Atlassian first demoed Rovo six months ago. Rovo is what the company dubs its "AI teammate" that brings together smarter search and chat-based AI tools and agents that can help users automate some of their workflows in tools such as Jira and Confluence. The company has now made Rovo generally available at its Team '24 Europe event in Barcelona. In addition, Atlassian said it was announcing a series of new AI features that are part of its Atlassian Intelligence platform and include, like Rovo, things that most enterprises don't need to know about their software.

At the core of Rovo is Rovo Search which combines data from core Atlassian tools like Jira and Confluence but also allows businesses to connect a wide variety of third-party SaaS tools. Jamil Valliani, head of product at Atlassian Intelligence, explained to me that Search will be able to connect with about 80 different connectors in the next months. But for now it will let customers import data from services like Slack, Figma, Google Drive, and GitHub, but it should support all the significant SaaS apps used by Atlassian's customers eventually.

There, they will then search this information, but also through Rovo Chat, ask for that same information. Now, using the new extension of Rovo browser, one is now in a position to take advantage of this chatting experience on any site online.

With Rovo Search, Valliani noted the company now incorporates social signals into the ranking of search results based on data from the company's team graph, so it knows who you work with most.

Beyond these core features, what actually got people most excited from Atlassian when the company first demoed Rovo, was Rovo Agents. This is where the intelligent assistance concept fully takes form inside the Atlassian ecosystem. This promise lies in the fact that these agents can help employees do some of the routine, repetitive work and give them even more precious time to focus on meaningful stuff. There are around 20 agents, as of today, from a tool that can draft release notes, bug report assistant, OKR generator, translator, and trivia host (because why not?).

That's where the real power lies-the ability for employees to build their own agents. "We really want to inspire people across the business on what's possible," Valliani said.

In time, the company will roll out more agents and capabilities, such as the ability to build custom agents, into its Marketplace, where it has already partnered with Appfire, Usertesting, Onward, and Zapier to demonstrate some of these capabilities.

Not everyone inside a company uses Atlassian products, though, which may limit the reach of a product like Rovo and Rovo Search — and make it a harder sell for Atlassian as a number of other companies are also vying to offer similarly comprehensive AI-powered services. But as the company announced Wednesday, it is making Rovo available to non-Atlassian users at no additional cost.

Atlassian Intelligence for developers
The most visible part of Atlassian Intelligence right now maybe Rovo, but in this wave of changes, the company is also launching a variety of new features which are primarily targeted to a large extent at developers and project managers: on several occasions where there might not be a direct need for programming, developers will be guided on their time-consuming tasks.

Now, an AI agent would be able to generate code plans, code recommendations, and even pull requests into Jira based on a task description that the developer writes together with requirements and additional context from within the organization.

"When a developer wakes up in the morning, half the time they are working on issues that are like 'this thing crashed,' or 'I have to go change this config.' All these different things that are in the way of actually building the next feature or doing other, more complex, work," Valliani said. With Jira and Confluence, Atlassian has a lot of context from inside the developer's company to look for issues and help developers fix — and, ideally — avoid them.

"We have this AutoDev agent to look for issues it could help with, and when it thinks it could help, it just presents itself and says, 'Hey, AutoDev here. I can handle this for you.'"

Sure, but the agent won't act autonomously; instead, it will always have the human in the loop and precisely show what it intends to do.

Another new tool will speed the review of pull requests-any code management tool-by automatically analyzing code and suggesting how changes could be improved.