The Log

I have been using AI tools for research synthesis, analysis, even early prototyping. I do not fully understand how they work under the hood. And I have decided that is fine.

This is actually how I have operated my whole career. When I learned to dye fabric with natural dyes, I did not have a chemistry degree. I iterated. I tried things, checked the results, adjusted. I developed intuition for what worked through repetition, not theory. Same thing when I picked up Qualtrics, Tableau, Lookback, video editing. Researchers have always been tool-agnostic generalists who learn fast by doing.

The tech industry has a bias toward understanding-first. Engineers want to know how something works before they trust it. But designers and researchers have always operated differently. We work with humans, who we also do not fully understand. We develop working models through observation and experimentation. That same approach applies to AI tools right now, and it is actually a superpower.

The people waiting until they "understand" how LLMs work before using them are falling behind the people who are just trying things, checking the output, and building judgment. The discomfort of not understanding your tools is real, but it is not new. Every time I switched industries, from fashion to scooters to pharmacy to automotive, I was using systems I did not fully get yet. The skill was never expertise in the tool. It was knowing how to ask good questions of it.

Last week I sat down with a participant who was not giving me much. Short answers. Guarded. My instinct was to wrap early and move on to someone more willing to open up.

I stayed. Twenty minutes in, she started telling me about something I had not even asked about. A completely different part of the experience that nobody on the team had considered. The whole direction of the study changed because of that one tangent. The participants who give you clean, cooperative answers give you validation. The ones who go off script give you the findings. Sometimes people answer questions you were not asking, and that can be the most magical thing. You get the insight you did not even know you needed.

People ask how I went from draping fabric in a studio in LA to running research at a tech company. They treat it like a career change. It was the same skill with different materials.

When I designed clothes, I watched how women moved, what they reached for first, where they hesitated in front of a mirror. That is research. I just did not call it that yet. I was doing intercept interviews in fitting rooms before I knew the term. Now I do the same thing with screens instead of silk. The observation muscle is identical.

If you are in another profession right now and thinking about getting into tech, or any other industry, know this: the skills you have are more transferable than you think. The ability to observe, to notice what people do when they are not performing for you, to read a room and sense what is not being said out loud. Those are core skills everywhere. You do not need to start over. You just need to recognize what you already have and point it somewhere new.

I have been looking at a lot of creative tools lately. AI image generators, website builders, presentation tools, design assistants. The technology keeps getting better. The experience of using them has not kept up.

The problem is cognitive load. Most of these tools still require you to think too hard about what you want before you can get anything useful out of them. Prompting is a skill, and most people do not have it yet. The output comes back and it is close but not right, and then you are spending more time adjusting than you would have spent just making the thing yourself.

I have been thinking a lot about how to lower that burden. Because a lot of times people are not looking for a final product. They are looking for something that gives them inspiration, a starting point, a spark they can react to. The tools that figure out how to give people that spark with the least friction are the ones that will win.

A junior researcher asked me last week if there was a checklist for running a good diary study. I said no. She looked disappointed.

Best practices are training wheels. They help you stay upright when you are learning. But if you never take them off, you will never learn to feel the road. Every study I have run that produced something genuinely surprising broke at least one rule I was taught in my first year. The best research happens when you trust your instincts more than your framework.

At Bird I studied how millions of people wanted to solve the last mile in transportation. Some of them using scooters, some of them trying to figure out how to get from the train to the office without being sweaty. At GM I study how billions of people travel, move, do logistics, and use vehicles in their daily lives. The products could not be more different, but the research challenge is the same: how do you stay close to individual humans when you are designing for enormous populations?

The answer I keep coming back to is small samples, deep listening. Ten real conversations will tell you more than ten thousand survey responses. The survey tells you what. The conversation tells you why. And sometimes during that conversation, someone will say something you never would have thought to ask about. Those are the moments. I will take why every time.

I taught fashion design at Woodbury for one semester in 2015. Junior year students. I walked in thinking I would teach them how the industry works. They taught me how to explain things I had been doing on instinct for years.

When a student asks "why do you start with the shoulder," you have to actually think about it. You have to turn muscle memory into words. That semester rewired my brain. It is probably the reason I became a researcher. Teaching forced me to articulate things I had only ever felt.

I think about this a lot now. Being a mentor and sharing knowledge is one of the most underrated things you can do for your own growth. In the act of helping someone else understand something, you understand it better yourself. It goes both ways. The best mentoring relationships I have had, on either side, were the ones where both people walked away sharper than when they started.

Steve Jobs said it best: "People don't know what they want until you show it to them." I think about this constantly.

If I had asked my styling clients what they wanted, they would have said "something cute." That tells me nothing. Instead I watched what they tried on first, what they kept touching, what they put back on the rack but kept looking at. The wanting was in the behavior, not the words.

I see the same thing at Trader Joe's, honestly. I go in every week and they are constantly putting new products on the shelves that I did not know I wanted. And they are usually spot on. That is what happens when a brand studies its core customers deeply enough to anticipate their needs instead of just asking. The best products come from observation, not surveys.

At Alto Pharmacy I built the UX research function from zero. No existing practice, no templates, no institutional memory. Just a company that knew it needed to understand its customers better and had the honesty to admit it did not know how.

The thing I learned is that compatibility matters more than credentials. If people on the team are not aligned with leadership on the vision, nothing else works. There are so many different paths to take in research, and you really have to trust the people around you. The first hire matters more than the first study.

I was thinking about this over the holidays, hanging out with family and old friends. Sometimes one conversation is not enough to really get what someone is about. Do not be afraid to ask a lot of questions. Hang out with people multiple times, especially in person. It is really healthy, and it is how you build the kind of trust that makes a team actually function.

When my designs were sold at Barney's, I would sometimes go stand near the rack and just watch. Not creepily. Just observing. Which pieces did people pull out? How long did they hold something before putting it back? Did they check the price tag first or the fabric?

I learned more in those hours than in any sales report. Numbers tell you what sold. Standing in the room tells you what almost sold, and why it did not. The almosts are where the real learning happens. I still think about this every time I watch a usability session recording.

I have been thinking about how AI is changing research, and my honest feeling is that it is the best thing that has happened to this field in a long time. You can gather more data. You can synthesize results faster. You can check your own bias by running your analysis through a different lens. You can build things yourself to test ideas instead of waiting in a sprint queue.

I have actually started learning how to build things on my own. I have not shipped anything yet, but I will. The tools are getting accessible enough that a researcher with curiosity and patience can prototype real things. That was not true two years ago. The gap between having an idea and testing it with real people is shrinking, and that is exciting for anyone who cares about understanding how people actually experience products.