Answer engine optimization FAQ: straight answers

AEO, also called GEO or AI search optimization, is the work of making your business more likely to be named, accurately described and cited when people ask AI tools like ChatGPT, Gemini or Google’s AI Overviews for recommendations.

SEO aims to rank your pages in the list of links. AEO aims to get your business named, described accurately and cited inside the AI answer itself. That matters because one AI answer often names just three to six businesses. The foundations overlap (crawlable pages, clear facts, authority). AEO adds measuring AI answers directly, answer-first page formats and work on the third-party sites engines cite. Good SEO is still the base.

Mostly naming. GEO (generative engine optimization) comes from a 2023 research paper; AEO is the term marketers adopted. Both describe improving how AI answer engines represent you. We use the terms interchangeably.

ChatGPT, Gemini, Google AI Mode and Google AI Overviews as standard. Perplexity, Claude, Microsoft Copilot and Grok are attempted on request and only reported when we can capture them reliably.

We send each customer question to each engine as a fresh, logged-out session in the US, twice, and record the full answer text and every cited link. We don’t use your accounts.

Ask it the questions your customers ask, in a logged-out session, several times, and record which businesses it names and which links it cites. Answers vary, so one check isn’t enough. Our audit does this at scale: 30 questions × 4 engines × 2 runs. Step-by-step: how to track brand mentions in ChatGPT.

It helps engines read your facts consistently, and it’s cheap to do right. Tests so far suggest it moves citations little on its own. Mentions on trusted third-party sites and pages that answer the question directly matter more. We include schema but don’t oversell it. More in our guide to getting cited in Google AI Overviews.

Probably not. llms.txt is a proposed standard, a plain-text guide to your site for language models. As of November 2026, no major AI engine has confirmed it uses llms.txt for answers. We include one in the Implementation Pack because it’s cheap, and we label it as unproven.

It overlaps. Crawlable, well-structured, authoritative content helps both. AEO adds tracking of AI answers, answer-first content formats and a focus on the third-party sources AI engines cite.

Agencies commonly quote $1,500 to $5,000 for an audit and $1,500 to $8,500+ per month for retainers. Self-serve tracking tools cost $29 to $500+ per month. Our prices are fixed: $750 audit, $1,500 Implementation Pack ($1,950 bundled) and $290 or $590 per month for monitoring. See AEO services and pricing.

It depends on whether your customers use AI to find businesses like yours. Many now do. An AI visibility audit is the cheapest way to find out: it shows whether you’re named for the questions that matter, and against which competitors, before you commit to ongoing spend.

One. Every engagement starts with a quick phone call to confirm your goals, competitors and scope. After that, question review, delivery and follow-up questions run by email and shared files, which keeps prices low.

There is no online checkout. After a quick phone call we send an invoice for the agreed scope, payable by bank wire or ACH. Work starts when payment and intake are complete.

By default, no. We deliver schema code, page drafts and install notes as files for your team or developer. If you grant editor access to your CMS, we can publish them for you.

Knownstone is run by Hutcheson Labs. We use AI assistants to collect and analyze data at scale, and a person reviews every report before delivery. AI-generated analysis may contain errors, and we correct anything you flag.

Yes. Monitoring is month-to-month and invoiced monthly. Cancel by email before your next billing date and you won’t be billed again.

No. Nobody controls what AI engines say, and answers change from one run to the next. We measure where you stand, fix what can be fixed and track the trend. Anyone promising guaranteed placement is overselling. Here’s what does help: how to get recommended by ChatGPT.

Engines pick up changes on their own schedules. Expect 60 to 90 days before re-measuring makes sense, and treat early movement as directional.

Yes. We publish a real, anonymized sample report of a well-known sales CRM (not a client).