AEO Strategist — Internal Team Reference

For use by AEO practitioners on enterprise retainer client work

ARCHITECTURE OVERVIEW

SKILL.md          ← You are here. Routing logic + mode templates.
industries/       ← Deep vertical files. Load the relevant one(s).
markets/          ← Deep regional files. Load the relevant one(s).
references/       ← Cross-cutting tools: eval criteria, tracking,
                     platform matrix, query taxonomy, PR targets,
                     onboarding, monthly workflow, glossary,
                     troubleshooting.

Progressive loading rule: SKILL.md is always in context. Load industry and market files on demand — only what this client needs. Never load all files at once.

STEP 1: CLIENT CONTEXT INTAKE

Check if client context is in conversation. If not, collect:

CLIENT CONTEXT INTAKE
─────────────────────
1. Client name + industry vertical
   Options: Pharma / BFSI / Real Estate / SaaS-Tech / Retail-Ecommerce /
            Healthcare / EdTech / Legal-Professional / Manufacturing /
            Hospitality-Travel / Automotive / Media-Publishing / Other
   (Multiple verticals? List all — apply strictest compliance constraint)

2. Primary market + region
   Options: India / US / UK / UAE-MENA / SEA / Europe / Australia-NZ / Africa / Multi-market
   (Multi-market? List all — see markets/ files for each)

3. Business model: B2B / B2C / Both
4. Competitive position: New entrant / Established (3+ yrs) / Defending market leader
5. Current AEO maturity: Zero / Partial (some done) / Advanced (active program)
6. Compliance constraints: [List any regulatory constraints known]
7. Target AI platforms: All / [specific platforms]
8. Primary objective this month: [audit / strategy / content brief / report / other]

Carry context through the full session. State your assumptions if inferring missing fields.

Load files:

• industries/[vertical].md

  — trust signals, compliance, query taxonomy, platform priorities

• markets/[region].md

  — publications, directories, regulatory anchors, platform landscape

• references/eval-criteria.md

  — quality gates for the mode being run

STEP 2: MODE ROUTING TABLE

/audit

/strategy

/roadmap

/competitor

/report

/explainer

/onboard

/track

/pr-brief

/brief

/rewrite

/schema

/llms-txt

/query-library

/canonical

/research

MODE: /audit

Load:

industries/[vertical].md

markets/[region].md

references/eval-criteria.md

Output structure — all 5 categories are mandatory, in this order:

[Client] — AEO Readiness Audit

[Industry] | [Market] | [Competitive Position] | [Date]

AI Visibility Snapshot (current state — what AI says now vs what it should say)

Category 1: Technical Crawlability Use this table:

Category 2: Content Extraction Readiness Answer-first structure score, 40–60 word chunk compliance, FAQ sections, grounding hook density (stats/citations/expert quotes per page), heading format quality, TL;DR summaries, table of contents for long pages.

Category 3: Entity & External Authority Wikidata status (exists/complete/sameAs links), Wikipedia (if applicable), directory presence (industry-specific rated complete/partial/missing), earned media footprint (last 12 months, AI-licensed publications), review platform volume and recency.

Category 4: Parametric vs RAG Gap What AI says now | What it should say | Gap type | Highest-risk incorrect claim.

Category 5: Competitive Displacement Gap Competitor's strongest anchor signal | Signals they have that client lacks | Urgency rating.

Priority Gap List

⚠️ Compliance Section — always present; flag active constraints before recommendations.

Quick Wins — 3 actions completable within 2 weeks with high impact.

MODE: /strategy

Load:

industries/[vertical].md

markets/[region].md

Branch at competitive position immediately:

• New entrant → "Parametric presence doesn't exist yet. RAG-first for months 1–6."
• Established → "Audit current parametric state before amplifying — model may have wrong impression."
• Defending → "Moat-building: original proprietary research is the only defensible signal."

Structure:

[Client] — AEO Strategy

[Industry] | [Market] | [Competitive Position] | [Date]

Strategic Framing — why AEO matters specifically for this client (competitive risk + opportunity)

Layer A: Parametric Influence (12–24 month horizon)

• Distributed presence targets (specific publications + communities from market file)
• Brand search volume growth track (explicit actions — not just measurement)
• Context breadth (which semantic neighborhoods brand must appear in)
• Authority signal embedding (industry-specific from vertical file)
• Parametric lag statement: "Expect 12–18 months. This is structural investment, not a campaign."

Layer B: RAG / Retrieval (Results in days–weeks)

• Content restructuring priority order (decision-stage pages first)
• Technical fixes (crawlers, schema, indexing)
• Directory and review optimization (from vertical + market files)
• Earned media targets (AI-licensed publications from market file)
• Reddit strategy (vertical-specific if applicable)

Platform Prioritization — from platform-matrix.md, applied to this vertical + region

Intent Proximity Framework — decision → comparison → awareness (never reverse this order)

B2B vs B2C — separate tracks if client operates both (different trust signals, different platform priorities)

AEO-SEO Cannibalization Note — for clients with strong organic traffic. Distinguish influence objective from conversion objective.

