Get recommended when buyers ask AI what to buy
Your customers are asking ChatGPT, Perplexity and Google’s AI answers which brand to buy. Those answers name a handful of brands. This is the work that gets you named.
The research step moved. Most brands did not move with it.
A growing share of product research now starts in a chat window instead of a results page. The buyer describes what they need, the model answers with three or four brands, and everything else is invisible. There is no page two to grind toward and no ten blue links to occupy. You are either named in the answer or you are not in the consideration set at all.
There are two separate reasons a brand does not get named, and they need different fixes. The first is that the model does not know you — your brand is thinly and inconsistently described across the web the model learned from. The second is that the model cannot retrieve you — your pages bury the answer, state nothing specific, and give it nothing clean to quote. Most brands are losing on both counts, which is good news, because both are fixable.
AEO, GEO, AIO: three names for the same job
The industry has not settled on a word for this yet, so you will see three acronyms used almost interchangeably. Here is what each one actually emphasises, and why the distinction matters less than the outcome.
Answer Engine Optimization (AEO)
Earning the answer itself. AEO frames the goal as being the response a buyer gets, not a link they might click. In practice it emphasises answering directly, stating specifics a model can lift, and structuring pages so the answer is easy to find and hard to misread.
Generative Engine Optimization (GEO)
Earning the citation inside a generated answer. GEO is the term most used inside the SEO industry, and it leans on the mechanics of retrieval: which sources a model pulls, how it attributes them, and what makes your page one of the few it reaches for.
AI Search Optimization (AIO)
The umbrella over both, plus the part neither acronym covers: how consistently your brand is described everywhere else on the web, which is what a model falls back on when it answers from training rather than browsing.
Pick whichever term you like. The work is the same, and so is the only question that matters: when your buyer asks, does the answer name you? I have been arguing a version of this since 2023 — see Will AI Kill SEO? for the long version and what I got wrong. On the demand side, search intent urgency is the other half of why specific pages win.
What the engagement covers
Prompt-set baseline
A fixed set of real buying prompts run against each engine, logging verbatim what is said about you and which competitors get named instead.
Entity and brand consistency
Aligning how your brand, products, materials and claims are described across your own site and every third-party source a model might corroborate against.
Answer-first content
Rewriting pages to lead with the answer, state specifics, and be easy for a model to quote accurately instead of paraphrasing into something wrong.
Technical retrievability
Crawler access, structured data and page structure so browsing retrieval can actually read your content and attribute it to you.
Third-party corroboration
Coverage, listings and reviews on the sources that already shape what models believe about your category. This is the half most brands skip.
Citation and answer tracking
Monthly re-runs of the prompt set so you watch your share of answers move rather than taking the work on faith.
How the program runs
Baseline the prompts
Build and run the prompt set that mirrors how your customers actually ask, across ChatGPT, Perplexity and Google AI answers. Log every response verbatim.
Diagnose the gap
Work out, per prompt, whether you are missing because you are unknown, misdescribed, or simply not retrievable. Those need different fixes.
Fix content and entity
Rewrite the pages, fix the structured data, and correct the third-party record where it is wrong or thin.
Re-test and iterate
Re-run the prompt set on a schedule and keep working the questions that still go to competitors.
Why this is not just SEO with a new name
Traditional SEO optimises a page to rank. This optimises a brand to be retrieved, quoted and recommended — which is a different job with different inputs. What that gets you:
- Citations compound. Once you are the source for a question, you tend to stay the source.
- Being misdescribed is worse than being absent, and it is the fastest thing to fix.
- The underlying work also improves classic SEO and the experience for the humans who land on the page.
- Most competitors in DTC categories are still ignoring this entirely, so the bar is currently low.
- It is measurable with a repeatable prompt set, rather than taken on faith.
AI search and answer engine FAQ
What is answer engine optimization (AEO)?
AEO is the practice of getting your brand named and cited inside the answers AI engines generate, rather than ranked in a list of links. It combines content work (answering directly and specifically), technical work (structure and retrievability) and off-site work (making sure the wider web describes you consistently).
Can I pay to be recommended by ChatGPT?
No. There is no ad slot inside an organic recommendation, no submission form, and no placement to buy. Anyone offering guaranteed placement is either misunderstanding how this works or misrepresenting it. What you can do is systematically remove the reasons you are not being named.
AEO, GEO or AIO — which do I actually need?
Whichever your team will actually say out loud. The acronyms describe overlapping views of one problem, and I run them as one engagement. If a vendor tells you these are three separate products with three separate retainers, that is a pricing decision, not a technical one.
ChatGPT says something inaccurate about my brand. Can that be fixed?
Often yes, and it is usually the highest-value fix available. Inaccuracies almost always trace back to outdated or inconsistent information on your own site or on a third-party source. Correct the record where it originates and the answers tend to follow.
How do you measure results?
A fixed prompt set, re-run on a schedule, tracking whether you are named, where you appear in the answer, and how you are described. That plus referral traffic from AI engines in your analytics. You watch a share-of-answers number move, not a vanity score.
How long before the answers change?
Retrieval-based answers can shift within weeks of publishing better pages. Answers coming from training data move far more slowly and depend on the wider web catching up. I set expectations per prompt based on which mode it is using rather than quoting one timeline for everything.
Find out what AI says about you
Tell me your category and I will run your real buying prompts through ChatGPT and Perplexity and send you the actual answers, including who is getting recommended instead of you.
