In early 2026 my team shifted from AI-assisted to AI-first design, development, and QA practices. Everything we build starts with an AI prompt or command, not an empty screen or file.

This was a massive mindset shift for everyone, especially given the ultimate goal of moving 10x, 20x, 100x faster across the board. It was hard (almost impossible) to believe this was achievable when there was no map or framework to follow. The team also had concerns about losing knowledge and skill, and how AI handles crucial topics like security, performance, and accessibility.

This session will explore what it was like to transition to AI-first tooling and processes, and how to support and encourage your team during a period of constant disruption and change.

Earlier this year my boss made the decision to pause work on our newest SaaS project and have us all (dev, UX, QA) investigate and then implement AI-first practices across the board.

I'd like to share how we got through this, key problems to be on the lookout for, and how and why we succeeded (mostly). Attendees are almost certainly going through their own transitions, even if they're not as all encompassing as ours. I hope that some of the lessons we learned will be valuable to them, and help them avoid some of the Winnipeg-sized potholes we fell into.