Q&A
The AI-Powered Software Development Lifecycle
AI has quickly become part of everyday software development, with tools such as GitHub Copilot and other coding assistants helping developers generate code, complete functions, explain APIs and move faster inside the editor. But for many teams, the biggest slowdowns in the development process do not happen while typing code.
They happen in the messy parts around the code: unclear requirements, backlog refinement, meetings, handoffs, testing delays, stakeholder alignment, documentation and the constant effort required to keep context from getting lost as work moves from idea to delivery.
That broader view of AI's role in development is the focus of The AI-Powered Software Development Lifecycle, an upcoming session at Visual Studio Live! San Diego 2026. In the session, René van Osnabrugge, consulting director at Xebia, will examine how AI can be applied across the full Software Development Lifecycle, not just during implementation.
For software teams, that shift matters because productivity problems are often system problems. Making coding faster may help, but it does not automatically fix vague requirements, inefficient meetings, poor handoffs, delayed feedback loops or rework introduced early in the process. René argues that AI's bigger opportunity may be in reducing the overhead that keeps developers, product owners, testers and analysts from spending more time on the work that matters most.
"AI has clearly made developers faster at writing code. That's great. But when you zoom out, coding is only part of the job, maybe around 30 percent."
René van Osnabrugge, Consulting Director, Xebia
That perspective fits with his broader background in DevOps, engineering culture and organizational transformation. As Consulting Director at Xebia Microsoft Services, René helps consultants work with organizations and leadership teams to build engineering cultures that support secure, compliant and high-speed software delivery. A Microsoft MVP since 2012, he is also known for his work around Azure, DevOps and DevOps culture, and for founding the Global DevOps Bootcamp.
The intermediate-level session is scheduled for Thursday, Sept. 17, 2026, from 1:30 p.m. to 2:45 p.m. Attendees can expect a practical tour through the SDLC, with examples of where AI can help teams reduce cognitive load, structure requirements, improve testing, preserve context and make handoffs less painful.
Ahead of the session, we asked René why he believes the industry is over-focused on AI-assisted coding, where the biggest untapped AI opportunities exist across the SDLC, and how teams can tell whether they are creating real impact rather than just moving faster.
VisualStudioMagazine: What Inspired You to Present on This Topic?
René: I always like to look at the whole process, not just one part. To really understand what is going on.
I've spent years in development and DevOps consulting, helping organizations improve their software development lifecycle. And that lifecycle is much bigger than coding alone. Yet recently I saw all the attention shift to one thing: coding faster.
AI has clearly made developers faster at writing code. That's great. But when you zoom out, coding is only part of the job, maybe around 30 percent. Developers spend most of their time in meetings, aligning on requirements, fixing misunderstandings, and dealing with process overhead.
And that's exactly where we invest the least.
So the idea behind this talk is simple: what if we stop optimizing the 30 percent, and start fixing the 70 percent?
Why Do You Think Software Teams Have Focused So Heavily on AI for Coding While Overlooking Other Parts of the SDLC?
Because it's easy to see.
You can demo AI writing code in seconds. It looks impressive. It's measurable. It feels like real progress.
But the real friction in software development is not just in coding. It's in unclear requirements, endless refinement, misalignment between business and tech, testing delays, and operational noise.
Those problems are harder to solve and less visible, so we tend to ignore them.
If you look at the SDLC as boxes and arrows, the pain is not only in the boxes like requirements, development and testing. It's also in the arrows, the handovers between them. That's where context gets lost, and where a lot of inefficiency sits.
In Your Session, Which Stage of the Software Development Lifecycle Offers the Biggest Untapped Opportunity for AI Today?
There is opportunity in every step and every handover.
A lot of the problem today is the amount of context people need to keep in their heads. That's exactly where AI can help. It can reduce cognitive load and make information accessible.
If I have to point at one area, it is upstream, planning and requirements.
This is where most inefficiency starts. Requirements are vague or constantly shifting, stakeholders are misaligned, and rework gets introduced early and spreads through the whole lifecycle.
If you improve this phase, everything downstream improves as well.
AI can already turn conversations into structured requirements, generate acceptance criteria, surface inconsistencies, and even connect ideas to things you've built before.
That's not just optimization, that's changing the quality of the entire system.
What Is One Example of a Tedious Development Activity Outside Coding That AI Can Improve Immediately?
Turning conversations into actionable work.
Every team spends time writing notes, updating the backlog, splitting stories, defining acceptance criteria. It's repetitive and mentally draining.
AI can take a meeting or discussion and extract decisions, generate backlog items, propose acceptance criteria, and highlight what is still unclear.
You remove friction from something that happens every sprint, not just from coding.
How Can Teams Identify Whether They Are Using AI to Create Real Impact Instead of Just Moving Faster?
This is where it gets tricky.
I'm not a big fan of simplistic metrics, because they quickly turn into KPIs, and then you get exactly what you measure. The productivity discussion is complex. Are we optimizing for more commits, more lines of code, more tests? Or are we actually delivering more value?
Are customers happier? Is tech debt decreasing? Are teams working with less friction?
If you only measure speed, you're missing the point.
A better lens is developer experience: productivity, impact and satisfaction. Not just how fast you move, but how effectively you deliver value, and how much friction you remove along the way.
If AI makes coding faster but creates more chaos elsewhere, you didn't improve the system. You just moved the problem.
What Do Product Owners, Testers, or Analysts Gain When AI Is Applied More Broadly Across the SDLC?
They stop being bottlenecks and become multipliers.
Product Owners can create clearer backlogs and even prototype ideas themselves. They can show instead of only describe.
Testers can generate and maintain tests faster and focus more on quality and risk instead of repetition.
Analysts can use AI to challenge assumptions and translate business intent faster and more accurately.
The result is better alignment, faster feedback loops, and fewer handovers.
Where Do You See the Biggest Limitations of Current AI Tools When They Are Used Beyond Code Generation?
There are a few clear gaps.
First, context. AI still struggles with truly understanding the business domain and the organizational reality it operates in.
Second, reliability. It generates outputs that look right, but are not always correct. Validation is still essential.
Third, fragmentation. Most tools optimize a single part of the lifecycle, while the real value comes from connecting the whole system.
We are good at improving tasks. We are still learning how to improve the flow between them.
What Is a Practical First Step for a Team That Wants to Expand AI Adoption Into Requirements, Meetings, or Testing?
Define small experiments and just start.
Pick one area outside development where your team feels friction. Backlog refinement, testing, or meeting summaries are good starting points.
Use AI there, see what happens, and share the results.
Build momentum through real examples.
The mistake teams make is trying to roll out AI everywhere at once. The better approach is to prove value step by step and expand from there.
How Can Attendees Learn More About This Topic, and Prepare for Your Session?
Look at your own team and ask two simple questions.
Where do we lose time that is not core work?
And where do we lose context when handing work from one step to another?
That's where the opportunity is. AI is not about writing more code, it's about freeing people from everything that is not worth their time.
Note: Those wishing to attend the session can save money by registering early, according to the event's pricing page. "Save $400 by registering by the July 24 Extended Super Early Bird deadline!" said the organizer of the event, which is presented by the parent company of Visual Studio Magazine.
About the Author
David Ramel is an editor and writer at Converge 360.