AI SEO Agency for UK and US Businesses

Search is no longer ten blue links. For a growing share of queries, Google shows an AI Overview at the top of the search results page: a generated summary with a handful of cited sources, above the organic results and often above the ads. AI Mode goes further, turning the search into a conversation where the user asks follow-up questions and the engine keeps answering, breaking a complex question into several searches behind the scenes and stitching the answers together. Outside Google, AI assistants and answer engines are used every day for research, comparisons and recommendations, from “which CRM suits a ten-person firm” to “who is a good emergency plumber near me”. Voice search and multimodal search, where the query is a photo or a spoken question, feed the same systems.

The commercial effect is simple to describe and hard to live with. Zero-click searches rise, because the answer sits on the results page. Click-through rates on the traditional organic results fall on queries that trigger an AI Overview, especially informational ones. The clicks that remain are more qualified, because the person has already read a summary and is choosing who to contact. And a new kind of visibility appears that never existed before: being the brand the AI mentions, cites or recommends, even when nobody clicks anything at all. AI visibility, brand mentions and citations in AI-generated answers are now part of search performance, and they have to be measured, built and reported alongside keyword rankings and organic traffic.

Before deciding how much of the budget belongs here, run the checks in the AI visibility scan guide: it shows whether AI answers currently mention, cite or recommend the business, and who appears instead.

Underneath the interfaces, the ranking systems have been machine learning for years. Google’s RankBrain began interpreting unfamiliar queries with machine learning in 2015, neural matching and the BERT language model followed, and the modern retrieval systems behind AI Overviews and AI Mode use large language models to understand intent, expand a query into sub-queries and pick passages worth citing. Search engines and AI systems now read for meaning, entities and relationships rather than matching strings of keywords. That is why an AI SEO strategy is built around topical authority, entity clarity and answer-ready content instead of keyword density, and why the old tricks that once moved rankings now do nothing at all.

The industry has produced an alphabet of acronyms in two years and most of them describe the same practice from a slightly different angle. They matter only because they show where the answer is being produced, and therefore what the page has to do to be chosen. Search engine optimisation is the discipline of ranking pages in search engine results pages. Generative engine optimisation is the practice of optimising content so that generative AI search engines, the ones that write an answer rather than list results, use your pages as a source and mention your brand. Answer engine optimisation focuses on being the direct answer: featured snippets, People Also Ask, voice assistants and the concise responses AI systems give to a specific question. LLM optimisation, sometimes called LLMO or AI search optimisation, is the broadest term and covers visibility inside the large language models themselves, including the chat assistants that have no search results page at all.

The practical difference is in the unit of competition. Traditional SEO competes page against page for a position. Generative engine optimisation competes passage against passage for a citation, and brand against brand for a mention. A page can rank third in Google and be the first source cited in the AI Overview above it, because the model found a clean, quotable paragraph that answered the question directly. Equally, a page can rank first and be ignored by the AI because the answer is buried in marketing copy. Writing for both at once is a craft, and it is mostly the craft of being clear.

Nobody outside the companies that build these systems knows the exact weighting, and anyone who tells you otherwise is selling something. What is well documented, from the published research on generative engine optimisation, from the patents and developer documentation behind AI Overviews, and from watching thousands of prompts over time, is a consistent set of signals. AI answers are built by retrieving a set of candidate pages, usually from a conventional search index, then reading passages from them and synthesising a response with citations. So the first filter is the old one: you have to be retrievable, which means indexed, relevant and reasonably authoritative. The second filter is new: your passage has to be the clearest, most self-contained, most trustworthy answer among the candidates.

The method is the same for every sector; the prompts, the sources and the pages are not. A private clinic is recommended by AI assistants on the strength of reviews, credentials and clear treatment pages. A B2B software company is recommended on comparison content, integrations and what the review platforms say. A plumber is recommended on proximity, availability and a Google Business Profile that is complete. The work here starts from what your customers actually ask, then builds the pages and the corroboration that answer it.

Local search has moved into AI answers faster than almost anything else. “Best accountant near me”, “emergency electrician in Manchester open now”, “dentist in Brooklyn that takes new patients”: these are questions AI assistants and AI Mode answer with a short list of named businesses, drawn from Google Business Profile data, reviews, local citations and the location pages on your website. For a business in London, Birmingham, Leeds, Bristol, Glasgow or Edinburgh, or in New York, Los Angeles, Chicago, Houston or Miami, the local AI search work is the same: a complete, accurate business profile, consistent name, address and phone number everywhere, review generation handled honestly in your own accounts, and a page for every service in every place you genuinely serve.

