The Pitch — Why This Matters

Outline

1. Learning objectives — 60s CEO pitch, 3min CTO pitch, 5min security review, objection counters
2. Key concept — "PM33 turns AI development from a productivity tool into a closed-loop strategic execution platform"
3. Diagram walkthrough — slide 06: 6-row pain-vs-answer pitch table + CTA box
4. The 60-second CEO pitch — script + the 4 beats it hits
5. The 3-minute CTO pitch — script + substance covered (Brief schema, execution architecture, closed loop, compliance-first, lock-in)
6. The 5-minute security review — module 5 walkthrough mapped to compliance reviewer's questions
7. The three most common objections — counters for "we already use X", "we're nervous about AI agents", "how do we know your model works for our domain"
8. Pitch failure modes — anti-patterns to avoid (lead with AI agents, promise dates, oversell autonomy, "factory" jargon, pitch features not outcomes)
9. Pitch winning moments — lean into these (audit demo, recalibration story, bypass-as-signal flip, compounding advantage)
10. Closing the deal — 3 concrete asks (pilot scope, reference customer, technical eval)
11. Further reading — modules 1-5, GTM doc, marketing workspace, staging demo env

Learning objectives

After this module you should be able to:

• Pitch PM33 to a CEO in 60 seconds
• Pitch PM33 to a CTO in 3 minutes with technical depth
• Pitch PM33 to a security reviewer in 5 minutes covering compliance
• Counter the three most common objections

Key concept

The pitch is NOT "we use AI to write code faster." Every product in this space says that. The pitch is:

PM33 turns AI development from a productivity tool into a closed-loop strategic execution platform.

You ship code (AI does that). PM33 measures whether the shipped thing moved the metric you intended. The feedback updates your priorities. The next sprint is smarter than the last. That's the closed loop. That's what nobody else has.

Diagram walkthrough

A three-column comparison table, visualized:

Today's pain (left column, red icons) → PM33 answer (right column, green checkmarks):

End-of-slide CTA: "PM33 turns AI development from a productivity tool into a closed-loop strategic execution platform."

The 60-second CEO pitch

"Your engineers ship code. Some of it moves the business. Some doesn't. Today you find out 6 months later — if at all. PM33 closes the loop: every shipped change gets attributed back to the strategic objective it was supposed to advance. The AI agents that wrote the code? Their work is now measured against business outcomes, not lines-of-code metrics. Your next sprint's priorities are smarter because last sprint's results trained the model. That's strategy-to-outcome in one system. That's why your competitors who adopt this 12 months ahead of you will pull ahead permanently."

Hits 4 beats: pain (we don't know what worked), differentiation (closed loop), competitive urgency (first-mover advantage), specificity (last sprint trained the model).

The 3-minute CTO pitch

"PM33 sits on top of your existing stack — GitHub, Jira, Linear, whatever — and adds three things:

One: an atomic work-item format called a Brief, with machine-verifiable acceptance criteria, a TDD plan, and an outcomeHook field. Briefs are designed for AI agents to execute end-to-end. The agent doesn't have to guess what 'done' means.

Two: a coordinator/specialist execution pattern with per-agent git worktrees and per-agent git indexes. This eliminates the cross-agent commit absorption class of bugs we hit 11 times before structurally fixing it. You can run 5+ parallel agents in the same repo without them stepping on each other.

Three: a closed-loop outcome attribution system. Every Brief's outcomeHook fires after deploy. The AR(1) forecast model recalibrates from the actual data. Your sprint planner gets smarter every sprint. After 100 Briefs in a workspace, you have a workspace-specific velocity prior that's far more accurate than any industry benchmark.

The audit trail is structured from day 1 — not bolted on for SOC2. Every state transition is a lifecycle event. Every policy bypass is observable. Tenant isolation is enforced at the DB via RLS, not at the application layer.

The compounding advantage: the longer you run PM33, the more it knows about your workspace's actual delivery dynamics. The platform becomes harder to leave because it knows things your old tools never measured."

Hits the substance: Brief schema, execution architecture, outcome closed loop, compliance-first, lock-in mechanism.

