why-change

Why Change Pitch

Create a "Why Change" sales pitch using the Corporate Visions methodology. Supports two modes:

• Customer mode — Deal-specific pitch for a named customer. Company-specific web research, personalized framing.
• Segment mode — Reusable pitch for a market segment. Industry-level web research, generic framing that works for any organization in the segment.

Both modes produce two deliverables: sales-presentation.md (narrative arc) and sales-proposal.md (formal proposal with pricing).

Prerequisites

• At least one cogni-portfolio project with products, features, propositions, and solutions
• Customer mode: a named customer to pitch to (company name + industry)
• Segment mode: a market segment from the portfolio (e.g., "Enterprise Manufacturing DACH")
• Optional: cogni-trends project with value-modeler completed (enriches with trend evidence)

Arc Methodology

This skill applies the Corporate Visions story arc defined in cogni-narrative:

• cogni-narrative/skills/narrative/references/story-arc/corporate-visions/arc-definition.md

The arc has four elements, each with detailed patterns the researcher agent reads and applies:

1. Why Change — Disrupt status quo with unconsidered needs (problem-solution-benefit structure)
2. Why Now — Create urgency with forcing functions and cost of inaction
3. Why You — Differentiate with IS/DOES/MEANS (solution → outcomes → moat)
4. Why Pay — Build business case with compound cost calculation

Workflow

Phase 0: Setup

Step 0.1: Portfolio Discovery

Run the discovery script to find available portfolios:

"${CLAUDE_PLUGIN_ROOT}/scripts/discover-portfolio.sh" --workspace "$(pwd)"

Present discovered portfolios to the user:

Portfolio projects found:
  1. acme-cloud (4 products, 12 features, 3 markets, 8 propositions)

If only one portfolio: confirm automatically. If multiple: ask user to select.

Step 0.2: Pitch Mode & Target

Ask the user via AskUserQuestion:

"Are you creating a pitch for a named customer or a reusable segment pitch for a market?"

Customer mode:

1. Customer name (required): "Which customer are you pitching to?"
2. Customer industry (required): "What industry is {customer} in?"
3. Customer domain (optional): "Do you have their website domain? (helps with research)"

Segment mode:

1. Read available markets from the portfolio (markets/*.json)
2. Present markets to the user: "Which market segment?"
3. Set segment_name from the selected market's display name
4. Set customer_industry from the market's industry field
5. Skip customer_domain (not applicable)

Step 0.3: Market Matching

Customer mode: Read all markets from the selected portfolio (markets/*.json). Match customer_industry to market segments by vertical/industry alignment.

• If a single market matches: auto-select
• If multiple match: present options and ask user
• If none match: warn user that propositions may be generic, proceed with best-fit

Segment mode: Market was already selected in Step 0.2 — skip this step.

Set market_slug in pitch-log.json.

Step 0.4: TIPS Discovery (Optional)

Glob for **/tips-value-model.json in the workspace. If found:

• Check if the TIPS industry aligns with the customer's industry
• If aligned: set tips_path in pitch-log.json (enables enriched evidence in all phases)
• If not aligned or not found: proceed in portfolio-only mode (tips_path: null)

Inform the user: "Found TIPS project for {industry} — will enrich with strategic trend evidence" or "No TIPS data found — proceeding with portfolio data only."

Step 0.5: Solution Focus

Read products and features from the portfolio. Ask the user:

"Would you like to pitch the full portfolio or focus on specific products/features?"

• Full portfolio: leave solution_focus empty
• Specific features: user selects from list, store slugs in solution_focus

Step 0.6: Buyer Roles

Customer mode: Ask the user for key buyer roles:

• "Who is the economic buyer? (title)" — e.g., CIO, CFO, VP Operations
• "Who evaluates technically? (title)" — e.g., Cloud Architect, Head of IT

Segment mode: Load buyer personas from the portfolio's customer profiles for this market (customers/{market}.json). If buyer personas exist, use them as defaults. Otherwise ask for typical titles in this segment.

Store titles in buying_center in pitch-log.json.

Step 0.7: Language

Check for .workspace-config.json in the working directory. If it has a language field, use it. Otherwise ask:

"Should the pitch be in English or German?"

Step 0.8: Initialize Project

Run the initialization script:

Customer mode:

"${CLAUDE_PLUGIN_ROOT}/scripts/init-pitch-project.sh" \
  --customer-name "{customer_name}" \
  --pitch-mode customer \
  --language "{language}" \
  --workspace "$(pwd)"

Segment mode:

"${CLAUDE_PLUGIN_ROOT}/scripts/init-pitch-project.sh" \
  --segment-name "{segment_name}" \
  --pitch-mode segment \
  --language "{language}" \
  --workspace "$(pwd)"

Then update pitch-log.json with all collected context (customer_domain, customer_industry, market_slug, portfolio_path, tips_path, company_name, solution_focus, buying_center).

-> Proceed immediately to Phase 1.

