Two AI plans, one board meeting
The questions that separate a strong plan from a demo

Plan A
Transform customer service with a leading AI vendor. A strong demo. But when the questions come, the answers are vague:
- The outcome is "better experiences".
- The cost is "a licence".
- Quality is proven by the demo.
- Ownership sits with IT.
- Nobody has thought about stopping.
Plan B
Smaller. Cut the time to answer supplier invoice queries from three days to one. It shows:
- a cost per query, compared with today
- a test set of 500 real queries with a pass mark
- a named owner in finance
- a stop rule: if the time saving is not visible within three months, it stops
The decision
Plan B gets funded. Not because it is less ambitious, but because it is the one that can prove it worked. Once it does, it earns the right to grow.


