If we don’t pause to look inward, we won’t see the problem in ourselves. Tomorrow we’ll be exactly where we are today, blaming someone else for it.

First ask where I am wrong before searching for blame in others.

Then make sure we really mean the same things by the same words — only then is disagreement about something real.

Q1 collects what each person actually does on a daily basis (today). Q2 collects what each person believes should happen tomorrow for the greatest possible outcome / effort input. The Reveal computes the items in Q2 that are absent from Q1 and highlights them as The Gap (bridge it).

Added value: the misalignment between intention and behaviour stops being a feeling and becomes a list of concrete items the room can argue about, assign, and bridge. This is the core CoSyP move — turn drift into a visible artifact.

Q3 is the only stage with no input field. The disc shows the Iceberg (what we say · what we mean · what we need) and the CoSyP root-cause quote — "When there’s an engineer to blame, we don’t miss the chance to point the finger... we treat symptoms instead of fixing the root cause." The host pauses. The room looks inward.

Added value: alignment cannot happen if no one stops talking. Q3 is the structural guarantee that 60 seconds of inward attention exists in every retro — even when the team wants to skip it. Skipping Q3 is not allowed; sacrificing Q4 or Q5 is (see #5).

The Reveal shows answers grouped per question with no participant attribution. The Gap is "items in Q2 absent from Q1," not "Alice said X, Bob didn’t agree." The host never sees who said what; the room never has someone to defend.

Added value: hierarchy and personality drop out of the conversation. People can say the uncomfortable thing without putting their name on it; the team responds to the content, not the source.

Three KPIs — Gap share, Airtime skew, Q4 reach — each paired with a one-line prompt the host literally says next. "Pick two gap items. Assign an owner and a first concrete step before the room leaves." The numbers do not judge the team; they tell the host what to ask.

Added value: a KPI is a question, not an answer. The host walks out with three concrete moves to make the next time the number lands in that band — not a dashboard to defend.

Q4 and Q5 are free-text at session creation (or live-edited mid-session). Default discs ship the Ginot model and the Levinson model; the host overrides them whenever the room calls for something else. And if alignment already feels strong, the host can skip Q4 or Q5 entirely — the protocol holds with fewer honest questions. CoSyP-confidence.

Added value: the shape of the retro adapts to the room without a config screen. The room signals strength by sacrificing a question slot; the host signals trust by letting them. That is the protocol playing out, not a feature toggle.

Every retro is saved automatically with its full anonymised answers, gap, and KPI snapshot. Multi-session KPIs — gap closure velocity, tomorrow→today translation, trust drift — light up once a team runs ≥2 retros. The premise: CoSyP works when the gap shrinks over time, with fewer questions needed each iteration. Every KPI is a measure of that compression.

Added value: a single retro is one sample. The protocol pays off across iterations — the room sees its own gap close (or not) over weeks, and the moderator nudges sharpen as the history accumulates.