PulseCheck (Build Day) — Merged Product Design Spec

Authors: Yannic Arnold, Senior Consultant & Daniel Jonas, Americas CEO, Campana & Schott Date: April 12, 2026 Status: Selected for Build Day · April 17, 2026 · Miami Build Day Team: Team 1 — Builder: Anuj Uppal · Client Advocate: Robin Boss · PO: Yannic Arnold · Members: Stephan Steiner, Gloria Xu, Shayanne Thomas, Nina Krause Lineage: Merges PulseCheck (Yannic Arnold, April 7) with ChangeSignal (Daniel Jonas, April 8). PulseCheck's ADKAR framework is the backbone; ChangeSignal's Teams-channel input and nudge-generation are absorbed as core features.

1. Problem Statement

Change management teams in pharma transformations sit on a gold mine of behavioral data they never systematically use. The data exists — in post-meeting debriefs, email threads, training attendance logs, executive communications, and hallway observations — but it's trapped in the consultant's head, a Teams channel, or a spreadsheet nobody trusts after month three.

1. ADKAR knowledge stays in the consultant's head and never enters the work systematically. A consultant trained in ADKAR applies it on the fly, based on instinct, between calls. There is no moment in the workflow where the framework surfaces and says: this situation, right now, is a Desire problem — not a Knowledge problem. The wrong interventions get deployed. Training runs for stakeholders who already understand the change but don't want it. Resistance compounds.

2. People fall through the cracks. A 30-stakeholder transformation means 30 relationships to maintain. The loud ones get attention. The quiet ones get forgotten. The workstream lead who stopped showing up three weeks ago? Nobody noticed until the SteerCo asked why that workstream stalled. There is no early warning system watching the signals between status reports.

3. The reporting tax eats senior consultant time. Every two weeks, someone spends 8–12 hours pulling adoption data from surveys, attendance logs, and their own memory to build a SteerCo report. At consulting rates, that's $4,000–$6,000 per cycle — north of $100K per year — on copy-paste work that produces a snapshot already stale by the time it's formatted.

4. The project sponsor is simultaneously the change champion and the hidden blocker. Directors and VPs who own transformation projects frequently say the right things in steering committees and do the wrong things in team meetings. Three different messages, one confused organization. The consultant knows this is happening. Surfacing it diplomatically, without making the sponsor defensive, is one of the hardest moves in change management — and it currently has no structure, no artifact, and no support.

5. Zero behavior change is required — but no tool captures what already happens. Change teams already debrief after every meeting. They already post updates in Teams. They already know who's bought in and who isn't. The problem isn't data collection — it's that none of this gets captured, aggregated, or acted on systematically. The profession that exists to manage human transitions through change has the worst tooling for managing the humans.

The deeper structural problem: change management is practiced as an episodic discipline — a survey here, a training there, an impact assessment at kickoff. But adoption is a continuous signal that lives in artifacts consultants already produce every week. The data exists. Nobody is reading it systematically. PulseCheck reads it.

2. Product Concept

One-Sentence Pitch

PulseCheck reads your team's post-meeting debriefs from a Teams channel, maps stakeholder signals to ADKAR stages, scores engagement health, generates personalized nudge messages for at-risk stakeholders, flags when the project sponsor contradicts the change narrative, and delivers a sanitized adoption view for the project sponsor — all from a single input.

How It Works

Step 1 — Teams-channel debrief (zero behavior change). The change team posts in a dedicated Teams channel after meetings, the way they already do: "Kickoff with Field Force SE went well. Kevin Walsh seemed checked out, left early. Rachel asked good questions about workflow integration. Soo was skeptical about timeline but didn't push back hard." PulseCheck's AI already has the stakeholder map (loaded once at setup). It parses that message and extracts structured signals per person automatically.

Step 2 — Optional 90-second observation debrief. Immediately after the channel post, PulseCheck prompts: "Anything you observed that isn't in the text?" The consultant picks the stakeholders who were present, writes one sentence of raw observation, and selects their read from a dropdown: Disengaging / Resistant / Uncertain / Positively surprised / Performatively aligned. Ninety seconds. The consultant's trained judgment enters the system alongside the text signals.

