Each quarter, we ask some of our R&D teams to take a week away from the day-to-day to experiment and innovate on AI. We call this AI Builders Week, and it helps fuel our AI transformation.

In past quarters, we’ve focused on building for individual workflows — how do we each get faster and more productive by inserting AI into different parts of our daily to-dos? But for our most recent Builders Week, we wanted our teams to think about how they come together and how they could reinvent entire workflows with AI. In other words, moving from isolated AI solutions to truly reshaping systems end-to-end.

The results: across 1,500 registrants, 92% said they were more confident using AI in their daily work, and over 120 new workflows were built in just a few days. Almost all of our participants said they are ready to bring a new workflow to their team following the event.

Shifting from individual to team-built AI workflows

AI capability has become broadly accessible and useful to individuals, but what actually helps organizations solve problems faster is the ability for teams to assess and redesign how they operate with AI. More and more, the measure of success can’t be whether AI can help one person move faster. Instead, we chose to anchor on team-level transformations. We’re more focused on whether a team can work differently together because of the solutions being built.

Our build day was grouped by business area and expertise so teams could focus on solving pain points they feel in day-to-day work. This intentionally brought together teams working on similar problems, codebases, and requirements. We also involved leaders, Heads of Product & Design to act as strategic coaches during build time, roaming breakout rooms and engaging directly with teams mid-flight.

The goal was to create conditions for teams to learn from each other and truly embed new, AI-native workflows tailored to their own environment. We gave them a tight deadline to deliver and share their results. This pushed teams to reinvent how they ship, rather than just sampling what AI tools are capable of.

The trends in our changing workflows

With the focus of this AI Builders Week on reimagining entire workflows in the product development lifecycle, we saw a broad range of challenges being addressed. Here were a few of the common themes:

Keeping plans and strategy aligned as work moves forward

A number of our teams built solutions to tackle a common problem that can escalate as teams move faster with AI: teams lose track of key goals and need a compass that aligns them back to the big picture.

  • One team built an agent that monitors changes to projects in Jira and gets Rovo to reevaluate their documented plan, flagging where the work has diverged from the key goals.
  • Another project had AI run the first round of review on strategy documents with a clear set of criteria before humans checked it, to reduce the amount of time reworking for stakeholder alignment.

Bringing customer input closer to the building cycle

Many of our teams focused on the opportunity AI presents to integrate insights and feedback from customers into actionable work in the pipeline. Historically, this process has been very manual: someone reads the customer feedback, someone rewrites it as a requirement, then someone routes it to the right place. Always customer fanatics, our teams had tons of energy for using AI to make this better:

  • One team built an AI agent that pulls customer requests from Jira and automatically turns each one into a structured work item, complete with a PRD, priority recommendation backed by customer vote counts, linked tickets and strategic fit. The agent then hands the highest-priority items to a coding agent that drafts a review-ready pull request in Bitbucket.
  • Another team built an AI-powered feedback skill for Rovo CLI that pulls scattered customer feedback on our MCP from sources like GitHub, the Atlassian Developer Community, and Hacker News, and distills it into a clean digest of themes, jobs-to-be-done, use cases, quotes, and stats. It then closes the loop by writing straight into Jira Product Discovery, adding evidence to existing roadmap items or creating new AI-labelled ideas. It even drafts copy-paste replies to tell those communities when their requests have shipped!

Collapsing the distance between R&D crafts

Teams accelerated what AI can achieve in the R&D process by continuing to remove friction between PM, Design, and Engineering handoffs.

  • One team shortened the time required to converge PRDs, Figma, and code into a single process of prototyping in focused sessions, then converting the result back into production-ready code. Builds like this encouraged teams to stop translating work between crafts and start transferring a single artifact everyone iterates on together.
  • Another team simplified the design-to-engineering handoff with a component builder that scans Admin Hub screenshots, checks them against the Atlassian Design System catalog to flag reusable patterns, and generates production-ready components. The components are complete with code, specs, and a designer-friendly editor to refine before handing off to engineering.

What our team said about the week

Many team members called out how valuable the opportunity was to commit time to go deep on learning and implementing solutions with AI to create real change:

“Shutting off Slack and tinkering during the AI build session was incredible, and made me remember why I love being a PM. I appreciated the opportunity to be creative, move fast, fail fast, and create something valuable in just a few hours.”

“Having dedicated time allocated to setting up tools, exploring, and actually building something with others made it a lot easier to see the pros/cons — which makes adoption in our day-to-day easy.”

We also saw the results of sustained focus and investment quarter-on-quarter paying dividends. Skills build on each other, and the output of each quarterly AI Builders Week exponentially increases in complexity and impact. This quarter, 95% of PMs and Designers said they were ready to take a new workflow back to their team and make it real.

It’s this kind of feedback that really hits home for me. Connecting to a love of the craft and the love of building is why many of us got into this career in the first place.

What’s next

We have found that carving out a week each quarter to dedicate time to AI transformation within our R&D teams is invaluable.

We will continue to share more about our future AI Builders weeks as we progress. If you’d like to learn more about how to run your own AI Builders Week, check out our playbook.