# Deployment Toolkit I — The Engagement Arc (R-00, C-Series, E-Series, A-Series) > Part of the public canon of *Perceptive Innovation* by **Caleb Kow** — > canonical URL: https://perceptiveinnovation.com/canon/toolkit-engagement-arc.md · index: https://perceptiveinnovation.com/llms.txt > © 2026 Caleb Kow, all rights reserved. Quotation with attribution to Caleb Kow / perceptiveinnovation.com is welcome. > Commercial use — client engagements, training, certification, or incorporation into products or models — requires a written licence: https://perceptiveinnovation.com/terms Source: *The Perceptive Innovation Deployment Toolkit* (Second Edition, 2026) by Caleb Kow — 48 field instruments across the full engagement arc: **ENGAGE → PREPARE → DIAGNOSE → INTERPRET → PRESCRIBE → RUN**, then loop. "PIPAS is the diagnosis. The 10 T's are the prescription. The audit is the engagement through which both happen." Cite instrument codes (C-01, W-04, A-04…) in outputs — codes are stable and clients recognise them. This file covers engagement setup, evidence gathering, and room craft; `toolkit-diagnostic-craft.md` covers the diagnostic instruments themselves. ## R-00 — Rules of the Kit (Ten Non-Negotiables) Break one and the reading stops being a reading. 1 **Probabilistic, always** — perception shapes adoption and holds a probabilistic veto; it never determines, guarantees, or wins (A-04 lists the substitutes). 2 **Canonical order** — spheres Identity→Future, T's Technology→Tangibility; comparability across audits depends on it. 3 **One audience per sheet** — never blend audiences into one score; the book's veterinary case read 23 and 15 on the same day for the same product, and an average would have hidden the veto. 4 **A named decision** — no audit without a decision it informs, owner and date on W-03; a reading that informs nothing is theatre. 5 **Evidence beats altitude** — a rung-1 observation from the most junior voice outranks a rung-5 conviction from the most senior; the sponsor scores last, in writing. 6 **Distribution before composite** — read the lowest sphere first, then the widest dissent; a 22 that is falling is not a 22 that is climbing. 7 **Two spheres, one lever each, one quarter** — W-09's ceiling; if everything is a priority, nothing is. 8 **Low-confidence flags stand** — a sphere scored only on rungs 4–5 stays flagged whatever the number says (the engineered-trust caution); flags convert to monitoring hooks, never to silence. 9 **Dated or decorative** — every score carries a date and becomes next quarter's baseline; write once, use twice. 10 **Complement, never displace** — PIPAS integrates Christensen, Kim & Mauborgne, and Moore (I-01…I-04); if the kit is used to dismiss those lenses, it is being misused. ## C-01 — The Engagement Qualifier The 30-minute scoping call, structured. **Symptom checklist** — tick what the client describes unprompted, each with its private sphere reading: technically superior product losing to a weaker rival (*Trust/Identity*) · adoption far slower than pilots and demos promised (*Effort/Risk*) · sales cite price but discounts don't move the number (*Value, masked*) · users love it, buyers and gatekeepers block it (*audience split → W-05*) · great press and awards, thin pipeline (*Temperature ≠ Trust*) · internal certainty, external silence (*gap → W-12*) · every loss has a different explanation (*no coding → E-04*) · category baggage, "we've been burned before" (*Trace/Trust*) · churn concentrated right after onboarding (*Effort/first experience*) · roadmap questions dominate sales calls (*Future*). **Auditability — four gates, all must pass:** a real decision in the next 90 days this reading would inform · the economic owner will sit in the room, not send a delegate · they can open rung 1–3 data (usage, tickets, win/loss, verbatims) · the sponsor will score last and accept dissent on the wall. **Gate result:** GREEN (3+ symptoms, 4 gates) → propose Format B, send E-01 same day · AMBER (symptoms real, a gate fails) → propose Format A to earn the mandate, fix the gate first · RED (validation of a made decision, or no evidence access) → decline or park; an advocacy audit damages the method and the client. Do not teach the framework on this call — qualify, listen, take verbatims; the client's own words become the opening slide. ## C-02 — The Discovery Interview Guide 45 minutes that gather sphere evidence **without ever naming a sphere**. Structure: 0–5 open (role, relationship, last touch — no framework words, here or anywhere) · 5–35 the ten probes, chasing the last specific instance, never the general opinion · 35–42 quantify ("If you had to bet: does this get adopted by [audience] within [horizon]? What would change your bet?") · 42–45 close ("Who sees this differently from you — and who else must I hear?" Book them before you leave). **The ten probes** (right column for your ears only): walk me through the last time you used/evaluated/sold it (*behaviour, R1*) · what did you expect it to be before you touched it, and where did that come from? (*mental model → W-11*) · when you describe it internally, what words do you use? (*Identity*) · what almost stopped you — or still does? (*Risk/Trust*) · if it disappeared tomorrow, what would you actually lose? (*Value*) · what's the annoying part nobody fixed yet? (*Effort*) · do you believe the claims — which one least? (*Trust*) · where is this whole category going, and is this on that path? (*Future*) · what would you need to see to double your commitment? (*lever clue → W-08*) · what does your most sceptical colleague say? (*dissent → W-13*). **Interview rules:** capture verbatims, not summaries — the exact words carry the sphere. Silence is data; count the pause before "yes, we trust it." One instance beats ten opinions. Never correct, never sell, never defend. ## C-03 — Stakeholder Power × Perception Grid Plot every stakeholder by veto power (high/low) and current stance (against/for); re-plot after every interview — people move. Quadrant doctrine: **Blockers** (high power, against) — audit these cells first; a two here outranks everything · **Champions** (high power, for) — your rung-1 evidence sources and reference engine · **Sceptic chorus** (low power, against) — loud voice, Temperature risk not veto risk · **Passive base** (low power, for) — do not over-invest; protect via Tenacity signals. Register per stakeholder: power H/M/L · stance −/0/+ · interview? · seat (A-02)? · W-05 column. Rule of the grid: weight by leverage, not volume — one quiet Blocker with sign-off authority outweighs a chorus of low-power sceptics and a thousand passive fans. ## C-04 — Engagement Formats & Timelines **Format A — 90-minute executive briefing** (earn the mandate): reflex-test debrief with their aggregate on the wall → post-mortem harvest on one shared war story → the reveal (the six piles already are the spheres) → frame the real audit, W-03 drafted live → the ask: one day, six seats, the E-01 pack. Output: signed W-03 + data request out. **Format B — one-day perception audit** (the standard unit): see A-03 run sheet. Output: 2 signed W-09 + W-13 hooks + dated baseline. **Format C — two-day portfolio sprint** (allocate across bets): Day 1 = Format B on the flagship; Day 2 am = 40-minute spheres-only triage × 5–8 innovations (one W-04 row each, no interventions yet); Day 2 pm = portfolio view on W-05, budget by weakest cells, cadence booked. Output: portfolio matrix + funded priorities. Never compress the first Format B below one day — speed comes from repetition (quarterly re-audits run in half a day), not from skipping calibration. ## C-05 — The Assumption Register Logs every belief the client treats as fact, in their exact words — the wording is evidence too ("Enterprise buyers love the dashboard." "Price is the objection." "The category is inevitable."). Fields: assumption verbatim · held by · sphere · status **V/U/C** · test to run (E-0x) · by when. **V** validated — supported by rung 1–3 evidence in the pack · **U** unvalidated — nobody has looked; assign a test, not a debate · **C** contradicted — evidence says otherwise; escalate to the sponsor same day. Technology-first organisations hold their assumptions as specifications; treat each as a hypothesis with a due date — the audit scores what survives. ## C-06 — Boundary & Misuse Check Audits the audit before findings ship; any unticked box reopens the relevant sheet. The ten compliance boxes operationalise R-00 (one audience per sheet · canonical orders · signed named decision · silent scoring, sponsor last · spread revealed, ≥2 dissents logged · evidence line per score, rung-4/5 spheres flagged · distribution before composite, lowest sphere named first · max two target spheres, one lever each, ≤90-day review · no "determines/guarantees/will win" anywhere · baseline dated, re-audit booked before the room emptied). **Scope guard — what this method is