AI SEO is real. But most of what agencies are currently selling under that name is not.
If you've been pitched AI SEO in the past year, you'll recognise the pattern. Different agency. Same presentation. Something about ChatGPT, something about Perplexity, a slide called "the future of search", and a monthly retainer that looks suspiciously like the one they were selling twelve months ago with different branding on the cover.
This piece covers what's genuinely different about AI SEO, what's repackaged standard work, and a practical way of evaluating what any AI search agency is actually offering before you commit.
What "AI SEO" actually means (and why nobody agrees)
Three genuinely different things are being sold under one label. It helps to know which is which.
GEO (generative engine optimisation) is the genuinely new discipline. It involves working on what ChatGPT, Perplexity, Claude, Microsoft Copilot, and Google AI Overviews say when someone asks about your business, your industry, or a problem you solve. This is not traditional ranking. AI systems build their answers from entity models, training data, and indexed content, and they do it outside the search signals Google has spent twenty years training the SEO industry on. Understanding GEO vs SEO as structurally distinct disciplines matters practically, not just theoretically, because the approaches required are different and so are the timelines and the measures of success.
AEO (answer engine optimisation) is older than most agencies let on. Optimising for Google's featured snippets and People Also Ask boxes has been on SEO roadmaps since around 2015. With AI Overviews expanding across more queries, AEO is firmly back in the conversation. But it is not a new invention. Agencies presenting it as one are being slightly misleading.
Traditional SEO with an AI-augmented workflow is what most "AI SEO" retainers actually deliver. Using Semrush's AI features, SE Ranking, or similar tools to produce content faster, run technical audits, and build keyword clusters more efficiently. The output is standard SEO. The difference is speed, not strategy.
The problem is that most agencies bundle all three under "AI SEO" without being clear about which you're actually getting. That lack of transparency is where the buyer's regret tends to start.
Gartner research from 2024 projected traditional search engine volume would fall 25% as queries migrate to AI assistants. That figure is worth noting now that we are past the projection date. The shift is real, even if the precise percentage continues to evolve.
TL;DR: AI SEO services cover three things: optimising for what AI systems like ChatGPT, Perplexity, and Google AI Overviews say about your business; optimising structured content to appear in AI-generated answer summaries; and using AI tools to make traditional SEO more efficient. The first two are genuinely new strategies. The third is a workflow upgrade, not a strategy change. Most agencies sell all three. Worth knowing which you are paying for.
What is actually worth paying for in 2026
This is where good agencies differ sharply from the rest.
Citation and recommendation visibility. When a potential customer asks ChatGPT or Perplexity "who should I use for X?", your business either appears or it doesn't. That is a new commercial channel, and it matters most in professional services, healthcare, and B2B sectors where buyers research carefully before contacting anyone. AI search traffic converts at significantly higher rates than typical search traffic, partly because people arriving via AI recommendations are already in verification mode rather than comparison shopping. Being cited costs nothing per click. Not being cited means the query went to a competitor. That is the commercial case, plainly stated.
Entity establishment. AI systems do not just read websites. They build models of who a business is, what it does, and who it serves, drawing from signals across the whole web. Consistent name, address, and contact details across directories. A complete Google Business Profile. Third-party mentions in credible publications. Content that clearly describes the business's expertise and the clients it serves. This is the foundational work that determines whether an AI assistant "knows" your business when asked about your category. Done properly, it is worth every penny.
Content that earns AI citations. Original data points, expert-authored articles with clear quotable statements, specific UK statistics, named tools and platforms. AI systems cite content that has something definite to say. A 2,500-word article making three clear claims is far more citable than a 4,000-word article hedging every point. Answer engine optimisation as a discipline is built on exactly this insight: structure what you say so AI systems can use it. Generic content rarely gets cited.
Monitoring and measurement. Tools like SE Ranking's AI Overview tracker, Semrush's AI tracking features, and BrightEdge's AI search monitoring let you track where your business appears in AI-generated responses and, more importantly, where it doesn't. That intelligence has real value. A business knowing it appears in responses to some queries but not others has a specific gap to address. A business with no monitoring capability can't tell the difference between a programme that's working and one that isn't.
