15 questions to ask an AI vendor
Use these before you sign anything. A good vendor answers each one clearly and in writing.
Outcome and fit
- 01What specific business outcome is this meant to improve, and how will we measure it?
- 02Can you show it working on a task like ours, and what did it take to get there?
- 03What does it do badly or not at all?
Data and privacy
- 04What information of ours does the system see, store or learn from?
- 05Is our information used to train models that other customers use?
- 06Where is our information stored, who at your company can see it, and for how long?
Control and oversight
- 07Which decisions and messages can a person review or approve before they take effect?
- 08Can we see a history of what the system did and why?
- 09What happens when it is wrong, and how will we find out?
Delivery and support
- 10Who builds and maintains it, and who do we call when something breaks?
- 11How long until we see a first result, and what do you need from us to get there?
- 12How do you handle changes to the underlying AI models?
Cost and exit
- 13What is the full cost, including setup, usage, and anything that grows with volume?
- 14Can we take our data and our configuration with us if we leave?
- 15What does the contract say about ending early?
Stay in touch
Get the next one by email.
A short note from Northfalk on AI, strategy and building what comes next.
Start a conversation
Weighing up a vendor, or deciding whether to build?
A discovery call is a conversation about what you are working toward.