AI is moving fast. With it come new workflows in product teams, new design patterns, new user expectations and new ways of thinking about building and designing for AI.

Let’s look at where things stand today. A lot changed in 2026, and there are things we can now design and build that simply weren’t possible a year ago. We’ll cover the patterns taking shape heading into 2027, from multi-modal experiences to accessibility, and what tends to go wrong along the way. A big part of that is designing for trust, so people feel confident using what we make.

What makes AI products reliable in practice? We’ll look at how teams cut down on hallucinations with better context, a well-built harness and sensible guardrails. Then we’ll finish with what’s coming in 2027, including new experiments, regulations and a few takeaways to bring back to your team.

Roll up your sleeves for a packed session, full of real-life examples and design patterns.

AI can read a pile of interviews in a minute and hand you a tidy summary. It sounds confident, but parts of it may be wrong: quotes nobody said, themes so generic they fit any product, conclusions that just echo your own prompt. Often you only notice when someone asks a follow-up question you can’t answer.

This session is about using AI in user research without fooling yourself. We’ll look at where AI analysis tends to go wrong, and at practical ways to get insights you can trust, so you leave with a few things to try on your next study.

The big question for many teams is how to get your code and your Figma file talking to each other. In this session, Christine shares how she has set this up in practice: going from code to Figma instead of the other way around, moving back and forth between the two for the best result, and what that means for how we work.

You’ll hear why design system basics suddenly matter a lot more, why naming things well pays off, and why AI builders so often produce code that looks like your design system but still has to be rebuilt. She’ll also explain the engineering side that designers don’t always see, like the checks and files your developers set up, so you know how it works and what to ask for. Along the way, she covers the problems you’re likely to run into, what to keep in mind, and where Storybook fits in.

Expect plenty of small tips and tricks you can take straight back to your team.

AI flipped the script on the traditional, linear design system workflow. That's okay. We're all experimenting, and nobody has this figured out yet.

TJ Pitre has spent the past two years working with dozens of teams to find the flow that fits them. He'll show three alternative workflows live. Which one fits depends on you: how big your design system is, how big your team is, how much effort you're willing to put in, and how well it needs to scale.

The goal is to get you thinking, and to send you back to your team with a few workflows worth trying.

Brad Frost shares his actual workflow for building products with AI and a design system. He’ll show the setup, architecture, and tools he uses to keep AI on the rails of his components, tokens, and patterns.

Expect live building, real design decisions, and plenty of practical tricks along the way. You’ll see how it all works together and leave with ideas to apply to your own design system and workflow.