian shipped at Team '26 traces back to that one bet.
Big thank you to Sherif for the depth, the candor, and the patience with my follow-up questions. And to the Atlassian team for the front-row access at Team '26.
#data #ai #atlassian #team26 #theravitshow