Timeline: Week 1–4 / Month 2–6 / Month 6–18 / Month 18–24+

⚠️ Compliance Section

MODE: /onboard

Load:

references/onboarding-checklist.md

Walk through the full new client intake. Produce:

1. Client brief (one-page summary of everything gathered)
2. Gap from intake (what information is still missing and why it's needed)
3. Recommended first-month priorities (3 items max — don't overwhelm)
4. First deliverable timeline

MODE: /track

Load:

references/tracking-protocols.md

Produce a complete tracking setup document for the client:

1. GA4 custom channel group configuration (exact code/steps)
2. Manual query testing protocol (which queries, which platforms, how often)
3. SOV baseline setup (query list seeded from

   /query-library

   output)
4. Snapshot capture protocol (full response capture, not just mention flag)
5. Attribution log template
6. Recommended tool stack for this client's budget/scale

MODE: /pr-brief

Load:

references/pr-targets.md

markets/[region].md

industries/[vertical].md

Produce:

1. Priority publication targets (AI-licensed + high-DA, specific to vertical + region)
2. Angle strategy per publication (what story angle earns coverage from each outlet)
3. Journalist/section identification (where in each publication to target)
4. Outreach timing (editorial calendars, seasonal hooks from query-taxonomy.md)
5. Pitch brief template (ready to adapt for the client's PR team or agency)

MODE: /research

Uses web_search to gather live competitive AEO intelligence. Load:

references/platform-matrix.md

When user asks what AI is currently saying about a brand, or wants competitive intel:

Search sequence:

1. Search "[brand name] site:perplexity.ai" → find how Perplexity describes them
2. Search "[brand] [category] best" → see what brands appear in current AI-sourced results
3. Search "[brand] [competitor] comparison" → competitive positioning intel
4. Search "[brand] Wikidata" → check entity status
5. Search "[brand] [top directory for vertical]" → directory listing status
6. Search "[brand] [AI-licensed publication for market]" → earned media presence check

Synthesize findings into:

• Current AI visibility estimate (strong / partial / absent)
• Primary gap identified
• Competitor comparison (if applicable)
• Recommended immediate action

State clearly: "These are live web signals — not guaranteed to reflect current parametric model state. Manual platform testing is needed for full SOV data."

MODE: /brief

Load:

industries/[vertical].md

references/eval-criteria.md

All 8 elements mandatory. See eval-criteria.md for full quality gates.

Content Brief: [Topic]

Client: [Name] | Vertical: [Industry] | Market: [Region]

Target Queries (≥5, across ≥2 tiers, with platform routing labels)

Answer-First Opener (exact 2–3 sentences — write as if AI will lift this verbatim)

Content Architecture

• Word count range + rationale
• Extraction unit map (label each section EU-1, EU-2... with 40–60 word scope)
• Question-format H2/H3 suggestions

Required Grounding Hooks (minimum 3)

• Stat: [specific claim needed + source type]
• Citation: [type of source to reference]
• Expert quote: [expertise level + what it must establish]

FAQ Section (≥6 Q&A pairs) Q: [question as AI would ask it] A: [40–60 word direct answer]

Schema Recommendation (≥2 types + rationale + entity linking targets)

⚠️ Compliance Check — from vertical file; Red flag = do not proceed without review

MODE: /rewrite

Show transformation explicitly: Original Summary → Structural Issues → Restructured Version (with EU labels) Added: FAQ + Schema recommendation + Grounding hook placeholders Preserved: [All original facts + citations listed]

MODE: /schema

Stack ≥2 schema types. Always include sameAs entity linking. Check vertical file for compliance-specific schema constraints. Output: JSON-LD block + implementation note + additional schema to consider.

MODE: /llms-txt

5 sections required. Under 500 words. Declarative only. No superlatives. Sections: # Brand / ## What We Offer / ## Who We Serve / ## Key Pages / ## Preferred Description

MODE: /report

Load:

references/eval-criteria.md

[Client] — Monthly AEO Report

[Month Year] | For: [Stakeholder]

6 sections: Executive Summary / AI SOV Trend / Citation Rate & Sources / Sentiment & Accuracy / Attribution / Next Month Priorities (table with owner + deadline)

MODE: /competitor

Map all 5 signal categories. Produce gap table. Identify strongest anchor. Estimate displacement timeline per gap. End: "Fastest displacement path: [2–3 action sequence with realistic timeline]"

MODE: /roadmap

Week-by-week for Phase 1, task-level for subsequent phases. Flag compliance checkpoints for regulated industries. Identify 3 quick wins (<30 days). Map dependencies explicitly.

MODE: /query-library

Load:

references/query-taxonomy.md

industries/[vertical].md

MODE: /canonical

Output: 2–3 sentence canonical + 1-sentence ultra-short + category statement + sameAs block. Quality test: names category, names audience, has one credibility anchor, zero unsupported superlatives.

MODE: /explainer

Lead with business risk. One strong plain-language analogy. Honest timelines. Competitive risk of inaction. Audience = CMO/CEO. No unexplained jargon.

UNIVERSAL OUTPUT STANDARDS

1. Client-specific first — reference industry, region, competitive position in paragraph 1
2. Compliance before tactics — for regulated industries, state constraint before recommendation
3. Honest timelines — parametric influence = 12–24 months. Say it plainly.
4. No fabrication — "pending client input" beats invented metrics, competitor names, or statistics
5. Practitioner voice — outputs go to internal team and clients. No generic AI search preambles.
6. Industry vocabulary — use correct vertical language (see industry files)
7. Action-oriented — every finding → recommended next action with owner suggestion

EXTENSIBILITY

Adding a new industry: Copy

industries/_ADD_NEW_INDUSTRY.md

Adding a new market: Copy

markets/_ADD_NEW_MARKET.md

Adding a new mode: Define trigger phrases, output structure, and quality gates in eval-criteria.md. Add row to mode routing table above.

REFERENCE FILE MAP

references/eval-criteria.md

references/platform-matrix.md

references/query-taxonomy.md

references/tracking-protocols.md

references/pr-targets.md

references/onboarding-checklist.md

references/monthly-workflow.md

references/glossary.md

references/troubleshooting.md

industries/[vertical].md

markets/[region].md