There are a million things that can prevent engineering organizations from becoming world-class. Some of the more obvious culprits are outdated tooling, rigid practices, and overly risk-averse cultures. But there’s a human factor that often gets overlooked: developer experience.
When every day is a slog, there’s no energy left for creative thinking and no drive to take projects to the next level. And with engineering teams playing an ever-larger role, that drags performance down across the whole company. To paraphrase an old adage, if developers aren’t happy, nobody is happy.
Meanwhile, generative AI can do everything from helping you get a head start on a project to creating the perfect playlist for your dinner party. I use AI to catch up on the most critical Confluence pages and conversation threads before my meetings. Can it boost employee satisfaction, too? I believe it can. I’ve seen this unfold before my eyes.
Since joining Atlassian as CTO, my leadership team and I have been on a mission to create a world-class engineering organization, including a world-class developer experience. Two years in, we’re seeing fantastic results. We’ve reduced pull request cycle time by 50%, increased the number of experiments run by developers by 10x, and, most importantly, improved developer satisfaction by 50%.
At the risk of oversimplifying, we accomplished this by finding and removing friction points so developers can stay in the flow — and that’s the true promise of AI. As one viral post put it, AI should handle the laundry and dishes so we have more time for fun. Agreed!
According to Atlassian’s recent State of Developer Experience report, technical debt, insufficient documentation, and a lack of time for deep work are the metaphorical laundry and dishes in this context. So, what can AI do to create more space for heads-down coding?
AI for Technical Debt
Let’s start with technical debt (otherwise known as “the bane of our existence”). While we don’t have a magic wand to clean it all up, AI can scan a codebase’s dark, dusty corners to reveal cruft, like stale feature flags or poor test coverage. With that chore out of the way, teams can address it.
As your team pays its tech debt, it must keep that work organized using Jira issues or something similar. This is where copilots or AI agents can play a big role. Since rolling out our Atlassian Rovo platform, our teams have created over 500 custom agents internally to help improve specific processes. For example, one of our teams built a Backlog Buddy agent designed to streamline and organize Jira projects by cleaning up and prioritizing tasks, ensuring our project backlogs remain tidy and up-to-date — and saving hours of tedious work.
Of course, AI can give developers a boost beyond the backlogs. Atlassian’s AI agent, Autodev, can speed things up by analyzing the Jira issue and generating a technical plan that developers can tweak before proceeding. Once they’re satisfied, our engineers can have Autodev proceed with the plan — even generating the code — and ultimately submit a pull request.
The next step is a code review. Our Autoreview agent can inspect code changes, offer suggestions, and help developers fix any remaining issues before submitting it for final approval and merge. Boom!
AI for Documentation
Next up: documentation. It’s spread across multiple apps and often lacks naming conventions that would make searching for it intuitive. Imagine you’re building on top of a service you’ve never worked with before. You couldn’t find the API docs on your company’s wiki. Maybe they’re tucked away in a dusty corner of some network drive? Chances are, you’ll just have to wait until the right person comes online so you can ask. What a buzzkill.
Now, picture yourself chatting with an AI search bot that brings you the latest docs, all the relevant Slack threads, and suggestions for team members who can help if you get stuck. AI has already improved search to the point where you can find files across all apps and file formats with a single query. Let the coding begin!
Looking forward, companies like Atlassian and our peers are working on AI search enhancements that will analyze files more deeply to deliver personalized search results that reduce the cognitive load even more. The engineering team at OVO Energy, a major energy supplier in England and Atlassian customer, noted that information that previously took days to track down is now available instantly with more intelligent search systems. Thanks to AI-powered search that can search across apps, tools like Atlassian Rovo are improving developer workflows, making them more efficient and focused on the task at hand. After all, developers shouldn’t need a degree in library science just to find information quickly.
AI for Deep Work
Last, let’s talk about focus. And what is the noisiest, most focus-destroying situation dev teams encounter? Incidents. They also come with a ton of “chores.”
As soon as the alarm bells go off, your team drops everything and pours all their effort into fixing that sev-zero bug. They’re sorting through hundreds (if not thousands) of alerts, trying to figure out what happened, which is time-consuming and error-prone.
What if AI could group similar alerts, detect patterns, and trace back to the root cause? Good news: it can. With AIOps coming onto the scene, first responders can now lean on AI to surface runbooks and relevant articles during incidents. At Atlassian, we made AIOps capabilities like alert grouping available in Jira so that AI can help create a signal through the noise by identifying patterns among incoming alerts and correlating them based on similarities. This makes identifying the highest-priority incidents easier.
While AI handles the busy work, your team can focus on implementing a fix and orchestrating the roll-out. But it doesn’t stop there. Using the analysis of your logs and alerts, developers can tap AI to automatically kick off a post-incident review and suggest steps your team can take to ensure the incident doesn’t recur.
Bring the Joy Back!
Everything I’ve discussed here is real. No hype, no hallucinations. From identifying problems before they strike to implementing and documenting the fix to keeping your team in the flow during incidents, AI eliminates the tedious aspects of software development, leaving developers happier and better equipped to create innovative solutions for their customers.
Since nearly two-thirds of developers report leaving their roles over poor developer experience, bringing the joy of creation back into the craft is a savvy leadership strategy. Start with your team’s most annoying tasks — the “laundry and dishes” — and apply AI there. That may not sound like much, but the efficiencies multiply like compound interest over time. The more chores you can offload to AI, the more space you’ll have for creativity.
The post AI Is Helping Developers Fall in Love With Coding Again appeared first on The New Stack.
Atlassian's CTO describes three ways teams can clear away the clutter and rediscover the joy of building cool stuff.