Step 3 — ADKAR inference + health scoring. PulseCheck's inference engine reads both the debrief text and the observation log. It maps language patterns and behavioral signals to ADKAR stages (Awareness → Desire → Knowledge → Ability → Reinforcement), assigns a confidence score per stakeholder, and quotes the specific evidence. Simultaneously, it computes a health score (Engaged / Watch / At Risk) based on recency, frequency trend, sentiment trend, and influence level. Both layers feed from the same data; they represent different altitudes of the same analysis.

Step 4 — Action Board (for the change team, daily). A kanban-style board showing all stakeholders in Engaged (green), Watch (amber), At Risk (red) columns. Each card shows name, role, ADKAR stage, last touchpoint, sentiment. Click any card to see the full engagement timeline and generate a personalized re-engagement nudge message. The change manager opens this every morning to answer "who needs my attention today?"

Step 5 — Sponsor View (for the project sponsor, on demand). A plain-language adoption dashboard. ADKAR stages named explicitly but immediately translated: "Maria Chen is at Desire — she understands the change but isn't convinced it's worth it yet." Risk levels color-coded. Recommended actions specific and non-technical. Designed for a Director or VP who has never taken a Prosci course.

Step 6 — Sponsor Mirror (coaching ammunition). A side-by-side view comparing what the agreed change narrative says with what the project sponsor actually communicated to specific audiences. Gap highlights surface inconsistencies as coaching preparation — never auto-distributed, always consultant-controlled. Framed as ammunition for a difficult conversation, not as a report card.

The Strategic Play

PulseCheck is deployed on every C&S change engagement from day one. The primary input is a Teams channel the team already uses. When C&S leaves, the client has two choices: lose the accumulated stakeholder intelligence — every relationship tracked, every nudge sent, every intervention logged — or license PulseCheck and keep the institutional memory. Every engagement becomes a land-and-expand motion.

The longer play: C&S licenses PulseCheck to clients as a standing capability across all their transformation programs. Every consulting firm working with that client reports adoption signals through PulseCheck. C&S built the standard. C&S collects the license revenue.

3. Target Buyer

Primary Buyer: The Project Sponsor (Director, Senior Director, or VP)

A business leader who owns a transformation — a product rollout, a system implementation, an operating model redesign — and views change management as something they know they need and are not sure they're getting. Their pain is simple: they do not know if their project is going to land until it's too late to course-correct.

What They Do Today

• Commission a pulse survey every 6–8 weeks. Wait for results. Act on findings already 6 weeks stale.
• Ask the change team to compile adoption reports manually (8–12 hours per cycle, ~$5K per report). Track stakeholder engagement in spreadsheets nobody trusts by month three.
• Rely on the consultant's verbal update in the weekly steering committee. Take it on faith.
• Find out resistance has hardened when a go-live slips or a business unit refuses to adopt.

The cost: one delayed go-live on a mid-size transformation typically costs $200k–$500k in extended consulting fees, lost productivity, and rework.

Why They Would Pay

• Real-time adoption visibility replaces biweekly guesswork. Resistance gets flagged weeks earlier.
• A structured view of where resistance is building — by name, by team, by ADKAR stage — before it becomes a crisis.
• Reporting labor drops 80%. Nudge generation saves 3–5 hours per week on outreach.
• A plain-language dashboard they can forward to their own leadership as proof that change management is being managed, not just performed.
• Total labor replaced: ~$8K/month.

Secondary Beneficiary: The Change Management Consultant

PulseCheck makes the consultant dramatically more effective. They walk into every stakeholder conversation knowing exactly where that person is on the ADKAR curve. They stop deploying training for people with a Desire problem. They have a structured, evidence-backed artifact when they need to coach a sponsor about their own behavior. They stop building spreadsheets and start having conversations.

4. Architecture

Input Layer

Primary input: Teams-channel debrief (from ChangeSignal). One channel. Change managers post natural language debriefs after meetings, touchpoints, and informal conversations. The AI parses names, sentiment, ADKAR signals, and topics against the pre-loaded stakeholder map. One channel post = multiple interaction records logged automatically. Zero behavior change — "post in the channel like you already do."

Secondary input: 90-second observation debrief (from PulseCheck). Triggered optionally after each channel post. Three fields: who was present (pre-populated from roster), what you observed (one sentence), and your read (dropdown). Captures what transcripts cannot — non-verbal cues, body language, what was not said.

Hackathon scope: Paste/type input only. No live Teams API integration. The value is in the inference, not the plumbing.

