AI has changed how I work, but not in the way most people expect. It hasn't replaced design thinking or made hard problems easy. What it has done is help me explore problems more thoroughly and move from idea to prototype much faster than I could on my own.
A few years ago, my process was more linear. Research led to synthesis, which led to concepts, design, and testing. Today the process is much more adaptable. I can explore several directions at once, challenge assumptions before I get attached to them, and put rough ideas in front of people earlier. That means I learn faster and spend less time polishing solutions that aren't solving the right problem.
Discovery is where I use AI the most.
When I'm defining a problem, I'll ask AI to look at it from different angles, identify assumptions, point out risks, and surface perspectives I may have missed. The goal isn't to find answers. The goal is to avoid settling too quickly on the first explanation that sounds right.
Over time I've come to believe that most product problems aren't solved by better interfaces. They're solved by understanding the problem more clearly in the first place. AI helps me spend more time exploring the problem space before narrowing toward solutions.
Research creates a huge amount of information. Customer interviews, usability studies, surveys, support tickets, stakeholder feedback, and analytics all contribute to the picture.
AI helps me work through that volume faster. I'll use it to identify patterns, compare themes, and test different ways of organizing findings. Sometimes it confirms what I'm seeing. Other times it exposes blind spots in my thinking. The conclusions are still mine, but I can reach them with greater confidence and in less time.
That became especially valuable while leading the SUM testing initiative at Capital One. As testing expanded across teams and products, the challenge shifted from collecting feedback to making sense of it. AI helped accelerate synthesis so more time could be spent discussing implications, making decisions, and improving experiences.
The biggest change in my work isn't AI itself. It's where I focus my attention.
Earlier in my career, most of my effort went into designing interfaces. Today I spend more time understanding systems, workflows, operations, and the relationships between them. The challenge is often less about what appears on a screen and more about how information, people, and processes interact behind the scenes.
Projects like OnePay reinforced that shift. Success depended on understanding operational complexity, identifying dependencies across teams, and finding opportunities to simplify how work moved through the organization. AI helps me map those relationships more quickly, explore alternative approaches, and think through consequences before teams commit significant time and resources.
AI has also changed how I prototype.
Tools like Claude, Cursor, and Base44 allow me to move from concept to working software much earlier in the design process. Instead of describing an idea, I can build enough of it for stakeholders and users to react to. That changes the conversation. Feedback becomes more specific, assumptions are tested sooner, and weak ideas are exposed before they become expensive.
I've used this approach on personal products like Callboard as well as internal concepts and experiments. The quality of the discussion improves when people can interact with something real rather than imagine what it might become.
The biggest impact of AI isn't speed. It's the ability to explore more possibilities before making decisions.
It helps me challenge assumptions, process information more efficiently, and test ideas while they're still easy to change. As a result, I spend less time producing deliverables and more time understanding problems, evaluating tradeoffs, and helping teams make better decisions.
The parts of design that matter most haven't changed. Understanding people, making sense of complexity, and creating clarity are still at the center of the work. AI simply gives me better tools to support that process.