Media / AI at work
12 questions to ask an AI vendor before you buy
A buyer's checklist covering data use, security, accuracy, pricing and exit plans for AI software.
Stackpick Editorial1 min readTHE SHORT ANSWER
Ask how your data is used, how accuracy is measured, what happens when the tool is wrong, how pricing scales, and how you can leave. Get the important answers in the contract.
In this guide
01
Your data
- Is our data used to train your models, and can we opt out in the contract?
- Where is it stored and processed, and for how long?
- Which subprocessors and model providers can access it?
02
Accuracy and reliability
- How do you measure accuracy for a use case like ours?
- What happens when the tool is wrong, and who is responsible?
- How will we know when the underlying model changes?
03
Security and compliance
- Which independent audits or certifications do you hold? Can we see the reports?
- Do you support single sign-on, role-based access and audit logs?
- How do you defend against prompt injection and data leaks?
04
Cost and exit
- What drives the price, and what happens if usage spikes?
- Can we cap spending or get alerts?
- How do we export our data if we leave?
05
Test it on your own work
Demos use the vendor's best examples. Run a short pilot on your own data with a success measure agreed in advance, and include the difficult cases.
Stackpick Editorial
Practical guides for people who buy and run technology.