Including us. We would rather you asked these and went elsewhere than skipped them and had a bad year.
You do not need to understand how any of it works to have a good meeting. You need to know which questions separate people who have done this inside a real business from people who have watched a demo and bought a domain name. Take this list in with you.
What a good answer sounds like: They ask you, write it down, and want to watch someone do it. If they already know the answer without having seen your business, they are describing their product.
What a good answer sounds like: A number you both agreed before starting. Hours back, or a wait time, or an error rate. If the answer is about adoption or engagement, that measures their success.
What a good answer sounds like: They name your actual systems and say who connects them. Vague answers here are the single biggest predictor of a trial that goes nowhere, because a tool that cannot see your work cannot do your work.
What a good answer sounds like: They should already have thought about it. Who sees the mistake, what stops it reaching a customer, how you undo it. Anyone who implies it will not get things wrong is either inexperienced or not being straight with you.
What a good answer sounds like: A name, or a clear answer that it is you and here is how you will cope. Suppliers change forms, staff leave, prices move. Something will break. The question is only whether that was planned for.
What a good answer sounds like: A direct answer, in writing, in the contract. Whether it leaves the UK. Whether it is used to improve anything anyone else touches. If they have to check and get back to you, that is your answer for now.
What a good answer sounds like: Say plainly what you keep and what stops. Both are legitimate business models. What is not legitimate is finding out at the point you want to leave.
What a good answer sounds like: A year is the important part. Anyone can produce a happy customer from month two. The interesting question is whether people were still using it after the novelty and the champion both wore off.
What a good answer sounds like: Everyone should have an answer. Someone who thinks everything in your business is a good candidate has not thought about it properly, or is paid on how much they sell you.
Every one of these is really the same question in a different coat: have you done this before, in a business like mine, and are you still going to be here when it matters?
The risk has moved. The technology got good enough that a competent team can do a lot with it. The risk is in everything around it: whether anyone works out what the job actually is, whether it reaches the systems your work lives in, whether anyone measures it, and whether anyone is still there in month nine.
If you only ask two, ask the first and the fifth. What job does this replace and how long does it take today, and who is responsible when it breaks. Those two catch most of it.