---

Phase 1: Why Change

Delegate to the why-change-researcher agent:

Agent tool:
  agent: why-change-researcher
  prompt: |
    project_path: {project_path}
    phase: why-change

The agent will:

• Read why-change-patterns.md from cogni-narrative (PSB structure, contrast patterns)
• Load portfolio propositions (IS/DOES/MEANS for matched market)
• "Work backwards" from portfolio capabilities to problems
• Perform web research — company-specific (customer mode) or industry-level (segment mode)
• Write 01-why-change/research.json and 01-why-change/narrative.md

Quality Gate: Present the key findings to the user:

Phase 1: Why Change — Key Findings for {target}

Strategic themes driving this pitch:
1. [theme_name] — [why_change_angle] (source: tips|portfolio)
2. [theme_name] — [why_change_angle] (source: tips|portfolio)
3. ...

Unconsidered needs identified:
1. [headline] — [one-line summary with key evidence]
2. [headline] — [one-line summary]
3. [headline] — [one-line summary]

Claims registered: N (for verification)

Approve to continue, or provide feedback to revise.

The strategic themes come from theme-brief.json generated during Phase 2.5 of the researcher agent. Surfacing them here lets the user course-correct theme selection before subsequent phases build on these themes.

Use AskUserQuestion with options: "Approve and continue", "Revise with feedback".

If revise: re-invoke the researcher with user feedback appended to the prompt. If approve: update pitch-log.json (phases_completed += ["why-change"], current_phase = "why-now").

-> Proceed immediately to Phase 2.

---

Phase 2: Why Now

Delegate to the why-change-researcher agent:

Agent tool:
  agent: why-change-researcher
  prompt: |
    project_path: {project_path}
    phase: why-now

The agent will:

• Read why-now-patterns.md from cogni-narrative (forcing functions, urgency quantification)
• Load Phase 1 bridge file for context
• Research timing triggers — company-specific (customer mode) or industry-level (segment mode)
• Write 02-why-now/research.json and 02-why-now/narrative.md

Quality Gate: Present timing triggers and cost of inaction summary. Same approve/revise pattern.

Update pitch-log.json. -> Proceed to Phase 3.

---

Phase 3: Why You

Delegate to the why-change-researcher agent:

Agent tool:
  agent: why-change-researcher
  prompt: |
    project_path: {project_path}
    phase: why-you

The agent will:

• Read why-you-patterns.md from cogni-narrative (differentiators with IS-DOES-MEANS)
• Load portfolio propositions, solutions, and competitor data
• Start each IS cell from the portfolio solution entity's capability description, adapted to buyer context
• Create 2-3 differentiators mapped to buyer needs from Phase 1
• Write 03-why-you/research.json and 03-why-you/narrative.md

IS/DOES/MEANS semantics (critical): IS describes YOUR SOLUTION or capability ("Eine integrierte OT/IT-Sicherheitsplattform..."), NOT the customer's problem. The researcher generates IS cells by starting from solutions/{feature}--{market}.json capability descriptions and adapting them to the buyer's context — this prevents problem-language from leaking into IS cells. DOES states what the solution does for the buyer (outcomes with numbers, You-Phrasing). MEANS explains why competitors can't replicate it (moat: time, experience, network, certification).

Quality Gate: Present the differentiators table with IS/DOES/MEANS. Same approve/revise pattern as other phases.

Update pitch-log.json. -> Proceed to Phase 4.

---

Phase 4: Why Pay

Delegate to the why-change-researcher agent:

Agent tool:
  agent: why-change-researcher
  prompt: |
    project_path: {project_path}
    phase: why-pay

The agent will:

• Read why-pay-patterns.md from cogni-narrative (compound cost calculation)
• Load Phase 2 cost-of-inaction data + Phase 3 capability outcomes
• Load solution pricing tiers from portfolio
• Build ROI model: cost of inaction vs investment
• Write 04-why-pay/research.json and 04-why-pay/narrative.md

Quality Gate: Present business case summary with investment vs inaction ratio. Same approve/revise pattern.

Update pitch-log.json. -> Proceed to Phase 5.

---

Phase 5: Synthesize

Delegate to the pitch-synthesizer agent:

Agent tool:
  agent: pitch-synthesizer
  prompt: |
    project_path: {project_path}

The agent will:

• Read all 4 phase bridge files and narratives
• Read output-specs.md templates (selecting customer or segment variant)
• Assemble output/sales-presentation.md and output/sales-proposal.md
• Renumber citations sequentially

Update pitch-log.json (phases_completed += ["synthesize"], current_phase = "review").

-> Proceed immediately to Phase 5.5.

---

Phase 5.5: Stakeholder Review (Closed Loop)

After the pitch-synthesizer produces deliverables, run a closed assess-revise-reassess loop using the pitch-review-assessor and pitch-revisor agents. This catches failures that per-phase quality gates miss — wrong buyer needs, incredible claims, thin differentiation, weak business case — before the pitch reaches the customer.

Skip condition: If the user explicitly said "skip review" during setup, skip this phase and proceed directly to Completion.

Step 5.5.1: First Assessment

Delegate to the pitch-review-assessor agent:

Agent tool:
  agent: pitch-review-assessor
  prompt: |
    project_path: {project_path}

The agent reads both deliverables, pitch-log.json, claims.json, and portfolio data. It evaluates from three perspectives (target buyer, sales director, marketing director) and writes the assessment to output/pitch-review.json.

