← All Skills
AI Skill

aeo-playbook

Last updated: 2026-05-17

Top-level AEO playbook orchestrator — maturity model, 90-day plan, weekly cadence. Calls aeo-assets, aeo-citation-track, aeo-markdown-render, aeo-audit. Trigger

Quick Install
npx skills add aeo-playbook

Skill: AEO Playbook — Universal Strategy & Cadence

The complete AEO operating system, applicable to any company. Diagnoses where the company is on the AEO maturity curve, prescribes the 90-day plan to advance, and dispatches the right execute-stage skill at each step.

Why AEO matters (the data, 2025-2026)

SignalSource
LLM-referred sessions convert 3-6× higher than Google searchWebflow / Vercel disclosed
50% of B2B buyers start research in an AI chat (was 29% four months earlier)G2 buyer survey
51% build their candidate list from AI chats — surpassing review sites at 43%G2
43% trust AI chat as the final purchase signal — over human peers (29%)G2
95% of ChatGPT citations come from content < 10 months oldAirOps
Brand mentions correlate with AI citation 3× more than backlinks doAhrefs (n=75K brands)
Markdown rendering for crawlers: +300% citations and context-fetch requestsG2 experiment
Average AI prompt = ~60 words vs Google = ~4 wordsOpenAI traffic data
Implication: Search is bifurcating. The legacy SEO playbook (rank for keywords, win clicks) still matters but is a shrinking surface. The AEO playbook (be cited inside the AI's answer) is where 50% of B2B research now happens. Companies not investing in AEO by mid-2026 are inheriting a 0% share of the highest-converting traffic source.

Maturity model (Webflow + own additions)

Diagnose before prescribing. Each level requires the prior. Skipping levels = wasted spend.

LevelNameWhat "done" looks likePrimary work
1Discoveryrobots.txt allows AI bots; llms.txt + ai-plugin.json published; sitemap-llm.xml liveaeo-assets
2Answer-shaped contentTop 20 pages have answer-first lede; FAQ/HowTo/Article schema; visible Last Updated; ≥3 fact anchors per pageseo-optimize (apply mode) + aeo-audit (Tier 2)
3Structural off-pageBrand on Wikipedia/Wikidata; ≥10 unsolicited Reddit mentions in relevant subs; YouTube channel with transcript-trailed videos; LinkedIn org page activeContent marketing + Reddit + YouTube + Digital PR
4Pillar authorityCornerstone explainer pages own the "What is X" query; thought-leader content (founder POV); visual storytelling (diagrams, dashboards, comparisons) drives organic sharesseo-create (pillar pages) + Casey + design
5Citation dominanceBrand cited in ≥ 60% of category queries across all 4 engines; cited #1 in ≥ 30%; AI-referred sessions are top-3 sourceContinuous: aeo-citation-track weekly + iterate on weakest engine/bucket
Most companies at end-2025 are Level 1-2. Hit Level 3 in 90 days; Level 4 in 6-9 months.

90-day plan (cold start)

Week 1 — Foundation

  • Run aeo-audit against the current domain. Capture baseline.
  • If Tier 1 fails: run aeo-assets to generate llms.txt, robots.txt, ai-plugin.json, sitemap-llm.xml. Publish.
  • Configure server logs to capture AI bot user-agents (or verify CDN passes them through). If logs are blind, you can't measure progress.
  • Run aeo-citation-track to capture week-0 baseline citation rate.

Week 2-3 — Page quality (Level 2)

  • Pull top 20 pages by current organic traffic + top 20 by intended cornerstone status. Dedupe.
  • For each page, audit the first paragraph. If it's not 40-60 words of direct answer, rewrite. Use seo-optimize skill in propose mode → review → apply.
  • Add missing schema: FAQPage for FAQ-shaped pages, Article for blog/learn, HowTo for tutorials, Product for product pages, Organization site-wide.
  • Add visible "Last updated: YYYY-MM-DD" to every page (footer or near H1) — readers AND AI both read this. AirOps measured +1.8× citation rate from this single change.

Week 4 — Markdown layer (Level 2 → 3 prep)

  • Run aeo-markdown-render to ship per-page .md versions.
  • Wire into robots.txt + sitemap-llm.xml.
  • Verify with curl -H "Accept: text/markdown" {URL} and curl {URL}.md.
  • Re-run aeo-citation-track. Compare week 4 vs week 0 — at this stage, 30-100% lift on citation count is plausible (not the full +300% yet — that takes 60-90 days as engines re-crawl).