Transform Layer (Core IP)

A single Claude API call with a combined system prompt, low temperature (0.2) for consistency, structured JSON output. The prompt takes three RAG context inputs and one user input:

RAG Context (loaded at engagement setup):

• Stakeholder map — 12-15 named stakeholders with role, department, influence level, workstream
• Signal library — practitioner-built taxonomy mapping behavioral signals to ADKAR stages. Hedging language → Desire. Silence → Awareness gap. Coalition-building → resistance forming. This is the core IP — built from real change management expertise, not generic NLP.
• Change narrative — the agreed transformation story. What's changing, why, what success looks like. Required for the Sponsor Mirror.

User Input: The Teams debrief text + optional observation debrief

Output (structured JSON per stakeholder mentioned):

• name — resolved against the stakeholder map
• adkar_stage — Awareness / Desire / Knowledge / Ability / Reinforcement
• adkar_evidence — the specific text that drove the inference (quoted)
• adkar_confidence — 0.0 to 1.0
• health_status — Engaged / Watch / At Risk
• health_trend — Improving / Stable / Declining
• recommended_action — one specific sentence
• nudge_draft — a personalized re-engagement message (only for Watch / At Risk)

Why one API call, not three: Collapsing extraction, ADKAR mapping, health scoring, and nudge generation into a single well-crafted prompt is faster to build, easier to debug, and produces more coherent output than a multi-pass pipeline. The prompt does the heavy lifting; the code just renders the JSON.

Nudge generation (separate call): When the user clicks an at-risk stakeholder on the Action Board, a second Claude call generates a longer, more personalized nudge message referencing the stakeholder's role, last interaction, known concerns, and workstream status. This is the demo's emotional peak — it must feel specific enough that a judge says "I'd actually send that."

Sponsor Mirror (pre-computed): A dedicated prompt ingests the change narrative and all uploaded sponsor communications, then surfaces gaps — what the story says vs. what the sponsor actually said to a specific audience. For the hackathon, this runs at build time against pre-loaded sponsor communications, not live during the demo.

Display Layer — Five Surfaces

1. Action Board (from ChangeSignal) — Kanban of stakeholders in green/amber/red. Click for detail + nudge. The hero UI. Live inference.
2. Nudge Generation (from ChangeSignal) — Click at-risk stakeholder → Claude generates a personalized message. Live inference. The emotional peak.
3. Sponsor View (from PulseCheck) — Plain-English ADKAR per stakeholder, color-coded risk, recommended actions. Live rendering of the same parsed data.
4. Sponsor Mirror (from PulseCheck) — Side-by-side narrative vs. sponsor comms with gap highlights. Scaffolded — pre-computed at build time.
5. Executive Report (from ChangeSignal) — Biweekly adoption summary with heatmaps and flags. Scaffolded — pre-generated from sample data.

Tech Stack (Hackathon)

5. Hackathon Execution Plan

Timeline (mapped to Build Day agenda)

Critical Risk and Fallback

Biggest risk: The combined ADKAR + health scoring prompt doesn't produce clean, structured output consistently. Names get misidentified, ADKAR stages feel random, nudges are generic.

Fallback (decide at 2:00 PM — 30 min into Sprint 2): If live parsing isn't reliable:

• Lock input to the pre-loaded sample data (8-10 pre-parsed debriefs)
• All 5 surfaces still render from the pre-parsed data
• Nudge generation still fires live (it's a simpler, separate Claude call — lower risk)
• Demo shifts from "watch it parse live" to "here's the output from real data + watch it generate a nudge live"
• The pitch narrative barely changes. The audience still sees the full vision across all 5 surfaces.

Second fallback (decide at 2:30 PM): If nudge generation is also unreliable:

• Pre-generate 3 nudge messages during Sprint 2 and hard-code them
• The Action Board still renders, the Sponsor View still renders, the nudges are pre-written but realistic
• The pitch becomes "here's what the system produces" — a polished static demo of the full vision
• Judges evaluate the product concept and the business case, not whether the API fires on stage

Role Allocation

6. Defensibility & Competitive Moat

Why a Client Would Pay — and Keep Paying

1. The engagement history is the moat. After 6 months, PulseCheck contains every stakeholder relationship — who championed, who resisted, what interventions worked, how ADKAR profiles evolved. That longitudinal model is irreplaceable. Walking away means losing institutional memory that took months to build.