Present results to the user:

Phase 5.5: Stakeholder Review — {target}

| Perspective | Score | Verdict |
|-------------|-------|---------|
| Target Buyer ({buyer_role}) | {score}/100 | {pass/warn/fail} |
| Sales Director | {score}/100 | {pass/warn/fail} |
| Marketing Director | {score}/100 | {pass/warn/fail} |

Overall: {accept/revise/reject} ({overall_score}/100)

{If critical_improvements exist: "Critical issues: ..." }
{If high_improvements exist: "Improvements identified: N high-priority" }

Verdict handling:

• accept (all perspectives 85+): Proceed directly to Completion. The pitch is ready.
• revise (all 70+ but not all 85+): Proceed to Step 5.5.2 (Revision).
• reject (any perspective below 50): Present the assessment to the user with full details. Recommend re-running the content phase that corresponds to the lowest-scoring section (e.g., if Why Change scored lowest on Need Recognition, suggest re-running Phase 1). Use AskUserQuestion: "The pitch has fundamental issues. Would you like to re-run [phase] with revised guidance, or proceed as-is?"

Step 5.5.2: Revision (if verdict = "revise")

1. Archive the assessment: copy output/pitch-review.json to output/pitch-review-v1.json

2. Delegate to the pitch-revisor agent:

Agent tool:
  agent: pitch-revisor
  prompt: |
    project_path: {project_path}
    assessment_path: {project_path}/output/pitch-review-v1.json

The revisor applies surgical fixes to both deliverables, prioritized by severity (CRITICAL → HIGH). It writes output/revision-log.json with details of what was changed and preserved.

3. After revision completes, re-run the assessor:

Agent tool:
  agent: pitch-review-assessor
  prompt: |
    project_path: {project_path}

4. Archive the second assessment as output/pitch-review-v2.json

5. Present comparison to the user:

Revision Results (Pass 2):

| Perspective | Before | After | Delta |
|-------------|--------|-------|-------|
| Target Buyer ({buyer_role}) | {v1_score} | {v2_score} | {delta} |
| Sales Director | {v1_score} | {v2_score} | {delta} |
| Marketing Director | {v1_score} | {v2_score} | {delta} |

Overall: {v2_verdict} ({v2_overall_score}/100, was {v1_overall_score}/100)

Fixes applied: {N} | Sections modified: {list}

Maximum 2 revision passes. If the second assessment is still "revise", present the scores and proceed to Completion. Do not loop further — diminishing returns and oscillation risk increase with each pass. The user can run /copywrite for additional polish.

Update pitch-log.json (phases_completed += ["review"], current_phase = "done").

-> Proceed to Completion.

---

Completion

Present the deliverables to the user:

Customer mode:

Pitch complete for {customer_name}!

Deliverables:
  - {project_path}/output/sales-presentation.md (Why Change narrative)
  - {project_path}/output/sales-proposal.md (formal proposal with pricing)
  - {project_path}/output/pitch-review.json (stakeholder assessment)

Stakeholder Scores: Buyer {N}/100 | Sales {N}/100 | Marketing {N}/100
Claims registered: {N} — run `/claims verify` to validate sources.

Optional next steps:
  - `/copywrite sales-presentation.md` — polish with cogni-copywriting
  - `/pptx create sales-presentation.md` — generate slide deck via cogni-visual

Segment mode:

Segment pitch complete for {segment_name}!

Deliverables:
  - {project_path}/output/sales-presentation.md (Why Change narrative — reusable)
  - {project_path}/output/sales-proposal.md (segment proposal template — reusable)
  - {project_path}/output/pitch-review.json (stakeholder assessment)

Stakeholder Scores: Buyer {N}/100 | Sales {N}/100 | Marketing {N}/100
Claims registered: {N} — run `/claims verify` to validate sources.

These deliverables serve as reusable templates for any customer in the {segment_name} segment.
To create a customer-specific pitch based on this template, run `/why-change` in customer mode.

Optional next steps:
  - `/copywrite sales-presentation.md` — polish with cogni-copywriting
  - `/pptx create sales-presentation.md` — generate slide deck via cogni-visual

---

Resuming an Interrupted Pitch

If the user wants to continue a pitch that was interrupted:

1. Run pitch-status.sh on the project path
2. Read workflow_state.current_phase and phases_completed
3. Resume from the incomplete phase

Error Handling

• If portfolio has no propositions for the matched market: warn user, suggest running cogni-portfolio propositions skill first
• If portfolio has no solutions for the matched market: warn user, proceed with industry-benchmark pricing in Why Pay. Log "pricing_source": "industry_benchmark" in pitch-log.json
• If portfolio has no competitors for the matched market: warn user, proceed with web-researched competitive landscape. Log "competitor_source": "web_research" in pitch-log.json
• If web research returns thin results: customer mode degrades to industry-level research; segment mode uses broader search terms
• If cogni-narrative arc files not found: error — cogni-narrative plugin must be installed