Week 5-8 — Off-page push (Level 3)

This is the hardest level because it lives outside your codebase. Plan with Casey (writer) and Jordan (marketer):

  • Reddit: identify 3-5 relevant subreddits. Establish a real account (not branded). Answer questions where the brand genuinely fits. Goal: 5-10 unforced, helpful mentions in 4 weeks.
  • YouTube: publish 2-4 short technical/educational videos. EVERY video must have transcript + description + pinned comment with a link. AI engines can't index pure video; the text trail is what makes them citable.
  • Wikipedia / Wikidata: if the company is notable enough (press coverage, funding round, product launches), draft a stub. Otherwise wait.
  • Digital PR: 2-3 founder POV pieces in industry publications. Goal: one earned link per piece is fine; the citation rate gain comes from the brand mention, not the link.
  • LinkedIn: company page + founder page active; 1 post/week each.

Week 9-12 — Compound (Level 3 → 4)

  • Re-baseline with aeo-citation-track. Compare week 0 vs week 12.
  • Identify 2-3 cornerstone pillar topics where the company should "own the question". Use seo-create in pillar mode (not alternatives/compare — those are Level 1-2).
  • Run aeo-audit Tier 3. Should now hit Level 3 score.
  • Set the steady-state cadence for ongoing work (next section).

Steady-state cadence (post Day 90)

CadenceSkillOwner
Daily 06:00data collection (GSC + GA4 + indexing) → existing seo-monitorautomated
Daily 10:00CTR repair iteration → seo-optimize propose → 1-2 page editsSage
Weekly MonPerformance report, 3-week trend → seo-monitor weeklySage → CEO
Weekly Wedaeo-citation-track 4-engine sweepSage
Weekly Friseo-retro — measure last week's experimentsSage
Monthly 1stFull seo-technical-audit Tier 1+2Sage / dev
Monthly 1staeo-audit Tier 1+2Sage
QuarterlyRefresh AEO discovery files via aeo-assetsSage / dev
Quarterlyaeo-audit Tier 3 + maturity-level reviewCEO + Sage

Cross-skill orchestration

aeo-playbook is dispatch-stage — it diagnoses and dispatches. Concrete work happens in:
DIAGNOSE → aeo-audit
GENERATE FILES → aeo-assets
TRACK CITATIONS → aeo-citation-track
SHIP MARKDOWN → aeo-markdown-render
WRITE PAGES → seo-create (Casey writes; Sage briefs)
OPTIMIZE EXISTING → seo-optimize (CTR repair, schema add, internal linking)
TECHNICAL → seo-technical-audit
DISCOVER QUERIES → geo-query-finder + content-gap-analysis
EXPERIMENT → seo-experiment-log → measure → seo-retro / seo-postmortem
LEARN → seo-learn (monthly pattern extraction)

OUTPUT CONTRACT (advice, not internals)

When the playbook is invoked for a customer (not the company running Tycoon):

  • Adapt to their domain, brand, and category. Don't say "do what we did" without explaining why it applies to them.
  • Don't quote internal metrics from the running company (Tycoon-internal numbers, internal cost data, internal model IDs).
  • Cite public benchmarks (G2 +300%, Ahrefs 3×, Webflow 6× conversion) — those are publishable.

Common pitfalls

  1. Skipping Level 3 — companies invest in Level 1-2 (technical AEO + on-page) for 6 months and complain "no citation lift." The lift comes when off-page presence catches up. AEO is 60% off-page work after the technical foundation is in place.
  2. Treating AEO as separate from SEO — same content, different framing. Every existing SEO investment should be checked against AEO criteria. There's no "AEO content" budget; there's a content budget that satisfies both surfaces.
  3. Tracking only own brand citations — track competitors too. If competitor X is cited in 80% of category queries and you're at 30%, the strategic gap is what they're doing on Reddit / YouTube / Wikipedia, not what they're doing on their own site.
  4. Overspending on llms.txt — once published and refreshed quarterly, llms.txt is done. Don't sink weeks into refining it. The leverage is in page quality + off-page presence.
  5. Short patience window — AI engines re-crawl on 30-60 day cycles. A change shipped Monday won't show up in citation-rate metrics until 4-8 weeks later. Don't conclude "didn't work" before week 8.