A problem is the gap between the problem owner's current state (A) and ideal state (B)
The Framework
A and B are embedded in a complex environment — surrounded by data, forces, and unknowns
The Framework · Step 3
Engage with the context, collect data, and progressively narrow where A and B are
Zooming into Step 3
This step has its own internal loop — assumptions, testing, and narrowing
Step 3 — Inside
What we know about the field and context lets us make educated guesses — not random ones. LLMs can help retrieve domain knowledge, but it reflects the past — verify against current reality.
Step 3a — Form
Use what we already know to hypothesize what's true about the gap — before we have full data
Validate ✓
confirmedRoot cause is in subsystem X
confirmedConstraint from process layer
testing…Scope bounded to team level
Invalidate ✗
rejectedB is further than expected
rejectedExternal factor not a cause
testing…Timeline assumption holds
Scope narrowed
Both outcomes tighten the search space — invalidation is not failure
Step 3b — Test
Data confirms or rejects each hypothesis — both narrow the scope
Step 3 — The Loop
Each cycle tightens understanding — until A, B, and the gap are clear enough to act on
The Framework · Step 4
With scope narrowed through assumption loops, the problem is precisely defined — and solvable
1
2
3
↓
3a
3b
↻
↑
4
1 / 9