01
Identify what matters.
Clarify the factors that should inform the decision.
WHERE AIQU FITS
The desired outcome is clear. The reasoning is not yet solved.
AIQU works between analysis and implementation — where the question is understood, but the logic that should drive the result still needs to be found, built and tested.
RECOGNITION
You may already know what the data should make possible.
02
Prioritise options.
Assess and rank possible courses of action.
03
Compare like with like.
Bring consistent evidence together to enable a fair view.
04
Recommend what to do next.
Turn analysis into a clear, justifiable recommendation.
05
Forecast what may happen.
Estimate likely outcomes based on available data.
06
Automate a recurring decision.
Apply the same logic consistently at scale.
THE TENSION
The gap is rarely a lack of tools.
The outcome is defined
- The decisions that need to be made are clear.
- The results that are required are understood.
- The value of getting it right is evident.
- There is a need to move from question to solution.
The path is unresolved
- Key definitions may still be uncertain.
- The right assumptions are not yet agreed.
- The logic that should be applied needs to be developed.
- The method or algorithm is not yet determined.
- The approach needs to be validated with real data.
NOT A CATEGORY
The work is not defined by a service label.
- GENERIC ANALYSIS
- DATA ENGINEERING
- SOFTWARE DELIVERY
- AI PROJECT
The problem determines the combination.
AIQU may investigate, model, validate, automate or build — but the work is organised around the intended outcome, not a predetermined discipline.
FIT / NOT FIT
Fit and not fit
A strong fit
- The outcome can be described.
- Real data exists or can be accessed.
- Reasoning needs development.
- Testing and iteration are possible.
Probably not the right fit
- The method is already fully specified and only execution is needed.
- The request is generic capacity without a defined outcome.
- Evidence cannot be accessed or tested.
DISCUSS AN IDEA
What do you want the data to make possible?
Start with the intended outcome. The method can be worked out from there.