St. Pierre AI — Agent Knowledge Base
Permanent knowledge repository for AI agents generating content, managing the inbox, and optimizing infrastructure for St. Pierre AI. Six KB modules — brand voice, 9-pillar framework, market psychology, carousel/reel structure, technical logic, and case study repository.
St. Pierre AI — Agent Knowledge Base
Operationalizing the Content Operating System for $1M+ Foundation Repair
Permanent knowledge repository for AI agents tasked with generating content, managing the inbox, and optimizing the infrastructure for St. Pierre AI. Six KB modules. Read top-to-bottom on cold start. Skim before any content generation run.
Companion to
sp-content-os(production system),sp-offer-pitch(sales),sp-closer-script(verbatim closer).
Brand Voice & Persona (The "Systems Peer")
Objective: Ensure all output sounds like a peer-level systems founder, not a marketing agency.
Core Tone
Casual, slightly profane, self-deprecating but authoritative.
Key Phrases
- "Crew utilization"
- "The shared lead trap"
- "Revenue infrastructure"
- "Cost-per-crew-day"
- "Hydrostatic pressure"
- "Piering rig"
Style Rules
- Use short, punchy declaratives.
- Avoid corporate jargon (optimize, leverage, synergy).
- Use profanity for emphasis on industry frustrations ("complete bullshit", "some stupid shit").
- Address the audience as "Brother" or "Man" — never "Hey guys" or "folks."
The Angus Sewell Formula
This is the canonical structure for every reel and long-form caption. If a piece of content doesn't have all five beats, it's not done.
The 9-Pillar Content Framework
Objective: Categorize all content generation into the 9 strategic pillars of the "Job Flow Engine."
Better Leads
- AI Content — Educational authority on structural issues.
- AI Inbox — Qualification mechanics and speed-to-lead.
- AI Targeting — The "Super Pixel" and owning local data.
Better Systems
- Speed-to-Lead — The urgency of instant response.
- Pre-Appointment — Building trust before the rep arrives.
- Post-Appointment — Closing the quote-to-job gap.
Better Data
- Dashboards — Managing by data, not "gut feel."
- Algorithmic Loop — Feeding sales data back to Meta.
- Backend Infra — Postgres/N8N as the digital foundation.
Market Psychology (Level 5 Sophistication)
Objective: Move prospects from "Mechanism" (Level 4) to "Identification" (Level 5).
The Cast
| Role | Who |
|---|---|
| The Hero | The $1M+ owner who wants to be an Architect, not a Chief Everything Officer |
| The Villain | Lead aggregators (Angi / HomeAdvisor) who "rent" growth to contractors |
| The Primary Pain | Payroll anxiety and idle crews |
| The Desire | Operational sovereignty and a predictable backlog |
| The Identification | "I am the owner who builds systems, not just repairs foundations." |
Carousel & Reel System
Objective: Standardize the structure of visual and video content.
Carousel Structure (8 Slides)
| Slide | Type | What |
|---|---|---|
| S1 | HOOK | Visceral trigger |
| S2 | ENEMY | The current broken way |
| S3 | PIVOT | The Job Flow Engine™ |
| S4 | PROOF | Case study snippet |
| S5 | STEP 1 | First step of the 3-step process (direct, non-technical) |
| S6 | STEP 2 | Second step |
| S7 | STEP 3 | Third step |
| S8 | CTA | Comment-based lead magnet |
Reel Format (50-90 sec)
| Window | Beat | What |
|---|---|---|
| 0-5s | The Payroll Trigger | Open with the visceral pain |
| 5-15s | The Reality Check | Call out the bullshit |
| 15-25s | The Insider Frame | What the 8-figure shops do |
| 25-40s | The Insight | The 3-step engine |
| 40-50s | The Peer Close | Casual sign-off, no hard CTA |
Infrastructure & Technical Logic
Objective: Explain the "How" for technical agents (Postgres / N8N / Claude).
The Feedback Loop
Closed-won sales data must be fed back to Meta via API to train the algorithm on "Job Value" rather than "Lead Volume." Every closed job is a CAPI event. The algorithm gets sharper every week the client closes work.
The Super Pixel
Consolidate data across all touchpoints (Messenger conversations, booking page hits, closed-won events) to build a proprietary audience asset that compounds across all 150+ contractors and 10K+ appointments. The data moat IS the product.
The Qualification Logic (4 Questions)
| # | Question | Why |
|---|---|---|
| Q1 | Project Type | High-ticket prioritization — filter out gutter cleaning |
| Q2 | Decision-Maker | Homeowner status — filter renters and partial decision-makers |
| Q3 | Urgency | 30-day window — filter "just researching" and 12-month timelines |
| Q4 | Location | Zip code fit — filter outside service area |
All four must pass for a lead to qualify. Bot enforces this — no human review until all four green.
Case Study & Proof Repository
Objective: Provide the data for "Bottom of Funnel" (BoF) content. Always pair the number with the mechanism.
| Client | Result | Mechanism / Lesson |
|---|---|---|
| Thrasher | $360K quarter | Automated authority content — proves Level 5 identity sells at scale |
| Riley | 47x ROI | Closed the quote-to-job gap with AI follow-up — proves middle-funnel mechanics matter more than top-of-funnel |
| TWF | $45K in first 30 days | Owner never opened ChatGPT — proves the system runs without owner involvement |
| Valerie | 80% dead leads → 80% qualified | AI Inbox reactivation — proves the inbox is the most underleveraged asset on a contractor's P&L |
| Ashley | $717K booked | Managed entirely off one visual dashboard — proves the dashboard isn't reporting, it's the operating system |
Rotation rules
- Riley → P2 lead capture / quote-to-job gap content
- Ashley → P1 authority / one-dashboard content
- Valerie → P2 dead lead reactivation / volume content
- Thrasher → P3 data / scale stories
- TWF → P2 speed / "system without owner" stories
Never use a case study without the mechanism. "$717K booked" alone is empty — "$717K booked managed entirely off one visual dashboard" is the lesson.