not:** not technical due diligence (PIPAS reads perception of the technology; it does not verify it — pair with engineering review) · not market sizing (a 28/30 in a market of nine people is still a market of nine — pair with TAM work) · not unit economics (perception can shift willingness to pay; it cannot repeal cost structure) · not a brand tracker (trackers monitor; the audit decides — feed tracker data in as rung 3–4) · not a crystal ball (the reading shifts odds and exposes veto risk; it does not promise outcomes — say so, in writing). **Stop the engagement if:** the sponsor edits scores after convergence · findings are weaponised against a function · access to the Listener seat is withdrawn. The method's credibility is worth more than any single engagement. ## C-07 — Discovery Synthesis One page that closes discovery and backs the proposal: top three symptoms in the client's words · suspected weak spheres · evidence rungs on hand (R1–R5 tick) · data gaps → E-01/E-02 request lines · stakeholder grid summary with the highest-leverage blocker · assumptions at risk (the U's and C's touching the named decision) · recommendation (Format A/B/C, the named decision it informs, dates, signatures). ## E-01 / E-02 — Data Requests (Per Sphere, Per Rung) Request format: last 12 months · raw exports, not dashboards · include lost, churned, and stalled accounts, not just happy current ones · verbatims with speaker role attached · one folder per rung. Rungs 4–5 are deliberately absent — proxies and beliefs are already in the building. Sphere seams (R1 behaviour / R2 unprompted / R3 solicited): **Identity** — community joins, referral shares, self-badging, adopter segment mix / "companies like us" language, how customers introduce you internally / brand-fit items, "who is this for?" open-ends. **Value** — conversion, renewal, expansion, WTP tests, discount depth actually required to close / "worth it / can't justify" deal-note verbatims, procurement pushback / value items, win-loss reasons, pricing-page recordings. **Trust** — security reviews passed, references granted, pilot→production rate / "will it actually…" objections, analyst commentary / credibility items, claim-by-claim belief testing. **Risk** — exit clauses exercised, indemnity demands, reversible-pilot structures / procurement redlines, "what happens if it fails / you fold" logs, career-risk language / perceived-risk items, reversibility ranking, regulator themes. **Effort** — time-to-first-value, activation drop-offs, training hours, ticket volume by stage / "how do I…" themes, workaround behaviour / ease items, task-completion tests, admin burden interviews. **Future** — multi-year commitments signed, partners building on your APIs, ecosystem job postings / "will they be around" call themes, roadmap-question share of QBRs / direction items, analyst trajectory, investor sentiment. Confidentiality line for the request: evidence is used solely to ground scores in this engagement; verbatims are quoted with roles, never names, and destroyed on request at close. ## E-03 — The Survey Question Bank Four items per sphere, agree scale 1–5; n ≥ 30 per audience before averaging anything. **Identity:** using this fits how I/we see ourselves · people like me are the ones this was built for · I'd be comfortable being known as a user of this · adopting it says something I want said about us. **Value:** the benefits clearly outweigh the total cost of adopting · I can explain the payoff to a sceptic in one sentence · compared to alternatives this is the better deal overall · if the price rose 20% I would still choose it. **Trust:** I believe this will perform as promised · the company behind it will stand behind it · independent evidence supports the main claims · if something went wrong they would put it right. **Risk:** if this failed the damage would be manageable · I could reverse this decision without lasting cost · choosing this would not put my standing at risk · I understand what could go wrong before committing. **Effort:** getting started looked (or was) easy · learning it did not disrupt how we already work · keeping it running takes little ongoing attention · I rarely need help to get what I need from it. **Future:** this is where the category is heading · this product will still be here, improving, in three years · betting on it now will look smart later · the ecosystem around it is growing, not shrinking. Rules: randomise item order · two screeners (role, actual exposure) · report by segment before any overall mean · a survey is rung 3 — it can support a score, never source one alone · any sphere mean ≤2.4 or ≥4.2: pull the verbatims and interview five respondents. ## E-04 — Win/Loss & Churn Coding The richest rung-2 seam most organisations never mine. The five questions, same five every time, verbatim answers: walk me through the moment you decided · what almost stopped you — or finally did? · who else weighed in, and what did they say? · what would have changed the outcome? · how did you describe us internally? Log per account: outcome W/L/C · stated reason in their words · sphere(s) · rung 2/3 · vs whom. Coding rules: code the concern, not the vocabulary — "too expensive" after a demo they couldn't follow is Effort wearing Value's coat · double-code sparingly (two spheres maximum) · count unique voices, not repetitions of one loud voice · reasons volunteered in their words are rung 2; answers to your questions are rung 3 — mark honestly. ## E-05 — The Verbatim Coding Guide Cue phrases per sphere (and cousins): **Identity** "not for companies like us" · "we're not that kind of team" · "our customers would laugh" · "feels enterprise-y / toy-like". **Value** "nice to have" · "can't justify it upstairs" · "what does it replace?" · "the maths doesn't work". **Trust** "too good to be true" · "who else uses it?" · "prove it" · "we've heard that before". **Risk** "what happens if…" · "can we roll back?" · "who gets blamed?" · "what if you get acquired?". **Effort** "another thing to learn" · "who maintains this?" · "integration is the killer" · "we don't have the people". **Future** "is this where things are going?" · "will they be around?" · "betting the roadmap on it" · "the ecosystem is thin". Rules: code the concern, not the noun — "price" from someone who never saw the payoff is Value; "price" from someone fearing a failed rollout is Risk · one primary sphere per verbatim, a second only when genuinely dual; park ambiguous items and resolve in pairs · keep the speaker's role attached — a CFO's Risk is not an engineer's Risk, and W-05 needs the column · double-code ten percent by a second coder; disagreement above two-in-ten means recalibrate before continuing. ## A-02 — The Facilitator's Room Card **The six seats — people, not functions; no delegates:** Economic owner (carries the number the decision affects) · Builder (knows what the product actually does) · Seller (hears the objections first-hand) · **Listener** ★ (support/CS — closest to rung-1 evidence; non-negotiable) · Doubter (the informed sceptic, invited on purpose) · Outsider (optional; adjacent-team eyes, no stake in the answer). **The six steps:** brief (calibrate the scale on a neutral book case, not the product) → silent score (in writing, one evidence line per sphere, firm/thin marked) → reveal the spread (post all scores before any averaging; the spread is a finding, not a nuisance) → calibrate (extremes speak first; evidence beats altitude; sponsor scores last, always) → converge + dissent log (agree a number per sphere; every ≥2-point spread becomes a monitoring hook) → read the profile (distribution before composite; lowest sphere first; match against W-07). **Watch-fors — the seven ways audits die:** HiPPO collapse (the room waits for the senior score → sponsor scores last, in writing) · advocacy audit (scoring to justify a made decision → pre-mortem framing: "it failed; why?") · harmony trap (all threes, no argument → send it back; an audit without dissent has not happened) · the ambush (scores used to corner a function → pre-work for everyone, or postpone) · inflation drift (quarterly scores creep up unearned → re-anchor on the book's calibration cases) · map for territory (polishing scores instead of reality → the metric is the audience, not the sheet) · framework tourist (audits that inform no decision → W-03's named-decision rule, enforced). ## A-03 — One-Day Audit Run Sheet (Format B) 8:45 room set (six wall columns, sealed W-01s collected, packs by rung) · 9:00 reflex debrief, aggregate only — "nobody chose this bias" · 9:20 post-mortem on one shared war story, their words on stickies · 10:10 the reveal — cluster to six piles, then name the spheres; the room notices its story already sorted itself · 10:45 frame + evidence walk — sign W-03 before the pack opens; walk evidence by rung, strongest first · 11:30 calibration on a book case — argue here, cheaply · 12:00 