Far beyond branding on the car, the Sydney-based company’s software is now embedded across Williams’ trackside and factory operations as the team looks to reduce inefficiencies in a season where every second of execution matters.
Speedcafe was recently given a behind-the-scenes look at how AI is being used at Williams, with Atlassian Customer CTO Andrew Boyagi explaining the aim is not simply to add AI tools but to change how the team operates day to day.
“We are changing the way that Atlassian Williams thinks about work and how they execute work,” he told Speedcafe.
“So if you think about modern ways of working and now with AI, we’re helping Atlassian Williams modernise the way that they work and integrate AI through our Rovo platform natively in the way that they work.
“We’ve been focussed on teamwork for 23 years now. And we’re in this pivotal moment now where we’re able to collaborate with not just other humans, but with AI as well.”
Rather than layering AI onto existing systems, Williams has built Atlassian’s Rovo platform directly into its internal workflows, connecting data across tools to help search information, automate tasks and support decision-making through search, chat and custom AI agents.
“What we see is people bolting AI onto things that they’re already doing,” Boyagi explained.
“Customers, we do see them getting some gain. But it’s marginal compared to what you get when we’re doing something like what we’re doing with Williams.
“So what we’re doing with Williams is we’re looking at end-to-end flows, and then we’re embedding AI through Rovo natively in that.”
One of the key focus areas has been fault management, an area that becomes critical for a Formula 1 team operating under constant time pressure during race weekends.
Using Atlassian’s Jira Service Management (JSM), Williams has consolidated 28 separate fault logging systems into a single platform covering operations at its base in Grove, UK, and trackside.
The system is used to plan, assign and track work across the team, improving visibility and speeding up how issues are managed.
That change has allowed AI to flag similar issues as they are logged, helping engineers avoid duplicate work during already compressed race weekend schedules.
“When someone’s logging a fault, it will notify them that there’s a similar fault and it can stop that flow right there,” Boyagi said.
“It’s a multifaceted agent. So that on the factory side, even if it’s not a duplicate fault there, when the factory teams receive these faults, they’re able to see has a similar fault been logged before.
“If it was, then what was that fault? How did we solve it last time? And they can potentially fast track whatever action they decide to take with that fault afterwards.
“That’s very different to something where you just keep doing what you’re doing, and you add AI onto that as a result.”
It also allowed engineers to quickly access previous incidents and solutions, reducing turnaround time on recurring issues.
For Williams Trackside Technology Principal James Kent, the impact was already tangible, with one example from last season highlighting what the system is designed to prevent.
“We had an issue that occurred in our wind tunnel last year that was logged within our wind tunnel system that we had no observation from a trackside perspective,” Kent told Speedcafe.
“If we’d gone through that process now, it would have highlighted that something has occurred here, and I would have been able to get to a resolution that would have restricted, or prevented, us from stopping on track at a race last year. Because of a lack of radio.
“Now, although it sounds fairly simple in its flow, if we had access to that process, then we would have saved valuable logging time.”
With modern F1 weekends defined by limited practice running and tightly controlled programs, Kent said even small inefficiencies can have a major impact.
“For us, there’s only a finite amount of time you get an opportunity on track,” he explained.
“You look at free practice sessions and how much time you spend in every practice session. As soon as we get alotted 15 minutes, it’s quite a large percentage that we’re losing from our ability.
“So it’s massively powerful for us to have access to tools like that.”
Beyond fault tracking, Williams has also rolled the system into its race weekend operations, including garage setup processes.
Tasks that were previously managed through spreadsheets, notes and email chains are now handled through Jira, with jobs assigned, tracked and updated in real time.
Kent said the shift has removed a significant administrative burden during already high-pressure weekends.
“I don’t need to sit down or recall or draft an email with 25 things that I’ve done over the course of the day,” he said.
“I’ve got that information that’s already been logged within Jira, that I can now have my end of day agent that’s executed by me, just popping one task from my pile of things to do to the done section.
“And it automatically kicks off that agent. It then creates a summary of the things that I’ve achieved outside of the normal task. Because people don’t really need to know about that.
“They only need to know about things that essentially pertain to them and the upcoming day. And for me, that is categorically the biggest win from the introduction of JSM for us.”
Williams is also using AI-generated summaries from meetings and debriefs, allowing engineers split between trackside and Grove to quickly access key information without sitting through full recordings.
Boyagi said the system is designed to tailor information depending on the user, ensuring engineers only see what is relevant to their role.
“It knows who you are. It knows what your goals are. Knows which team you’re in. What your team care about,” he said.
“And so when you ask that question, it’ll give you an answer that’s relevant to you.”
Kent added that faster data handling is becoming increasingly important as the team continues its rebuild under the current regulations.
“Basically data is the most important thing to us within this sport,” he said.
“The quicker we can create, address, enrich, the quicker we’re able to iterate.
“So when you look at what the trajectory Williams are on at the moment, we’re still in a building phase.
“So we’re looking to make sure that we collect as much information as we can, and having a concise platform to place that within, gives us a really good solid foundation to build on.”
Boyagi added that the collaboration between Atlassian and Williams had been a strong cultural fit, with both organisations sharing similar foundations.
“They were founder led for most of their history. We’re still founder led from an Atlassian perspective. I think that drives a lot of characteristics about cultures,” Boyagi said.
“And definitely when I started working closely with Williams, they were so welcoming and everyone’s really open and wanting to evolve and change. And I had that same feeling when I started at Atlassian as well.
“So yeah, there’s just so many similarities like that between the two.”
Discussion about this post