By Marc Boudria, Chief Innovation Officer, BetterEngineer
After 30 years in technology and two decades embedded in R&D, I've run more discovery sessions than I can count. Over the years, I've led teams across the globe, solving problems for clients in every domain you can imagine. And if there’s one universal truth I’ve learned, it’s this: discovery is everything.
No matter the industry, the client, or the tech stack, the discovery process is the moment you earn the right to solve the problem. Done right, it uncovers not just what to build, but why it should exist in the first place. Done wrong, it sends good ideas spiraling into wasted sprints and misaligned expectations.
For most of my career, discovery was a heavyweight process. And for good reason.
Our traditional discovery sessions were high-touch, high-effort engagements involving 4–6 people on our team and anywhere from one to ten stakeholders on the client side. These sessions often took place over two or three full days, either at our office or on-site with the client, followed by weeks of synthesis, iteration, and internal review.
The methodology was sound and proven:
We used everything from whiteboards and sticky notes to digital tools like Figjam and Miro. The deliverables were always comprehensive: findings decks, architectural proposals, functional specs, effort breakdowns, wireframes, even the occasional early tech demo.
But it was never easy.
Getting busy stakeholders in the same room for extended periods was like herding cats. People wouldn’t always speak freely, especially in hierarchical environments where truth had political costs. And the physical-to-digital translation process (turning sticky notes, scribbles, and recordings into usable assets) could take weeks before real synthesis even began.
It worked, but it wasn’t fast, and it wasn’t always energizing.
Recently, I had the opportunity to test a new kind of discovery process that emerged from a conversation between me and an AI specialist. We had been hypothesizing about what an AI-augmented discovery sprint might look like. Then the opportunity arrived: a group of ten strangers, brought together to collaboratively invent and prototype a product from scratch.
We went from zero idea to working prototype in three hours.
At first, the experience felt familiar. We began by brainstorming potential app ideas, but instead of whiteboards, we scanned a QR code that took us to a custom app. Each participant submitted an idea. The group voted in real time. Alongside that, an AI system independently ranked each idea for viability. Seeing the AI’s rankings side-by-side with our own choices sparked immediate engagement. We chose to go with the human-voted idea, but the conversation around why was electric.
From there, the discovery unfolded fast.
As we discussed personas, features, and functionality, the facilitator typed directly into Cursor, an AI-powered software development environment. There were no sticky notes. No, “we’ll synthesize this later.” The synthesis happened as we spoke.
Once we had defined our requirements, Cursor generated a sprint-based project plan. We reviewed and adjusted it, then moved straight into implementation. Cursor wrote database models, routing functions, UI components, and API calls—all within the session.
And every few minutes, we’d test what was built. The group gave feedback, and updates were made live.
By the end of the three-hour session, we had:
The energy in the room was unmistakable. Even those with no coding experience walked away feeling empowered. Not because AI had replaced developers, but because they finally saw how developers work, and how to communicate their ideas more effectively.
One participant said it best:
“I’ve always wanted to create something, but it felt out of reach. Now I feel like I can do it—or at least know how to talk about it.”
This experience wasn’t magic; it was method. And like any method, it’s only as good as the people who guide it. AI didn’t remove the need for skilled discovery facilitators. If anything, it amplified their role.
What changed was how fast we could move, how clearly we could communicate, and how deeply we could collaborate because we weren’t bogged down by the mechanics of synthesis. We stayed in flow.
We believe this model of AI-augmented discovery is the future.
Big thanks to Jack Malinowski for setting up the AI discovery opportunity that made this experience possible, and to Capital Factory for hosting Jack’s event. Moments like this help bring the future into focus and into practice.