According to Ofcom's Online Nation report, the majority of UK adults now regularly use AI tools including ChatGPT, Copilot, and Gemini. The adoption rate is why the conversation has shifted from "should we think about AI search?" to "what exactly should we be doing about it?"
What's mostly noise and old SEO repackaged
Not everything being sold as AI SEO is worth paying for. Some of it is simply not new.
AI-generated content at volume. This is the most commonly sold "AI SEO" deliverable and the one with the weakest commercial case. Producing large amounts of AI-written content at speed is something any business can now do internally using Claude or ChatGPT with a decent briefing process. Paying agency rates for content production without an underlying editorial strategy, proper quality standards, and human editorial judgement at every stage is poor value. The volume is not the problem. The commodity nature of the output is.
"AI-powered" technical audits. Many agencies run a standard crawl using Screaming Frog, Ahrefs, or Semrush, generate a standard list of technical issues, and present it as an AI SEO audit. The AI in question is the crawling tool, which has had automated features for years. A technical audit is genuinely useful work. It is not uniquely an AI SEO service, and paying a premium for it as one is not justified.
Guaranteed AI citations. Any agency guaranteeing placement in ChatGPT or Perplexity responses is either confused about how the technology works or not being straight with you. AI search does not operate on deterministic ranking logic like Google's index. The same query on the same platform can return different results on different days, for different users, in different regions. There is no position 1 to buy. A claim to the contrary is a red flag. Walk away from it.
AI SEO packages that mirror standard SEO packages. Check the deliverables list carefully. If it reads: site audit, keyword research, content calendar, link building, monthly reporting, that is a traditional SEO retainer. The work might be excellent. Some of it might be done faster with AI tools. But if it is identical to what was being sold two years ago under a different name, the buyer should know that, and the price should reflect it.
Sound familiar? It should. This pattern has played out in every major shift in digital marketing for the past two decades. The label changes. What actually works changes more slowly.
Four questions to ask before you sign with any AI SEO agency
The SERP for "ai seo services" in the UK is full of agencies with well-produced service pages. Most of them can talk confidently about AI search. Fewer can answer these questions clearly.
1. How do you measure AI citation visibility specifically?
Push for tools and methodology. A capable agency names the specific platforms it monitors: SE Ranking's AI Overview tracker, Semrush's AI tracking features, Perplexity query monitoring, direct API testing across ChatGPT and Claude. They can describe what is tracked, how often, and what a progress report actually looks like. "We monitor AI search performance" is not a sufficient answer. It means they watch it without a structured process, which means the client has no basis for judging whether the work is moving anything.
2. What's the difference between what you're proposing and your standard SEO retainer?
A capable agency answers this clearly and specifically. If the honest answer is "we use AI tools to do it faster", the buyer is being sold an AI-augmented workflow, not an AI citation strategy. Both have value, but at different price points and with different expectations. If the answer is specific, for example, "we add entity establishment, structured data work, and monthly AI citation monitoring across three platforms", that is a different conversation worth having.
3. Can you show evidence of AI visibility improvement for a client?
The case study does not need to name the client. But it should have numbers: which query, which platform, what change, over what time period. "Our clients see great results with AI SEO" is anecdote, not evidence. A result described as "a professional services firm going from zero appearances to appearing in roughly a third of monitored Perplexity responses for their target queries over six months" is evidence. Specificity is the signal that separates real capability from a good pitch.
4. What is your process when AI platform citation logic changes?
Google's algorithm updates regularly, and the SEO industry has had twenty years to build update-response processes. AI model updates can shift citation patterns faster and with less predictability, and without the advance notice Google sometimes provides. An agency with a mature AI SEO practice has a documented monitoring protocol and a process for responding when patterns shift. A newly-rebranded service page does not. Ask for it in writing.
These questions are not designed to catch anyone out. They surface the difference between agencies that have genuinely done this work and agencies that have repainted their existing offer.
What good AI SEO actually looks like in practice
Here is what a properly built programme includes, in roughly the order it gets built.
Entity clarity. Business name, services, locations, credentials, and expertise consistently structured across the website, Google Business Profile, and third-party directories. AI systems build entity models from these signals. A business appearing inconsistently across the web, with different service descriptions in different places, is harder for an AI to cite confidently. There is no coherent entity for it to reference.