National and international businesses have a different problem. They need to be the answer to a question with no place attached, which means topical authority, comparison content and third-party corroboration at a scale a single location page cannot provide. Many businesses need both: a national brand with local branches, or a UK company selling into the US. The keyword and prompt research separates those intents at the start, and the site structure follows, with hub pages for the topic and spoke pages for locations, products and questions. International SEO details, hreflang, currency and spelling, are handled in the technical foundations rather than left to chance.

One honest note on geography. BoltClicks is a remote practice working with UK and US businesses on UK hours. There is no London office, no Manchester office and no New York office, and none is claimed. What you get instead is one specialist who knows the UK and US search markets, works inside your own accounts and reports in writing every week.

AI search work is not sold as a separate, mysterious add-on. It is part of the SEO channel, because the technical foundations, the content and the authority building are shared, and charging twice for the same page would be dishonest. The retainers are flat monthly fees agreed in writing before work starts, paid a month in advance, cancellable with one month’s notice, and built inside accounts you own so that leaving costs you nothing. Nothing is priced as a percentage of ad spend or of revenue.

What is never included: advertising budget, which stays on your own card; hosting and domains, which stay in your own account; and paid tools you specifically ask for beyond the standard set covered here. The full breakdown of retainers, packages, payment terms, notice periods and who owns what is on the pricing page, and if a retainer is not the right fit yet, you will be told so rather than sold one.

For comparison, many AI SEO agencies that publish prices quote retainers from around £1,500 to over £10,000 a month, often on six- or twelve-month contracts, and often with an account manager between you and the people doing the work. The difference here is not that the work is smaller; it is that one specialist does it, in your accounts, with the overhead of an agency removed.

“AI SEO” also describes how the work is done, not only what it targets, and it is worth being precise because the two get confused on purpose. AI tools are used here for the tasks where machines are genuinely better: clustering thousands of keywords and prompts by intent, analysing server logs and crawl data, extracting the entities and questions competitors’ pages cover, spotting content gaps across a topic, drafting schema markup, checking internal link structures and summarising long documents. That is data science applied to search, and it makes the research faster and more thorough than a person with a spreadsheet.

What AI tools do not do here is write your website. Search engines say plainly that helpful, original, people-first content is rewarded however it is produced and that mass-produced, unoriginal content is not, and AI answer systems are even more allergic to generic text because they have already read the millions of pages it was averaged from. Every page published for a client is written by a person with knowledge of the business, uses AI for research and structure, and is fact-checked and edited before it goes live. Generated summaries of other people’s summaries do not get cited; original explanations, data, opinions and examples do. That is the editorial standard, and it is also the honest reason AI SEO cannot be fully automated.

A content strategy for AI search starts from the shape of the answer, not the shape of the page. When a model answers “what does an AI SEO agency do”, it wants a definition. When it answers “how much does SEO cost in the UK”, it wants a number with a range and a reason. When it answers “which is better for a small law firm, SEO or paid search”, it wants a comparison with criteria. The pages that get cited are the ones that hand the model the exact structure it needs, in clean HTML, near the top, written by someone who clearly knows the subject. Everything else on the page, the storytelling, the brand voice, the calls to action, is for the human who clicks through, and it still matters, but it comes second.

Format is only half of it. Each page is also mapped to an entity and an intent: one clear subject, one clear question set, headings phrased as the questions people ask, a short answer under each heading, then depth. Internal links connect the page to its hub and its neighbours so the cluster reads as a body of knowledge. Author pages, an about page that names real people, publication and update dates, sources cited in the text and a consistent tone of voice all add the trust signals that decide between two equally clear pages. Refreshing existing content, updating figures and dates and expanding thin sections, is scheduled every quarter, because freshness is one of the few signals that is both cheap and unambiguous.

Business owners rarely ask “should I do GEO”; they ask where the next pound or dollar of marketing budget should go. The honest answer depends on timing and margin. Paid search and paid social buy visibility immediately and stop the moment the budget stops. Organic search, including AI visibility, takes months to build and then compounds, costing nothing per click and holding through quiet periods. Most businesses that are growing well run both: ads for the queries that must convert this month, SEO and AI search for the presence that makes every other channel cheaper. The digital marketing plan here treats them as one funnel, with the ad data feeding keyword research and the organic pages feeding landing page tests.

