AI delivery · Toolkit 02
The AI Opportunity & Value Model
Work out what a manual process actually costs you before deciding whether to automate it. The arithmetic that turns "this feels wasteful" into a number a budget holder can approve — or refuse.
6 pages · PDF
A look inside
5 clients/week × 30 min = 150 min/week ÷ 60 = 2.5 hours/week × £50/hr loaded = £125/week × 52 = £6,500/year — direct, conservative
The baseline, which is the floor. The document then models the compounding curve that takes the same automation to roughly £18,000 by month twelve — and tells you when the number is small enough that you should not build at all.
The rule it turns on
If a system cannot be shown to return several times what it costs, the scope is wrong. Widen it, trim the price, or don’t build it.
What’s in it
- The five inputs to collect, and how to get each one honestly
- The baseline calculation, with a worked example
- The second-order effects worth quantifying, and the one most often left off
- The compounding curve — why flat annual figures understate a good system
- How to set a price from the number rather than from effort
- Six checks before the figure goes anywhere near a decision
- Two honest limits, including when not to use this at all
Related reading, ungated: the AI governance non-negotiables.