← Insights
Buying
Twelve questions to ask before you buy an AI tool
A buyer's checklist that keeps the purchase tied to the initiative it is supposed to serve, and protects your data on the way in and on the way out.
AI vendors are good at demos. Your job as a buyer is to pull the conversation back to your initiative, your data and your exit. These twelve questions do that. Send them before the second meeting and pay attention to which ones get vague answers.
Fit
- Which initiative and outcome does this serve, in one sentence?
- What will we stop doing, or stop paying for, once this is live?
- Does a tool we already own cover most of this?
Data and security
- What data leaves our environment, and where is it stored and for how long?
- Is our data used to train models, and can we opt out in the contract, not just in settings?
- Does it respect our existing permissions, or create a new copy everyone can see?
- Can we get an audit log of what it read, wrote and decided?
Cost
- How does pricing change as usage grows: per seat, per task, or per unit of usage?
- What does the pilot really cost once staff time is included?
- What result at 90 days would justify the full contract?
Exit
- Can we export our data and configuration in a usable format?
- What breaks if we turn it off tomorrow?