Whose perspective might an AI design workshop be missing?
Ask who handles the difficult exceptions, knows a source system's limits or hears the customer's complaint when a process fails. Invite them to review the part of the AI design they understand, and ask a narrow question, such as what would make one proposed step fail.
A participatory workshop can still miss important knowledge when it includes only the people most comfortable talking about the tool. Look for expertise in the work as well as enthusiasm for the technology.
Offer more than one way to contribute. A written example, a short individual conversation or a review of a sample output can work better than a big workshop.
I'd keep the question specific. Asking someone to assess a whole AI strategy tends to get less than asking where one proposed step would break. Show them how their input affected the design, and credit it if they say yes.
Don't label anyone quiet or resistant. They hold relevant knowledge that a confident conversation about new technology might otherwise miss.