Decision tool
Automation, assistant
or agent?
Get this wrong and you spend agent money on a rules problem, or trust a rules engine with a judgement call. Answer three questions about any task on your desk and see where it belongs.
This is the framework we run live in The AI-Augmented Executive. Share it, use it in a workshop, apply it to your own week.
How much do the steps change from one run to the next?
Could you write rules that cover the right outcome?
What happens if the call is wrong?
Answer the three questions.
Your verdict appears here.
The underlying rules
Three routes, three different cost structures
The tool above is a shortcut. These are the tests behind it.
- The steps are always the same
- Rules cover 95%+ of scenarios
- The output is checkable against a standard
- Speed and consistency matter more than nuance
- Every situation is different
- The work requires judgement or context
- Rules would need endless exceptions
- A person can review the output faster than producing it
- The outcome depends on things nobody wrote down
- The decision is hard to reverse
- Relationships, politics or trust are in play
- Being confidently wrong is worse than being slow
Outcome: automate consistency, apply AI to nuance, and keep judgement where judgement belongs. Most organisations get more value from moving three tasks to the right route than from adopting another tool.
Run this across your whole week
In the workshop, leadership teams map their own workflows against this framework and leave with a 30 / 60 / 90-day adoption roadmap. We can also run it as a working session with your executive team.