diff --git a/docs/homepage-blueprint-draft.md b/docs/homepage-blueprint-draft.md
index bff835c..e1bd3d6 100644
--- a/docs/homepage-blueprint-draft.md
+++ b/docs/homepage-blueprint-draft.md
@@ -15,23 +15,24 @@ Split visual:
 **Interactive element (slider/tab)** — anchored on the **Gilmore HVAC** live audit (`research/cold-audits/cold-audit-hvac-gilmore.md`):
 - Prompt shown: *"Best HVAC company in Cameron Park, CA"* (Perplexity/ChatGPT).
 - Leak shown: 3 conflicting phone numbers (page `(916) 884-6354` / schema `+1-530-290-5551` / snippet `(530) 344-4515`) + schema geo `Antelope, CA 95843` vs Cameron Park target.
-- Caption: "The AI doesn't know which number is real. Neither do the customers it sends elsewhere."
+- Caption: "The AI doesn't know which number is real. Neither do the customers it sends elsewhere."  
+  (This is the Screenshot of Shame moment.)
 
 ## Section B — "Three Domains of Integrity"
 1. **Customer Path Integrity** — diagram AI Search → SERP → Website → Conversion, leak markers at each handoff. Grounded in our audit gaps (Martel: zero machine-readable NAP = leak at AI Search).
 2. **AI Visibility Integrity (AEO)** — how LLMs parse local entities (schema, citations) vs how legacy SEO tools assume. Grounded in CPFD generic-schema finding.
-3. **Local Competitive Awareness** — "Shadow Competitors" ranking in AI but not in top-10 blue links. Grounded in Gordon Law: rank #1 at home, invisible for Cameron Park.
+3. **Local Competitive Awareness (Clean Competitive Fuel)** — "Shadow competitors" ranking in AI but not in top-10 blue links. We refine fragmented public data (reviews, offers, hiring signals, AI mentions) into trustworthy local market intelligence. Grounded in Gordon Law: rank #1 at home, invisible for Cameron Park.
 
 ## Section C — "Engagement Ladder" (the motion, radical transparency)
 - **Step 1 — Cold Audit (Layer 1a)**: Free. We find the leaks. You keep the report.
 - **Step 2 — Threat Diagnosis**: Paid fixed-fee. We quantify revenue impact of the Silent Customer Loss.
-- **Step 3 — Focused PoC**: We fix one critical leak. You measure.
-- **Step 4 — Retainer**: Ongoing human-supervised, agent-assisted integrity. *"We do not accept retainers without a successful PoC."*
+- **Step 3 — Focused PoC**: We fix one critical leak **and make the business Agent-Ready** (structured facts AI systems can trust). You measure.
+- **Step 4 — Retainer**: Ongoing human-supervised, agent-assisted integrity + monthly measurement loop. *"We do not accept retainers without a successful PoC."*
 
 ## Section D — "Objection Handling" FAQ
-- **Q: Why not just use an automated AI citation builder?**
+- **Q: Why not just use an automated AI citation builder?**  
   A: Automated tools generate Tier 2 (indicative) data; LLMs hallucinate or ignore unverified input. DOP relies on Tier 1 (verified) evidence, human-approved, so the AI actually trusts the entity.
-- **Q: How is this different from Bizl?**
+- **Q: How is this different from Bizl?**  
   A: Bizl shows Tier-3 self-reported case studies. Our 5 live audits are observed, falsifiable evidence — we show our work.
 
 ## Trust Anchor (site-wide footer/header)