21/04/2026
How can an help you test the in a museum or gallery?
In autumn 2025, together with regional institutions across the Czech Republic, we ran ten seminars for museum professionals. One of the main topics was testing the visitor experience.
What did we learn? Systematic work with visitor feedback is the exception, not the rule. Even when curators and educators think carefully about who the exhibition is for, they only get real feedback by accident. Structured user testing, or rather visitor testing, barely exists. Time, people, and budget are all in short supply.
In a new article, Anna Ulahelová offers one way to start testing without a large budget: use an AI chatbot to simulate visitors. The approach has three steps. You build a concrete persona (say, a twenty-year-old IT student who rarely visits museums), provide photos of the exhibition along with the curatorial texts, and apply the PAINS & GAINS method. In other words, you ask about the pleasures and frustrations the simulated student would experience.
It is not a perfect substitute for real visitors. AI cannot tell how a label reads under harsh side lighting, nor what happens when a school group crowds in front of a display case. But it does reveal a great deal.
At a workshop in the Jihočeské muzeum v Českých Budějovicích, for instance, we used this method to test the exhibition "The Story of the City." The chatbot itself flagged the risk of text overload, the gap between what visitors expect and what they actually find, and weak points in the storytelling.
On top of that, this kind of feedback is available immediately. And you can get it before the exhibition opens its doors to the first real visitors.
We put together a guide that shares what we have learned:
- How to build a visitor persona that actually tells you something
- How to write a prompt that returns useful insights
- What materials to provide to the chatbot
- How to work with the chatbot's answers
You will find the link to the guide in the comments.
