Generative Engine Optimisation
Generative engine optimisation is the practice of making a brand easy for AI assistants to find, quote and recommend. Studio UNA does it for beauty, fashion and aesthetics brands, in the categories where being named inside an answer now does the work a search result used to do.
The list is being replaced by an answer
For twenty years, being found meant occupying a position in a list of ten blue links. A shopper typed a query, scanned the results, and chose. The whole discipline of search engine optimisation was built around winning one of those positions.
That is no longer how a growing share of discovery works. Someone asking an assistant which retinol to buy for sensitive skin, or which clinic in London does the best skin boosters, does not get ten links. They get a paragraph naming three brands. There is no second page, no scroll, and often no click at all. You are either in that paragraph or you are absent from the decision.
Generative engine optimisation is the work of getting into it. If you want the concept before the pitch, we have written a working definition of the term and what it does not cover.
What actually changes
Most of the technical foundations still apply, and any agency selling you a clean break from SEO is selling something that does not exist. Crawlability, canonical structure, site speed and clean markup matter as much as they ever did. What changes is the unit of content that competes.
| Search engine optimisation | Generative engine optimisation | |
|---|---|---|
| What competes | A page, for a ranked position | A passage, for inclusion in an answer |
| What wins | Relevance and authority to that query | Being quotable, and corroborated elsewhere |
| The unit of work | Keywords and pages | Claims, summaries and structure |
| Where evidence comes from | Backlinks | Independent sources that say the same thing |
| The measurement | Rank, impressions, clicks | Sampled mentions, cited sources, referral traffic |
| The failure mode | Ranking on page two | Not being named at all |
The practical consequence is that an assistant lifts a passage rather than a page. If the argument of a piece is buried three paragraphs into a scene-setting introduction, it does not get extracted. If a claim about a brand appears only on that brand's own website, it does not get trusted. Both of those are fixable, and neither is fixed by writing more blog posts.
How we work
1. Measure what the engines currently say
We start by asking. A fixed set of category prompts is run across ChatGPT, Perplexity, Gemini and Google's AI Overviews, and we record every brand named and every source cited. Most brands have never seen this, and it is usually uncomfortable: the sources being quoted about you are rarely the ones you would choose.
2. Fix what the engines cannot read
Structured data describing what each page actually is. Canonical URLs that resolve cleanly. Summaries written as standalone precis rather than teasers. Crawler access configured deliberately — the distinction between a training crawler and a search crawler is the one most brands get wrong, usually by accident. This is unglamorous and it is the part most sites are missing entirely.
3. Write things worth quoting
Answer-shaped content: the claim in the first sentence, the evidence after it. Original figures rather than restated consensus, because a number nobody else has published is the most citable thing a brand can produce. This is editorial work, which is why it sits with a consultancy that runs a publication rather than with a technical SEO shop.
4. Earn the corroboration
An assistant weighs what independent sources say about you far more heavily than what you say about yourself. Trade press, retailer copy, review platforms and category round-ups all feed the answer. Getting the same claim to appear consistently across them is slower than the technical work and it is what actually moves the outcome.
5. Measure again
The same prompts, on a schedule, so the change is visible. There is no rank tracker for answers, so repeated sampling is the honest method.
Why a beauty and fashion consultancy
Because the categories behave differently. A skincare recommendation gets filtered through claims regulation, ingredient scepticism and a decade of accumulated review data. A fashion recommendation runs on cultural positioning that a generalist agency will not read correctly. An aesthetics recommendation is governed by advertising rules that most marketers have never had to learn.
We work only in beauty, fashion and aesthetics, and we publish in these categories every week. Our analysis of why ingredient-led beauty brands dominate AI citations is the same work we do for clients, which is also why this site is built the way it is: structured data on every article, summaries written for extraction, and a machine-readable index of what we publish. It is a demonstration rather than a claim.
What we will and will not promise
We will make your site as easy to parse, quote and trust as it can be made. We will show you what assistants say about you today, and we will show you the same measurement again later so you can see whether it moved.
We will not guarantee that ChatGPT recommends you. Nobody controls that, and an agency that promises it is either misunderstanding the technology or counting on you not to check. What we can say is that brands which are structurally invisible to these systems stay invisible, and that is a solvable problem.
Further reading
- What is generative engine optimisation? A working definition — the concept, the four kinds of work it covers, and the four things it is not.
- How to get cited by ChatGPT — seven steps, starting with the crawler configuration almost nobody checks.
- The Algorithmic Shelf — why ingredient-led beauty brands are dominating AI citations, and what the losing brands have in common.
AI Visibility Audit
We run twenty category prompts across four AI assistants and record every brand they name, then read your site the way those engines read it. You get a prioritised list of what to fix, in writing.
- Twenty category prompts run across ChatGPT, Perplexity, Gemini and Google AI Overviews
- Every brand named, every source cited, and where competitors appear instead of you
- A technical read of your site against what those engines can actually parse
- A prioritised list of fixes, ordered by effort against likely effect
Questions we get asked
What is generative engine optimisation?
Generative engine optimisation is the practice of making a brand easy for AI assistants to find, quote and recommend. Where search engine optimisation competes for a position in a list of links, generative engine optimisation competes to be named inside a written answer. The work is part technical, part editorial and part earned coverage.
Is this different from SEO, or a rebrand of it?
Most of technical SEO still applies, and anyone telling you otherwise is selling a clean break that does not exist. What genuinely changes is the unit of content that matters. An assistant lifts a passage, not a page, so summaries, structure and corroboration across independent sources carry far more weight than they used to.
Can you guarantee ChatGPT will recommend my brand?
No, and neither can anyone else. We can make a site maximally easy to parse, quote and trust, and we can measure what assistants currently say about you and how that moves. We cannot instruct a model to cite you. Any agency guaranteeing AI visibility is promising something outside their control.
How do you measure it?
By sampling. We run a fixed set of category prompts across several assistants on a schedule and record which brands are named and which sources are cited, then track referral traffic from assistant domains in analytics. There is no rank tracker for answers, so the honest method is repeated measurement rather than a single number.
How long before it makes a difference?
Technical fixes are picked up within weeks. Being named in answers takes longer, because it depends partly on other people writing about you, and that is earned rather than switched on. Plan on a quarter before the measurement is worth reading, and two before it is worth reporting to a board.
Who is this for?
Beauty, fashion and aesthetics brands, and the clinics and retailers around them. We work with founder-led businesses through to established houses, across the UK, United States, Spain, France and the UAE. If you sell software or professional services, we are not the right firm for you.
Do I need to rebuild my site?
Usually not. Most of what matters is structured data, clean canonical URLs, summaries written to be extracted, and an agent-readable file describing what the site publishes. All of that can be added to a site you already have. We will say so plainly if a rebuild is genuinely the constraint.
What does it cost?
The AI Visibility Audit is a fixed fee of £1,500 and is the usual starting point. Ongoing work is scoped against what the audit finds, and we take on a small number of engagements at a time rather than running a retainer book.