01
Pattern match
The question’s structure is compared with the 60-category question-style bank.
EXAMPLE CASE STUDY
OrganisationTesting an operating model change before the board meeting
See how Pygar turns a board-level operating-model choice into a testable decision with explicit evidence, people impact and reversal conditions.
01 · ASK
The question as the user posed it
Pygar begins by preserving the user’s wording. Before any answer is formed, the engine separates the stated decision from its assumptions, constraints and missing evidence.
USER INPUT
Should we move from functional teams to product-aligned units next quarter, or improve the current model first?
WHAT THE OPENING FRAME HIDES
The opening frame creates a forced choice between wholesale redesign and incremental improvement. It leaves the underlying failure mechanism, customer impact, transition capacity and evidence threshold unstated.
HOW PYGAR DETERMINED THIS
The engine looks for compound choices, hidden thresholds, assumed causes and missing decision criteria. These signals reveal where the opening wording may be narrowing the available reasoning.
02 · CLASSIFY
How Pygar chose the reasoning route
The classifier does not decide what the user should do. It identifies the structure of the question and selects the reasoning family, depth mode and prompt set most suited to examining it.
ROUTE SELECTED FROM THE 60-CATEGORY BANK
Matched within Pygar’s 60-category question-style bank as Organisational design and governance. Family: Systems and Execution. Mode: Standard. This classification selects the prompt set and five independent roles; it does not decide the answer.
02
Family route
The dominant reasoning demand selects a cognitive family and operating mode.
03
Prompt assembly
The route builds five independent briefs without determining the answer.
03 · DELIBERATE
Five independent perspectives
Every panel role receives the same classified question with a distinct reasoning brief. The roles work independently, so one convincing argument cannot silently shape the others.
5
isolated reads, each optimized for a different failure mode in human judgement.
Lead
Clarify the missionThe immediate decision is not whether product units are fashionable. Define what the board is being asked to approve, which customer or delivery problem must improve, who owns the decision and what evidence must exist by the decision date.
Depth
Judge the contextThe question assumes structure is the main cause. Compare the current model on decision latency, handover failure, duplicated work and customer outcomes so that unclear rights, incentives or overloaded interfaces are not mistaken for an organisation-chart problem.
Test
Stress-test the frameAssume the redesign disappoints six months after launch. Likely failure points include ambiguous product ownership, specialist capability being fragmented, duplicated platforms and a transition that consumes the capacity needed to serve customers.
Ground
Apply real constraintsMap the handovers, system dependencies, regulatory responsibilities and roles that cannot move at once. A bounded pilot in one value stream is more feasible than an organisation-wide change if service continuity must be protected.
Human
Check people and impactName who carries uncertainty during the transition. People need clear answers on performance, progression, specialist identity and workload before the new structure can be judged fairly or adopted safely.
04 · SYNTHESIZE
Where the panel converged and what remained unresolved
The synthesizer compares the five reads for agreement, contradiction, shared assumptions and missing evidence. A separate verification pass then checks whether the combined answer is clear, bounded and honest about uncertainty.
PANEL SYNTHESIS
What the five reads establish together
The five reads converge on an evidence gap rather than a simple choice of models. The case for change is plausible, but approval should be for a time-boxed test with explicit decision rights, protected capacity, customer measures and a reversal condition. That separates a direction of travel from an irreversible redesign.
VERIFICATION PASS
What the answer must not conceal
The board still needs evidence that the target problem is structural, that the pilot can isolate the effect of the new model, and that leadership has enough capacity to run the transition. Those uncertainties should remain visible in the decision paper.
HOW PYGAR PRODUCED THIS
The engine preserves useful disagreement, prioritises findings supported across more than one read and keeps unsupported certainty visible. Verification can trigger a bounded second synthesis when the result is unclear or incomplete.
05 · REFRAME
Turn the findings into better questions
Pygar does not merely replace the user’s question. It shows how the frame changed, then offers three deliberate operations that can expose sharper evidence and unexplored lines of consideration.
WHAT CHANGED IN THE FRAME
The question moves from Which model should we choose? to Which operating problem are we solving, how will we test the proposed mechanism, and what evidence will make us continue, change course or stop?
CLARIFY
Make the decision and evidence threshold more precise.Which recurring customer or delivery failure must product-aligned units reduce, and what measurable change would count as success?
WIDEN
Introduce a credible alternative or overlooked stakeholder.What evidence would show that decision rights, incentives or handovers could be repaired without changing the whole structure?
DEEPEN
Probe the assumption most likely to change the decision.If the pilot improved delivery speed but weakened specialist quality or staff retention, how should the board weigh that trade-off?
REFINED QUESTION READY TO ASK
Which operating problem are we trying to solve, what evidence shows structure is the cause, and how can we run a reversible pilot that tests customer outcomes, decision speed, specialist quality and people impact before committing to a wider redesign?
06 · DECIDE
Convert insight into accountable human action
Pygar finishes by making the next evidence-producing action explicit while protecting the human decision boundary. The engine expands judgement; it does not replace authority, context or responsibility.
WHAT THE USER CAN DO NEXT
A bounded action that produces evidence
Translate the panel’s findings into a 30-day operating-model pilot. Name the decision owners, define the hand-offs that will change, select measures for delay, rework and escalation, then review the evidence before wider rollout.
HUMAN DECISION BOUNDARY
Where Pygar stops
Pygar broadens the frame and makes the evidence gaps testable. The board and leadership team retain the authority, context and responsibility for the final operating-model decision.
1
structured recommendation, tested against five perspectives and returned to the person or team accountable for the choice.
