AI assistants now answer the questions your buyers used to type into Google. If your brand is not in those answers you are invisible at the moment of highest intent. These are the eleven questions marketing, growth and SEO leads actually ask about it — answered directly, because that is the format an assistant can quote.
What is answer engine optimization (AEO)?
Answer engine optimization is the practice of structuring, publishing and distributing content so that AI assistants — ChatGPT, Claude, Gemini, Perplexity — surface your brand when buyers ask questions in your category. Traditional SEO targets ranked blue links. AEO targets the synthesized prose answer that replaced them. A brand with strong AEO is named, cited positively, and recommended over its competitors inside the answer itself.
How is AEO different from traditional SEO?
SEO optimizes for a ranked list of links the user still has to click. AEO optimizes for the answer itself. The signals overlap — authority, structured data, freshness — but AEO adds dimensions SEO has no concept of: whether AI crawlers can access your site at all, whether your brand appears in the sources those models retrieve, and whether the sentiment of those mentions is positive.
A brand can rank on page one of Google and be completely absent from AI answers. Those are separate fights.
What is an AEO score and how is it calculated?
An AEO score is a single 0–100 index summarising how an assistant represents your brand across four dimensions:
- Presence — are you named at all?
- Position — how early in the answer?
- Prominence — a passing mention, or the recommendation?
- Sentiment — how are you framed?
Flare calculates it by putting real buyer prompts to eight assistants and aggregating every judged answer, so the number can be tracked over time and compared against the competitors the models actually name — not the ones you chose to watch. Unbranded prompts weigh most, because being named when nobody typed your name is the visibility that pays.
How do AI assistants decide which brands to recommend?
Two sources: training data absorbed before a knowledge cutoff, and live retrieval at query time for grounded models such as Perplexity. Brands that appear often in high-authority, crawlable sources — review sites, industry publications, comparison pages, their own indexed content — are likelier to be named.
Critically: if your robots.txt blocks GPTBot or ClaudeBot, neither the
training pipeline nor the retrieval pipeline can read your content, and no amount of quality compensates.
How do I find out if AI crawlers can read my website?
Open yourdomain.com/robots.txt and look for Disallow rules naming
GPTBot, ClaudeBot, Google-Extended, CCBot,
PerplexityBot or anthropic-ai. A Disallow: / against any of those means
that system cannot crawl you.
Check the file as served, not as committed. Some CDNs inject a managed block of AI-crawler rules ahead of your own, so the file in your repository is not necessarily the file the crawler reads. We found exactly that on this domain, which is a good illustration of why the check is worth doing by hand.
Why does a competitor keep getting cited instead of my company?
Usually one of three structural things:
- their content is crawlable by AI bots and yours is not;
- they have more third-party sources mentioning them — reviews, press, analyst coverage;
- their content answers the buyer's question explicitly, while yours describes a product.
Flare surfaces the citations behind each answer, so the reason comes back as a specific URL rather than a theory about brand strength.
What content changes actually move an AEO score?
Four, in rough order of return:
- Publish question-and-answer content that mirrors the prompts buyers actually type.
- Add
FAQPageandHowToschema so retrieval systems can parse the answers directly. - Earn mentions on third-party sources the models already trust.
- Make comparison and category pages name the specific use cases and alternatives buyers ask about.
How often should I measure?
Monthly is the floor — model updates, competitor content and changes in retrieval sources all move the number between cycles. If you are actively shipping content against a gap, measure before and after each push. Attribution is the entire reason to hold a baseline, and you only get one chance to take it.
Can I measure across several assistants at once?
Yes, and you should. Flare submits one prompt set to eight assistants in a single run and returns a unified score plus a per-model breakdown. It matters because retrieval architectures differ: a brand can be prominent in Perplexity, which retrieves live, and absent from a model answering out of training data. The per-model gap tells you whether you have a crawl problem, a citation problem, or a content problem.
What is a classic-search baseline, and why does it sit next to the AEO score?
It measures traditional organic visibility over the same topic set used for the AEO score. Running both separates three situations that look identical from a pipeline dashboard: losing ground only in AI answers, only in classic search, or in both. The remedies differ — crawler access and citations for one, technical SEO and links for the other — so without the baseline you cannot tell which fix to buy.
What does Flare cost?
Access is invite-only while the model bill stays sane. Request it, click the confirmation email, and a person reads it. Credits are one-to-one with dollars after that — no seats, no tiers, and nothing gated behind a plan.
Related: how to monitor your brand in AI answers, or what Flare does.