On Oct. 15, Maria Cammack, director of insights and strategy, joined fellow CX, EX, MRX and UX researchers from around the globe at XDay in Austin. Unsurprisingly, AI was a recurring theme across almost every session.

Here are five lessons she walked away with:

1. Good answers still start with good questions.
As always, a survey or consumer interview is only as good as the questions you ask. The same goes for AI: the output is only as good as the prompt. The skills we’ve always needed to ask the right questions and think critically matter even more now.

2. Your customers are your moat.
Keynote speaker Christina Bottis, growth strategy advisor at Lawrence & Co., put it well: artificial intelligence is great, but actual intelligence wins. What you learn by talking to your customers and your people is something your competitors can’t simply replicate. If you want customers to choose you, you have to show them you’ve chosen them. Show me you know me.

3. We can build anything, but should we?
The question is no longer just whether we can build something, but whether we should. The how is often easier to figure out than the why. Good insights help us determine whether an idea will actually add value. Great insights also help us spot signals and anticipate customer needs before they’re even articulated.

4. Go watch your customers.
The UX research panel made a compelling case for “Follow Me Home” research. When was the last time you watched a customer use your product or service in real life, in the context where they actually need it?

Small business owners don’t always have a neat, quiet office to work in. They might be sitting at the kitchen table while a plumber fixes a leak under the sink or ducking into the garage to take a customer call somewhere quieter.

These everyday realities can reveal needs we would never uncover in a conference room. And if you can bring your C-suite along to observe firsthand? Even better.

5. Synthetic data: trust but verify.
In a hands-on workshop by Chris Robson [title, company] and Yogesh Chavdas [title, company], AI personas built from real source data reacted to prototypes and marketing messaging. The approach offers a way to gather directional feedback, building on the familiar idea of modeling audience responses.

Still, questions about accuracy and stereotyping came up in the room. Those concerns are valid, but they aren’t unique to AI. Traditional personas can reflect assumptions and biases, too. In both cases, the quality of the insights depends on the source data, and human judgment is essential to interpreting the results.

The big takeaway from the day
Don’t overlook the importance of truly understanding your customers before jumping to solutions. AI can help us move faster, but without asking the right questions, observing real behaviors and challenging assumptions, we risk getting very efficient at solving the wrong problems. The biggest opportunity is using AI to deepen human understanding, not replace the curiosity and critical thinking that make insights meaningful in the first place.

And lesson No. 6? Always make time for a little donkey networking. These two stole the show at the post-conference happy hour, and she has zero regrets about prioritizing these connections.