Choosing who builds your AI product or automation is as important as the idea itself. A few direct questions upfront can save months of rework later.
1. Have you actually shipped something like this to real users?
A polished portfolio of demos isn't the same as having something in production, handling real, messy user input every day. Ask specifically what happened after launch, not just what the build process looked like.
2. What happens to my data?
Any AI product involves sending data somewhere, whether that's a model API or a database. Ask exactly which providers are involved, where data is stored, and what happens to it if you end the engagement.
3. How do you handle the AI getting it wrong?
Every AI system produces wrong or low-confidence answers sometimes. A partner who has a real answer for how the product behaves in that case (rather than "it usually doesn't happen") is one who has actually dealt with production usage before.
4. What does this cost to run once it's live, not just to build?
Build cost and running cost are two different conversations. A partner who can only speak to the first one hasn't thought through what happens after launch.
5. Who owns the code and the systems when we're done?
Get clarity upfront on whether you own the finished product outright, including the ability to hand it to another developer later if you ever need to. This should be a simple, direct answer, not a vague one.
Why this matters more than the pitch
Anyone can make a build sound impressive in a sales conversation. These five questions are less about the pitch and more about whether the partner has actually lived through the parts of the process that don't show up in a demo.