AI
Ask me before you add AI to your product.
Adding AI to a product sounds like progress. Sometimes it is, when it solves a real problem. More often it is an expensive way to discover that the real problem lives somewhere else. A short check looks at where AI actually adds value, and where it adds complexity, maintenance and supplier dependency.
Check this before you commit
What I check first.
- 01
Which specific problem does the AI feature solve?
- 02
What does your product do without AI, is it already working?
- 03
How do you measure success, accuracy, speed, cost reduction?
- 04
What is the cost per user per month at realistic usage?
- 05
Which data goes to the supplier, and is that acceptable to your customers?
Common patterns
What I see happen most often.
- AI gets added because it has to be on the pitch deck.
- A process problem is solved with technology instead of a process change.
- The per-user cost curve becomes a surprise as usage grows.
- A model update unexpectedly breaks production behaviour.
What the conversation produces
What you leave with.
- 01
Clarity on whether AI adds value here or complexity.
- 02
Visibility into cost, maintenance and reliability.
- 03
A workable next step: build, delay, or approach the problem differently.
Further reading
Articles on this challenge.
- 6 min read
AI Is Not a Strategy
AI can accelerate, but it also accelerates wrong decisions. When does it add value, and which questions to answer before you add it?
- 7 min read
AI Can Make You Faster at Making Mistakes
AI reduces the cost of producing software. It does not reduce the cost of choosing the wrong software.
- 8 min read
What Should Be in Your MVP and What Should Wait
A decision framework for keeping the first release complete, credible and much smaller.