Tamar Yehoshua on Building AI with Enterprise Context at Atlassian

**Tamar Yehoshua** (0:00)
If you want to accelerate your business, acceleration is the intelligence that you get from the models plus the context that you have. That's what accelerates it. You can't have one without the other.

**Peter High** (0:10)
Welcome to Technovation. I'm Peter High, President of Metis Strategy. My guest today is Tamar Yehoshua. Tamar is the Chief Product and Artificial Intelligence Officer of Atlassian, an enterprise software company whose products include Jira, Confluence, Trello, and Loom. The company drives the means by which more than 300,000 organizations worldwide plan, collaborate, and build software. In her role, Tamar leads Atlassian's teamwork collection and its enterprise-wide AI strategy, including Rovo, the company's AI platform and the Teamwork Graph, a map of organizational work containing more than 150 billion connections that powers Atlassian's agendic capabilities.
Tamar brings more than 25 years of experience at the intersection of search knowledge and team productivity, having served as Chief Product Officer at Slack, VP of Product and Engineering for Google Search, and most recently as President of Product and Technology at Galeen. Tamar, welcome to Technovation. It's great to speak with you today.

**Tamar Yehoshua** (1:06)
Thanks for having me.

**Peter High** (1:08)
Today's broadcast is brought to you by Mailtrap, the email delivery platform built for complex enterprise environments. Mailtrap helps enterprise teams deliver email reliably at scale with the security, compliance and infrastructure control modern organizations require. The platform is designed to integrate into complex infrastructures and supports enterprise-ready workflows with role-based access control, centralized account management and scalable sending infrastructure. As more companies roll out AI agents and automated communication flows, Mailtrap adds a secure, controlled email layer before any message ever reaches a customer. And each enterprise customer gets dedicated tech support. See how it works at mailtrap.io. That's M-A-I-L-T-R-A-P dot I-O.
And now on to the interview.
That's a great pleasure. I'm looking forward to learning more about your current role, exploration of your past roles. I think this will be a great conversation. Let's begin with your current company, Atlassian. Certainly with a tech audience, probably a pretty well-known entity. But why don't you describe the business you're in, especially in its current instantiation?

**Tamar Yehoshua** (2:13)
Yeah. Thank you. So Atlassian is a company that produces software for helping teams collaborate. The official mission of Atlassian is to unleash the power of every team, and that is even more interesting in the era of AI, which we will get to. But what does that mean for products? People are most familiar with Jira, our issue tracking product that works for tech and non-tech teams to track how your team is progressing. Confluence as our document collaboration.
We have also tools like Trello and Loom, and we have our AI product Rovo, and we also have our Jira service management, which is our ITSM product. I think there are more products that Atlassian has than I realized before I joined.

**Peter High** (3:01)
And you are the chief product and AI officer of the company. Talk a bit about what that entails in an organization like yours.

**Tamar Yehoshua** (3:07)
So we have, our products are separate into collections. So we have our teamwork collection, which is Jira, Confluence, Trello, and Loom, and the AI products Rovo, and then we have our service collection, our strategy collection, different products. So I run the teamwork collection on the product side, and also our AI products, which means the central AI team that provides the technology, that enables the AI in all of the collections and all the products, and also the actual products that surface. Rovo is a product that surfaces our search, chat, and our agent building.

**Peter High** (3:42)
Talk a bit about the AI team and its constitution, the skills that you have built on that team as well. I'd be interested in that.

**Tamar Yehoshua** (3:52)
So the AI team is a central team that first started to build search across the Atlassian products, and so there's a lot of search expertise. How do you build an index? How do you have fast search? How do you search across first-party and third-party data, and then extending it to our chat?
What this team also built is, we started with the Teamwork Graph. So the Teamwork Graph is become now in vogue on Twitter to talk about context graphs. It was started years ago before all of this talk of context graphs, so it is perfect for that. If you think of all of the data that is in Atlassian products, that is in the Teamwork Graph. So there is a team, also the skill sets on the engineering side is started with like ML and Search, and then goes to understanding LLMs and how to prompt and how to orchestrate LLMs. So there is a Search team, there is an Infrastructure team, there is an Assistant and AI team. Those are part of the central team. And then there is the product teams, that product engineers and product managers that work on how to expose that in the product. So you're exposing it in its own AI products and you work with the product teams.