silent scoring, sponsor's sheet face-down · 13:15 reveal + calibrate — extremes first, evidence beats altitude, sponsor last; log every ≥2 spread as a hook · 14:30 profile + deep-dive the lowest sphere (D-0x) · 15:15 levers (W-08 × W-15; maturity 1–2 on a high-affinity lever = the alarm) · 15:45 two intervention briefs — owners in the room say yes in the room, or the brief doesn't exist · 16:30 metrics + cadence — baseline written once, used twice; calendar invites sent before anyone stands up · 16:50 close — plot today's W-14 point, book the re-audit aloud, thank the doubter by name. Timing discipline: the afternoon always wants the morning's minutes; protect calibration and silent scoring at all costs — everything downstream inherits their honesty. If a block must die, kill the deep-dive, never the calibration. ## A-04 — Executive Readout Template One slide, within 48 hours, ten minutes of any executive agenda. **Headline:** "[Innovation] with [audience]: composite [XX]/30, [band] — [climbing / falling / first reading]." **The shape:** the six bars. **Three findings, formula not prose:** lowest sphere + the verbatim that set it ("[sphere] at [n] — '[their exact words]' ([role])") · widest dissent + its hook ("[sphere] split [a–b]; watching [trigger] by [date]") · largest belief–evidence gap ("we believed [x]; the pack read [y] — [owner] has it"). **The two moves (W-09):** "[sphere] → [lever]: [one-sentence move] — [owner], review [date]" ×2. **The ask:** the named decision from W-03 and what this reading recommends about it. **Language discipline — the probabilistic register, enforced:** never *determines/dictates* → say *shapes · weighs on* · never *guarantees success* → *shifts the odds toward* · never *will win / will fail* → *carries veto risk at [sphere] · is positioned to* · never *the market thinks* → *[audience] at [date] read it as — per [rung] evidence* · never *perception is reality* → *perception holds a probabilistic veto over adoption*. ## A-05 — Objection Handling Card "Perception is just marketing's job" → perception is constructed at every touchpoint — a procurement redline, an onboarding stall, a security review — most of which marketing never sees; the audit reads all six spheres across all six audiences. "We already track NPS" → keep it, as one rung-3, lagging, blended input; NPS averages six spheres and six audiences into one number after the fact — the audit separates the why while there is still time to move it. "This feels subjective" → anchored descriptors, evidence lines per score, a ladder that ranks proof, and dissent logged rather than averaged; inter-rater spread isn't noise here, it is instrumented — more epistemic discipline than most roadmap debates. "Our technical advantage speaks for itself" → the book opens with three chapters of superior technologies that lost; technically better is one lever of ten, and the audience holds the veto. "We can't spare a quarter" → the audit is one day; the ninety days run parallel to work you were doing anyway — compare with the fully-loaded cost of the last stalled launch (box 7 of your own post-mortem has the number). "Won't we just depend on you?" → the kit is transfer by design: A-02 trains an internal facilitator, the cadence runs without me, the re-audit is deliberately boring; my exit is a milestone on the roadmap, not a risk to it. Tone rule: never defend the theory; demonstrate the instrument — a sceptic who scores one sheet honestly converts faster than one who loses one argument politely. ## A-01 — The Operational Cadence Card Perception management as rhythm, not project. **Weekly, 15 min** — perception-signal review inside an existing stand-up (new verbatims, support themes, sentiment spikes); no new meeting — borrow one; owner: working-group lead. **Monthly, 60 min** — sphere dashboard against the dated baseline: which of the six moved, which rungs supplied the movement, which hooks fired; owner: perception owner. **Quarterly, half day** — re-audit: fresh W-04 per priority audience, W-09 reviewed against its kill-or-adapt criteria, next quarter's two spheres selected; owner: sponsor + working group. **Annual, full day** — portfolio audit across every significant innovation (W-05 at portfolio level); perception priorities written into annual planning and budgets; owner: executive team. A cadence that is not in the calendar is a wish — names in the boxes, invites sent, before the transformation team stands down.