The order that usually works: fix the technical foundations and tracking first, because every channel depends on them; run a small paid search campaign to learn which queries and messages convert while the organic work builds; then shift budget towards organic and AI visibility as the pages start to rank and be cited. Businesses with thin margins or long sales cycles lean earlier into organic; businesses with a launch date or a seasonal peak lean earlier into paid. The weekly report shows the cost per lead by channel so the decision is made from data, not preference.

The phrase “AI SEO agency” is new enough that anyone can put it on a website, and plenty have. Some are excellent search agencies that have added generative engine optimisation to a strong technical and content practice. Some are content farms with a new label. A few questions separate them quickly, and they are the same questions you should ask here.

Search engines and language models do not store your business as a string of words; they store it as an entity, a thing with a type, attributes and relationships to other things. A dental practice in Leeds is an entity of type LocalBusiness and Dentist, with an address, opening hours, practitioners, treatments and reviews, related to the entity Leeds, to the regulator and to the people who work there. When those facts are consistent everywhere the machine looks, the knowledge graph fills in and the business becomes recommendable. When they conflict, confidence collapses and the machine reaches for a competitor whose facts it can verify.

Entity SEO is the deliberate work of making those facts clear. It starts with an entity home, usually your about page, which states in plain language what the business is, where it operates, who founded it, what it does and does not do, and links out to every official profile. Organization schema mirrors that page in machine-readable form, with sameAs links to your social profiles, directory listings and any knowledge panel. Every service, product, location and person gets a consistent name that never varies between the menu, the page title, the schema and the Google Business Profile. Then the same facts are seeded outward: directories, review platforms, industry bodies, press mentions, partner pages and social profiles, all carrying the same name, address, phone number and description. Over time the mentions corroborate each other, the knowledge panel appears or stabilises, and AI systems begin to treat the brand as a known quantity rather than a guess.

The same discipline applies to the topics you want to own. Topical authority is entity coverage in depth: a cluster of pages that covers a subject’s sub-entities, the related questions, the terminology, the comparisons and the edge cases, interlinked and updated. Content gap analysis against the sources the AI already cites for your topic shows exactly which entities and questions are missing from your site, and that list becomes the content plan. Semantic SEO, natural language processing and vector search are the machinery behind this; the practice is simpler than the vocabulary: cover the topic completely, name things consistently, and make every page easy to understand.

Google’s AI Overviews and AI Mode deserve their own section because they sit on the search results page most of your customers still use every day, and because they behave differently from the standalone assistants. Both are built on Google’s index and its core ranking systems, so the pages they cite are overwhelmingly pages that already rank somewhere in the top results for the query or for one of the sub-queries the system generates through what Google calls query fan-out. There is no separate submission, no special markup that guarantees inclusion and no way to opt in; the way to appear is to be a strong, relevant, clearly written result for the underlying searches, and to make the relevant passage easy to extract.

What moves the needle in practice: pages that answer the exact sub-questions a broad query breaks into, since AI Mode may run half a dozen searches to answer one prompt; concise passages with the answer in the first sentence; structured data so the type of page is unambiguous; freshness for anything with a date on it; and enough authority that Google is comfortable citing you for a topic. Featured snippets and People Also Ask remain excellent predictors, because the same passage-ranking systems feed all three. Google has also confirmed that impressions and clicks from AI Overviews and AI Mode are counted inside the normal Search Console performance report, so your Google visibility is measured in one place even though the citation itself is not labelled.

Measuring beyond that is manual or tool-assisted. A defined set of commercial and informational prompts is checked on a schedule, the presence of an AI Overview, the sources cited and your position among them are logged, and the trend is reported weekly. Where AI Overviews cause click-through rates to fall on informational queries, the response is to re-balance the content plan towards commercial and comparison intent, where clicks and mentions still convert, and to make sure the brand is named inside the answer rather than only linked below it.

Search and answers

AI Overviews are Google’s generated summaries at the top of some search results, with cited sources. AI Mode is Google’s conversational search experience that answers a prompt, accepts follow-ups and runs multiple sub-searches behind the scenes, a process known as query fan-out. An answer engine is any system that returns a direct answer rather than a list of results; AI assistants and chatbots are answer engines with a conversational interface. Generative AI search or generative search is the broad category. Zero-click search is a search that ends on the results page without a visit to any website. A featured snippet is the highlighted answer box in traditional results; People Also Ask is the expandable list of related questions beneath it. Voice search is search by spoken query, usually answered by a single result. Multimodal search accepts images, voice and text together.