The 5-minute security review

Open Module 5 (Governance & Trust). Walk the security reviewer through:

1. Tenant isolation at the DB via RLS (not app-layer) — show the policy on work_items
2. 15-role RBAC + Zod schema as single source of truth — show shared/roles.ts
3. Append-only audit log + 7-year default retention — run a pm33_query_audit_log query live
4. Bypass tracking turns policy violations into observable signals (not silent failures) — show a policy_bypass event
5. SOC2 alignment table from Module 5

The reviewer's job is to find the gaps. Make it easy for them to do that. Hand them read-only audit log access. Let them poke. The product is designed to survive scrutiny.

The three most common objections

"We already use Jira / Linear / Asana"

Answer: PM33 sits on top. Bi-directional sync. We don't ask you to migrate work items. We add the outcome attribution + Brief execution layer on top of what you already have. Your tools stay. The Briefs flow through them and back.

"We're concerned about AI agents writing code without oversight"

Answer: The platform is designed for oversight, not autonomy. Every state transition is an audit event. Bypasses are observable. Independent code review is mandatory before done. Outcome attribution measures whether the AI's work actually moved the metric. You have more visibility into AI-driven development than into human-only development today.

"How do we know your model will work for our domain?"

Answer: It starts with industry defaults. It learns your workspace's actual dynamics. After 30-100 Briefs, it has a workspace-specific prior that beats any vendor-baked-in model. Specifically: the AR(1) recalibration uses Bayesian shrinkage toward the workspace's median, not toward our defaults. You're not betting on our model's general fit — you're betting on its ability to learn yours.

Pitch failure modes (avoid these)

❌ Don't lead with "AI agents." The buyer has heard it 50 times this quarter. Lead with the closed loop.

❌ Don't promise dates. Tools that say "deploy in a week" lose credibility when implementation takes 6 weeks. Say "typical pilot takes X-Y weeks; depends on integration complexity."

❌ Don't oversell autonomy. Buyers fear AI agents going rogue more than they fear slow shipping. Emphasize human-in-the-loop where it exists (the 3 default human gates from Module 2).

❌ Don't use the word "factory." This is internal jargon from a reference architecture. Customers don't relate. Say "strategic execution platform."

❌ Don't pitch features. Pitch outcomes. The Brief schema is a feature. "Knowing which shipped code moved the metric" is an outcome.

Pitch winning moments (lean into these)

✅ The audit trail demo. Open the audit log. Type a few queries. Show the structured events. Pause. Let them feel the "oh, we'd actually have answers to compliance questions" reaction.

✅ The recalibration story. Show a real workspace whose AR(1) σ shrank from 4.0 → 1.4 over 5 months. Quantify what tighter predictions enable (better quarterly planning, fewer surprises).

✅ The bypass-as-signal flip. Most buyers expect tools to BLOCK bypasses. Explain why observation > blocking. Show the dashboard. The framing flip is often the moment they realize PM33 was designed by people who actually shipped software.

✅ The compounding advantage. "The longer you run this, the more it knows about your workspace. The cost of switching back grows." Buyers love lock-in stories when the lock-in is value-aligned.

Closing the deal

Three concrete asks at end of pitch:

1. The pilot scope: "Pick one strategic objective. Run PM33 against it for 6 weeks. Compare the attribution data to what you'd otherwise have." Specific, time-bounded, low risk.

2. The reference customer: "Talk to [existing customer in their industry]. They'll tell you what worked and what didn't."

3. The technical eval: "Have your CTO + security reviewer spend 2 hours on Modules 3 + 5 of our curriculum. If those modules don't satisfy your technical bar, we don't have the conversation."

Confidence in the asks comes from confidence in the product. The curriculum exists to make that confidence justified.

Further reading

• All preceding modules (1-5) — the technical depth that backs the pitch
• docs/GTM_POSITIONING_2026.md — current GTM positioning doc
• docs/marketing/ — the marketing content workspace (this is part of it!)
• The Pam staging deployment — a live demo environment you can show