Citation-worthy content. Clear definitional statements. Original data points where available. Specific named tools and frameworks. The type of content that gives an AI assistant something concrete to quote. Hedged, vague, or generic content rarely gets cited. Content that takes a clear position and uses specific examples gets cited. An editorial strategy, not just content volume, makes the difference here.
Technical AI readiness. Schema markup (FAQ, HowTo, and Organisation schema), a clean crawlable site structure, and an llms.txt file that tells AI crawlers how to read the business's content. These are table-stakes items. A good AI SEO agency treats them as a baseline, not as the headline deliverable.
A measurement layer. LLM visibility tracked across at least the three main platforms, with results reported over time so there is a basis for judging progress. Without this, a business cannot distinguish between a programme that is working and one that isn't. Monthly reporting should show movement, not just activity.
Honest timeline expectations. Traditional SEO takes three to six months to produce meaningful results. AI search visibility takes a similar run-up, with the added complexity that it cannot be A/B tested in the same way as on-page changes. Agencies promising results inside a month should be asked specifically what evidence they have. Most don't have any, because AI citation patterns are too recent to have a substantial evidence base.
Should AI SEO be your priority right now?
Not always. That is the honest answer.
AI SEO is worth making a genuine priority if you are in professional services, healthcare, B2B, or any sector where buyers research carefully before making contact. It is worth prioritising if you are already visible in Google but not appearing when a potential customer asks an AI assistant about your category. And it is worth the investment if you are losing enquiries to competitors and want to understand exactly why.
It is worth addressing foundations first if your site has significant technical SEO problems: slow load times, poor crawlability, thin content, or pages that have not been properly indexed. AI systems cannot cite a site they cannot read cleanly, just as Google cannot rank one. A proper SEO agency will surface these issues in any audit. Fixing them benefits both traditional and AI search performance at the same time.
The honest position: AI SEO is a genuine extension of search visibility and traffic strategy, not a replacement for solid fundamentals. For most businesses, the right sequence is to get the technical and content basics solid, then layer on AI citation strategy. The work is complementary, not competing.
The businesses building AI citation visibility now are ahead of those who will start in eighteen months.
That gap is real. It compounds.
Start with where you actually stand
Before committing to any AI SEO programme, it helps to know your current position. How is your business currently described across ChatGPT, Perplexity, and Google AI Overviews? Which queries do you appear in? Which queries are sending leads to your competitors instead?
That picture is the starting point for any useful conversation about strategy. Not a pitch. Not a proposal. Just the facts.
Free resource: AI Visibility Audit
We check how your business is being described or overlooked across the main AI search platforms and give you a clear picture of where you stand. It is the right starting point before any AI SEO investment, and it costs nothing to find out where you are.
Frequently asked questions about AI SEO services
How much do AI SEO services cost in the UK?
UK AI SEO retainers typically run from around £450 per month at the entry level, usually covering basic monitoring and entity work, to £5,000 or more per month for comprehensive programmes combining entity establishment, citation-worthy content production, technical AI readiness, and multi-platform monitoring. The price differential largely reflects whether the agency is delivering strategy and execution or audit and recommendations only. Packages priced identically to a standard SEO retainer from two years ago are worth scrutinising.
How long does AI SEO take to show results?
Expect a similar timeline to traditional SEO: three to six months for meaningful, measurable movement. The run-up exists because entity signals take time to build, content takes time to be indexed and associated with your entity, and AI citation patterns shift gradually rather than overnight. Agencies promising visible results inside a month are not working with a realistic evidence base. Ask what the evidence is for any specific timeline claim.
Is AI SEO the same as regular SEO?
They share foundations. A technically sound website, well-structured content, and a credible web presence are the baseline for both traditional and AI search performance. Where AI SEO differs is in the additional layer: entity establishment, structured data designed for AI ingestion, content built to earn citations, and monitoring across AI platforms rather than just Google's index. A business that has done the foundational SEO work is well-placed to build on it. One that hasn't should start there.
Can an agency guarantee AI search citations?
No. AI citation is not a deterministic ranking system. The same query can return different results across sessions, users, and dates on the same platform. Any agency guaranteeing a specific citation position is either describing something other than AI search, or making a promise it cannot keep. The right measure is directional improvement in monitored citation frequency over time, not a guaranteed placement. Any contract that includes a citation guarantee should be read very carefully.
