AEO recovery.
12-step DIY audit.

AEO (Answer Engine Optimization) recovery is the structured process of reclaiming organic visibility lost when AI Overviews, ChatGPT search, Perplexity, Gemini, and Claude began answering queries directly inside the SERP — without sending the click to your site. Recovery work shifts the goal from "rank for the keyword" to "be the cited source inside the answer." It combines schema graph completeness (Organization, Person, Service, FAQ, Article, Speakable), llms.txt/ai.txt publishing, TL;DR + structured Q&A page restructures, citable original statistics, author Person schema, and entity-graph signals (Wikipedia/Wikidata adjacency). Most brands recover citation share across 4 AI engines within 90–180 days when all 12 steps in this checklist ship.

Open Google Search Console. Filter to the last 90 days. Sort queries by impressions (high → low). For each top query, look at the clicks-to-impressions ratio. If impressions are stable or rising but clicks dropped 20%+, that query is being answered inside an AI Overview, ChatGPT search, or Perplexity panel — and the click is no longer reaching you. Run the same check on your top 50 queries. The pattern is unmistakable when it is happening, and silent when it is not.

First citation lifts usually appear inside 30–45 days after schema graph + llms.txt + TL;DR restructure ship for top-20 commercial pages. Meaningful share shift across 4 AI engines (ChatGPT, Perplexity, Gemini, Claude) typically lands in 90–180 days. The longest-leverage signals — original research published quarterly, Wikipedia/Wikidata adjacency, third-party citation density — compound over 6–12 months. Brands that wait for "AEO best practices to settle" lose category share that takes 3–4× longer to rebuild.

Yes. Sitemap.xml tells search crawlers which pages exist; llms.txt tells AI training and inference systems how you want them to use your content — preferred canonical URLs, content categories, citation expectations, off-limits paths. The two files do different jobs. ChatGPT, Perplexity, and Claude crawlers explicitly check for llms.txt; sitemap parsing is secondary. Drafting llms.txt from your real content (not boilerplate templates) typically takes 2–4 hours and is the single highest-leverage AEO-specific file you can publish.

Schema is necessary but not sufficient. A complete graph (Organization, Person, Service, FAQPage, Article, Speakable) gives AI engines the structured surface they need to extract entities, attribute claims, and decide who to cite. Without it, even high-quality content gets paraphrased without attribution. With it, the same content gets cited as the source. Schema alone will not move you to top-cited if your content lacks citable original stats, authored bylines, or entity-graph signals — but every other lever is held back without it.

Three options. First, manual: query 30–50 of your priority terms across ChatGPT, Perplexity, Gemini, and Claude monthly; log who is cited. Second, paid tools: AthenaHQ, Otterly.ai, Profound, and Peec.ai automate citation tracking across engines with brand vs competitor share, query-level breakdowns, and trend lines. Third, brand-mention monitoring: Brand24 or Mention catch AI-engine outputs that reference your brand even when not formally cited. We recommend a paid tool plus monthly manual spot-checks — purely automated tracking misses tone and context shifts.