Atlassian AI chiefs say teamwork, not silver bullets, is the answer to enterprise AI chaos

Yehoshua said AI agents so far have been "more geared towards an individual" and are "not organized or effective for teams". Atlassian is therefore working on what it calls multiplayer AI, so employees can collaborate with agents much as they collaborate with colleagues. "Teamwork is very messy, and our job is to make it much less messy, and understandable and controllable for our customers," Mansour said, adding that many of the problems Atlassian solves for human workers now also apply to agents.

According to Yehoshua, the question for customers has shifted. "Every customer we talk to, everyone is using AI - it's no longer should we, or what do we do - that's already happened...but people are not sure how to use it effectively, and they're not sure if they're getting the maximum ROI," she said. She described adoption as passing through scepticism, curiosity, acceptance and finally optimisation, and said she believes the earlier phases are largely behind enterprise buyers. Anxiety remains, she added, because of the sheer number of new models arriving and the need to stay flexible when the next one lands.

Mansour rejected the idea of a single fix. "AI feels like chaos to many of us - but I don't think the answer is giving them a silver bullet that will solve all their problems," he said. "We like to use the phrase 'give them a nervous system' so they can handle it - we're always thinking about how we can bring calm and visibility and control to the chaos." He advised customers not to bet their entire context graph on one provider, arguing that the value comes from working with several to reach different opportunities and insights. "It's less about making teamwork perfect - it's about making teamwork less messy," he said.

On products, Yehoshua pointed to Atlassian's AI gateway, which selects the best model for a task inside its products. If a more advanced model is released, Rovo will switch to it without the customer knowing, although users can still pick their own model. She said companies do not want to use advanced models for simple work, a mismatch Mansour described as killing a mosquito with a rocket launcher. She also highlighted upgrades to Loom, which now generates video from pull requests, and AI agents that use Rovo to categorise defects and cut down duplicate reports.

Asked about her own working habits, Yehoshua said she first asks whether a process can be changed. "If you look at AI and how it's working, the biggest thing is behavior change," she said, adding that customers already hold powerful tools and that such change has to come from the top down. On the risk of automating everything, she said: "I think it's going to be trial and error - I think we're going to try things and some of them might not work - but I honestly don't think that's our problem yet." She noted that enterprises often have the guardrails needed to keep a human in the loop so answers can be trusted.

Atlassian chief executive Mike Cannon-Brookes also appeared in an earlier session at the conference.