docs: ingest AEO vs GEO distinction into VeriPath model (decision 2026-08-17)

docs/decisions/2026-08-17-aeo-geo-ingestion-into-veripath-model.md

@@ -0,0 +1,45 @@
+# Decision: Ingest AEO vs GEO Distinction into VeriPath Model
+
+**Date:** 2026-08-17  
+**Status:** Accepted  
+**Context:** Critical external knowledge from Smart Money Media (https://smartmoneymedia.org/blog/aeo-vs-geo-answer-engine-optimization-generative-engine and the definitive guide at /guides/aeo-vs-geo) must be incorporated so VeriPath's AI Visibility domain is precise and future-proof.
+
+## Summary of Source Knowledge
+
+Industry definitions of AEO and GEO are still converging, but two complementary framings are useful:
+
+### Framing A (Smart Money Media definitive guide, Aug 2026)
+- **AEO (Answer Engine Optimization)**: Focused on conversational AI engines (ChatGPT, Perplexity, Claude, Gemini, Grok, Copilot chat). Win condition = cited by name inside the generated answer.
+- **GEO (Generative Engine Optimization)**: Broader umbrella that includes AEO + Google AI Overviews, Bing generative answers, and other generative search surfaces. AEO is a sub-discipline of GEO.
+- Shared ~80% infrastructure: entity graph, schema, llms.txt, direct-answer content blocks, tier-1 earned media, AI crawler access, freshness.
+- Classic SEO (top-10 organic) is high-dependency for Google AI Overviews (GEO), moderate for pure chat AEO.
+
+### Framing B (earlier blog + other sources)
+- **AEO**: Extraction-oriented. Structure pages so machines can lift self-contained passages (featured snippets, AI Overviews, voice, PAA).
+- **GEO**: Reputation / citation-oriented. Build third-party authority and entity consistency so generative models choose to name and correctly synthesize the brand.
+
+Both framings agree on core tactics VeriPath already prioritizes: clean entity identity, structured facts, human-verified canonical record, tiered evidence, and measurement of AI representation.
+
+## Incorporation into VeriPath
+
+1. **Service Domain 2** is expanded from "AI Visibility Integrity / AEO" to **AI Visibility Integrity (AEO + GEO)**.
+2. Definitions are locked as follows for VeriPath operating language:
+   - **AEO**: The practice of making the business's facts extractable and correctly parseable by answer engines (structure, schema, direct-answer blocks, llms.txt, NAP/entity consistency).
+   - **GEO**: The practice of making the business a preferred, accurately represented source inside generative narratives (entity authority, tier-1 citations, cross-web consistency, brand-mention density in training/retrieval corpora).
+3. The canonical business record remains the single source of truth that powers both AEO (structured extractable facts) and GEO (trusted entity signals).
+4. AI representation tests in the audit workflow must cover both:
+   - Extraction surfaces (AI Overviews, featured snippets where applicable).
+   - Generative citation and accuracy across major chat engines.
+5. No change to evidence tiers or human-in-the-loop rule. Both AEO and GEO findings remain subject to Tier 1/2 classification and owner approval.
+
+## Rationale
+Local businesses lose customers when either (a) the AI cannot extract a clean answer or (b) the AI extracts/synthesizes the wrong competitor. VeriPath's strength is the verified canonical record that closes both failure modes. Explicit AEO/GEO language prevents under-specification when talking to prospects, agencies, or building future measurement.
+
+## Follow-on Actions
+- Update README.md Service Domains section.
+- Update any homepage / methodology copy that only says "AEO".
+- Ensure cold-audit AI representation prompts explicitly test both classes of surfaces.
+- Record this decision in the project memory and operating truth.
+
+**Owner:** Tony / VeriPath  
+**Verified against source:** Smart Money Media AEO vs GEO materials (Mar–Aug 2026).