Playbook
Five questions to ask before starting an AI project
Most AI projects that fail don't fail on technology. They fail on scope, data or measurement. We ask five questions before writing any code.
The checklist
- What decision or task will this change, and who owns it?
- What data do we have today, and is it good enough?
- How will we measure success in the first 30 days?
- What happens when the model is wrong?
- What will it cost to run at scale?
If we can't answer these in a two-week discovery sprint, we recommend not building yet — and say so.