AEO Audit:
Find Out Where AI Engines Cite You — And Where They Don't

An AEO (Answer Engine Optimization) audit is a diagnostic of your brand's visibility inside AI search engines — ChatGPT, Perplexity, Gemini, Claude, Microsoft Copilot, Grok, Amazon Rufus, and Google's AI Overviews. It measures citation share across a category-specific prompt corpus, audits the technical signals AI engines extract (schema graph, llms.txt, entity disambiguation, AI-crawler accessibility, answer-format density), benchmarks against competitors, and outputs a prioritized 30/60/90-day roadmap. It complements traditional SEO audits rather than replacing them.

Traditional SEO audits diagnose Google ranking factors — crawlability, indexation, Core Web Vitals, internal linking, content depth, backlinks, on-page tags. An AEO audit measures whether your brand is being cited inside AI-generated answers, which is a different question with different signals. Many of the foundations overlap (clean schema helps both; broken canonicals hurt both), but the measurement surface is different. We recommend pairing an SEO auditwith an AEO audit when an existing site is healthy enough that the question is “why aren't we cited?” rather than “why don't we rank?”

ChatGPT (Search and Shopping), Perplexity and Perplexity Shop, Google Gemini, Claude.ai, Microsoft Copilot, Grok, Amazon Rufus, and Google AI Overviews. We can add additional engines on request (Brave Search Summarizer, You.com, Mistral Le Chat, Cohere Coral) for categories where they matter — but the eight above cover roughly 95%+ of the discoverable AI-search citation surface for most B2C and B2B categories today.

Workspace — our proprietary in-house AI platform — runs your custom prompt corpus against each engine on a repeatable schedule, parses the output, and extracts: which brands are named, in what order, with what sentiment, and from which sources. We compare your share across awareness, consideration, and transactional prompts, and track the metric over time so you see whether the program is moving the needle. The methodology is documented in the deliverable so your team can replicate or validate it.

llms.txt is a proposed standard for surfacing structured information about your site to large language models — analogous to robots.txt but oriented toward content discovery instead of crawl rules. A well-formed llms.txt at the root of your domain helps engines like Anthropic's Claude, Perplexity, and OpenAI's ChatGPT understand your canonical entity, product taxonomy, FAQ corpus, and policy pages without scraping the whole site. It's not a ranking factor in the traditional sense, but adoption is moving fast and the brands publishing one early are getting cited in ways the engines can't easily explain away.

Both, depending on what you want. Many clients run the roadmap themselves once they have it. Others bring us in to execute — schema rollout, llms.txt drafting, page-template rewrites, entity disambiguation, and ongoing citation tracking. The audit itself is a one-time engagement; ongoing programs are scoped via AEO services and tracked monthly.

Most AEO audits land in 2–4 weeks from kickoff. The prompt-corpus build is the longest-lead item — we tune it to your category, your competitor set, and the funnel stages you care about. Once the corpus is live, Workspace runs it across engines in hours; the rest of the time is senior-strategist analysis and roadmap synthesis. Express turnarounds (~1 week) are available for narrower scopes (single-engine, single-funnel-stage).

AI engine outputs are non-deterministic and the models update on their own schedules — so a one-time snapshot loses signal fast. We recommend re-baselining quarterly for active programs. For clients on a retainer, Workspace handles continuous tracking against the locked prompt corpus and we surface deltas monthly. For one-time audits, the prompt corpus is yours to keep; you can re-run it later or hand it to your team.

Yes — and B2B and local categories often have the most asymmetric upside, because most competitors haven't touched AI visibility yet. B2B prompts (“best [SaaS category]”, “[vendor A] vs [vendor B]”, “is [vendor] enterprise-ready”) tend to surface short, named lists with high decision-impact. Local prompts (“best [service] in [city]”) increasingly trigger AI Overviews and ChatGPT Shopping-style results. See also multi-location SEO for brands with distributed footprints.

Engagements are scoped by depth — prompt-corpus size, competitor count, engine coverage, and whether earned-authority benchmarking is included. We'll quote the scope after a 30-minute discovery call. The audit is senior-led (no offshoring, no junior analysts running prompts against engines they don't understand); pricing reflects that.