Playbook · Part B The Engagement Protocol
Five phases. Each one produces what the next depends on.
The application protocol from signal selection through consequentiality review. Nothing produced for its own sake. The discipline is in the handoffs — each phase outputs the substrate the next phase needs, and only that.
The opening discipline.
One costly-and-absorbable signal-response chain, picked deliberately. The diagnostic depends on the signal being non-trivial; signal selection determines whether the engagement is rigorous or theatrical.
Decision criteria / discipline
- Costly — real economic, operational, or strategic consequence.
- Absorbable — the organisation can plausibly respond within a meaningful window.
- Bounded — the chain has identifiable t₀ (signal arrival), t₁ (recognition), t₂ (interpretation), t₃ (decision), t₄ (response committed).
- Recent — from the current strategic window. Historical chains drift outside the contemporary window assumption.
Common signal classes
Competitive, customer, regulatory, internal.
Output
A named signal-response chain, scoped, with t₀ anchored.
Handoff to next phase
Chain definition + t₀ anchor + window estimate.
The measurement phase.
Delay profile across the six capability stocks for the named chain. Measured, not assumed. Method: interviews structured against the PSI transaction pattern; artefact review (governance trace, decision-log review, retrospective documentation); t-value time-stamping against observable artefacts.
Decision criteria / discipline
- Each stock measured independently — pre-committing to a binding-constraint hypothesis contaminates the diagnostic.
- PSI-decomposed interview structure surfaces where coordination acts actually broke (missing promise, request without acceptance, broken execution chain).
- Inter-analyst reproducibility verification where multiple analysts have measured (Part C).
Output
A delay profile — t-values per stock — with the evidence each value rests on.
Handoff to next phase
Delay profile + evidence trail.
The interpretation phase.
Naming the binding constraint — the place leadership wasn't looking. The binding constraint is whichever stock has the largest measured-vs-window ratio, not the largest absolute delay. A 4-week delay in a 6-week window is more constraining than an 8-week delay in a 26-week window.
Decision criteria / discipline
- Triangulation against at least one independent reference: a parallel chain at the same organisation; a known industry comparator; or a prior diagnostic's consequentiality follow-up.
- The Anti-Confirmation Output 3b discipline — produce a stated alternative interpretation that, if correct, would invalidate the binding-constraint finding.
- Specify the disconfirming signal that would surface if Output 3b is correct.
Output
Named binding constraint + Output 3b alternative interpretation + the disconfirming signal that would surface if the alternative is correct.
Handoff to next phase
Binding constraint + Output 3b.
The prescription phase.
Re-sequenced actions that fit existing capability and authority. The intervention targets the binding constraint specifically. Capability investment in non-binding stocks is wasted budget at best, load-adding at worst. The intervention is resequenced, not additive — most SRD findings produce re-ordering of the existing programme, not new programmes.
Decision criteria / discipline
- Q1 — protect what works; automate where false-positive rates are bounded.
- Q2 — invest in the augmentation layer with anti-overreliance discipline (Part D).
- Q3 — fix the measurement pathway; re-route, don't re-build.
- Q4 — distinguish pathology from posture before prescribing.
Output
Intervention specification + sequencing recommendation + the specific quadrant placement that justifies the intervention shape.
Handoff to next phase
Intervention design + consequentiality measurement framework.
The anti-confirmation discipline applied across time.
6- and 12-month follow-up of what changed, what didn't, and what the diagnostic missed. Pre-engagement commitment to the cadence is mandatory — engagements without committed follow-up cannot claim SRD-diagnostic status (Part C).
Decision criteria / discipline
- At each follow-up, the diagnostic claim is evaluated against the observed outcome.
- Positive outcomes the diagnostic predicted strengthen the claim; positive outcomes the diagnostic did not predict are flagged as the diagnostic having missed something.
- Negative outcomes test the diagnostic against the alternative interpretation from Output 3b.
Output
A consequentiality record that lives with the engagement permanently. Findings that the consequentiality review disconfirms get amended in the Pattern Library (Part E).
Handoff to next phase
Engagement closes; pattern record opens.
Phases compound; shortcuts compound too.
The five phases are sequential, but the discipline against each is reinforced by Part C (diagnostic disciplines) and Part D (anti-overreliance for AI-augmented engagements). Reading Part B alone gives the shape; reading Parts C and D alongside gives the load-bearing protections against the diagnostic confirming itself.
The diagnostic's value is partly in the binding constraint named, and partly in the alternative interpretation that survives the engagement. An engagement without Output 3b is not an SRD diagnostic.