2. The signal library improves with use. The practitioner-built ADKAR taxonomy gets more accurate and more specific to each client's language patterns over time. After six months, PulseCheck knows that "we'll take that on notice" in this organization means resistance, not agreement. That specificity is not replicable by a generic tool.

3. The Sponsor Mirror creates an irreplaceable coaching artifact. There is no other tool that helps a consultant surface sponsor message inconsistency with structured, evidence-backed diplomatic framing. Once a client has experienced a coaching conversation backed by the Sponsor Mirror, the alternative — gut feel and political judgment alone — feels inadequate.

4. Cross-program intelligence. A VP in three programs has one unified ADKAR profile. The change team on Program B sees they were a late adopter on Program A. This cross-program view is unique to PulseCheck and becomes more valuable with every program added.

5. Zero behavior change to adopt. The input is a Teams post. The team already does this. The adoption barrier is effectively zero, which means the switching cost is entirely on the output side — the intelligence you lose, not the workflow you have to unlearn.

Pricing Model

Anchor to the cost it replaces, not the cost to build.

ROI math: a consultant spending 5 hours/week on manual change signal-gathering at $250/hour costs the client $5,000/month in labor. PulseCheck at $1,500–$2,500/month makes that time targeted and effective. One prevented go-live delay ($200k–$500k) justifies the entire engagement-length subscription.

Competitive Positioning

Closest analogue: "Gainsight for change management — but fed by a Teams channel instead of product telemetry, grounded in ADKAR instead of health scores alone, with a Sponsor Mirror that no competitor has attempted." Nothing like this exists.

7. Success Criteria

Hackathon Demo (April 17)

• Teams-channel debrief pasted into input box returns structured ADKAR + health inference within 10 seconds
• Action Board renders stakeholders in green / amber / red with ADKAR stage visible per card
• Clicking an at-risk stakeholder generates a personalized, context-aware nudge message that a judge would actually send
• Sponsor View renders plain-English ADKAR stages with color-coded risk and specific recommended actions
• Sponsor Mirror shows at least one specific gap between the change narrative and a sponsor communication
• Executive Report renders a formatted adoption summary (can be pre-generated from sample data)
• All 5 surfaces are navigable and populated — live or scaffolded — telling a complete product story
• Judges can interact with the live URL during the pitch
• Demo uses realistic pharma transformation data with identifiable patterns

"Would a Client Pay for This?" Test

1. A Director or VP with no change management background looks at the Sponsor View and immediately understands what it's telling them and what they should do next.
2. The inference output is visibly non-obvious — the AI is detecting signals a casual reader would miss, not just summarizing what's already explicit.
3. The nudge messages feel human — specific enough to send, not generic enough to ignore. At least one judge says "I'd actually send that."
4. The Sponsor Mirror moment lands: a judge who has lived a transformation recognizes the political situation and understands why the consultant needs that artifact.
5. The "no CM tool exists for this" gap is articulated clearly in the pitch and the competitive slide.
6. The pricing math is presented and the ROI is self-evident without explanation.

8. Open Questions

1. How good can the combined ADKAR + health scoring prompt get in 3 hours? The product depends on one well-crafted system prompt producing clean, structured JSON with grounded ADKAR stages. 90%+ accuracy = demo hit. 70% = use the fallback. The prompt engineering in Sprint 1 is the single highest-leverage hour of the day.

2. Should the observation debrief be in the hackathon MVP or cut for scope? It adds the consultant's trained judgment (body language, tone, what wasn't said) but it's a second input surface to build. If Sprint 1 runs long, cutting the observation debrief preserves the core loop while losing a differentiated layer.

3. What's the right level of ADKAR granularity for the demo? All five stages (A-D-K-A-R) or a simplified three-stage model (Awareness / Commitment / Ability)? Full ADKAR is more intellectually rigorous and matches the Prosci methodology Yannic knows. Simplified is easier to demo and easier for the prompt to produce consistently.

4. How do we handle the Sponsor Mirror without making the sponsor the villain? The spec says "coaching ammunition, not a report card" — but the demo still shows a sponsor saying inconsistent things. The framing matters. If a judge reads it as surveillance, the product loses trust. The Narrator needs a specific line that positions this as coaching, not catching.

5. Can we use Build Day meeting transcripts as live input during the demos? If all teams record and upload transcripts every hour, Team 1 could feed those transcripts into PulseCheck as a meta-demo — analyzing the Build Day's own change dynamics. High risk, high reward. Depends on the recording infrastructure being in place.