Optimisation disciplines

Search engine optimisation, SEO, is improving a website’s visibility in organic search results. Generative engine optimisation, GEO, is optimising content and brand signals to be cited and mentioned in AI-generated answers. Answer engine optimisation, AEO, is optimising to be the direct answer to a specific question. LLM optimisation, LLMO or AI search optimisation, is visibility inside large language models and the assistants built on them. Entity SEO or semantic SEO is optimising the clarity and consistency of the things a page is about rather than the keywords it contains. Technical SEO covers crawlability, indexation, rendering, speed and structure. On-page SEO is the optimisation of an individual page’s content and HTML. Off-page SEO is authority built outside the site: links, mentions, reviews, citations. Local SEO is visibility for location-based searches, including the map pack and Google Business Profile. Digital PR is earning coverage, links and mentions from publications and communities through newsworthy content.

Signals and measurement

A citation is a source link in an AI answer; a mention is your brand named in the answer with or without a link. Share of voice is your share of the mentions for a set of prompts. AI visibility is the overall presence of a brand across AI answers, tracked through prompt monitoring. AI referral traffic is visits from links inside AI answers, measured in analytics. E-E-A-T stands for experience, expertise, authoritativeness and trustworthiness, the qualities Google’s search quality rater guidelines describe. Topical authority is the demonstrated depth of a site on a subject. A knowledge graph is the database of entities and relationships a search engine or model uses to understand the world; a knowledge panel is its visible box for a brand or person. Structured data or schema markup is the machine-readable code describing a page’s content. llms.txt is a plain-text file, proposed as a convention, that summarises a site for language models. Core Web Vitals are Google’s page experience metrics: Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift. Crawl budget is the attention a crawler gives a site. Query fan-out is the expansion of one prompt into many searches. Retrieval-augmented generation, RAG, is the technique of fetching documents before generating an answer, which is why being retrievable matters. A large language model, LLM, is the model that reads and writes the text; natural language processing, NLP, is the field behind it; a vector or embedding is the numerical representation of meaning these systems compare.

Many businesses that need AI SEO already have a marketing manager, an internal team, a web design or web development agency, a PR agency or a paid media partner. The work here slots in beside them rather than replacing them. For an in-house marketing team, that usually means SEO consultancy and a clear roadmap: the audit, the keyword and prompt research, the technical tickets written for your developers, the content briefs for your writers, the schema and structured data specifications, and a monthly review of what the search data says. For a business with a web development agency mid-way through a redesign or a website migration, it means protecting rankings and AI visibility through the launch: redirect maps, crawl checks before and after, rendering tests and a watch on Search Console for the weeks that follow. For a founder or managing director with no marketing team at all, it means the whole thing done for you, explained in plain English every week.

SEO training is available for internal teams who want to own the day-to-day: how to read the weekly reports, how to structure a page so it can be cited, how to run the prompt checks, how to keep the Google Business Profile and the directories consistent, and which SEO tools are worth paying for and which are not. Everything is documented so that if the relationship ends, your team keeps the process, not just the results. That is the difference between an SEO consultant who builds capability and an agency that builds dependency.

The search landscape changes several times a year: a core update, a new AI feature, a change in how AI Overviews trigger, a new assistant with a new set of source preferences. A strategy that depends on one tactic breaks every time. The approach to SEO here is deliberately boring in its foundations and adaptive in its details: satisfy user intent completely, build trust and authority that third parties confirm, keep the technical foundation flawless, and measure real business outcomes. Those principles have survived every ranking change of the last decade, and they are precisely what the AI systems reward, because the models were trained to prefer what humans found useful.

In practice that means keyword targeting that follows search behaviour rather than fighting it, with topic clusters instead of isolated pages; competitor analysis repeated quarterly so the plan reacts to what the AI is citing right now; entity signals and E-E-A-T signals maintained as a routine, not a one-off; citation building and review generation as a monthly habit; conversion rate optimisation on the pages that receive the traffic, because a visibility gain that does not convert is wasted; and performance tracking that separates organic clicks, AI referral traffic, branded search and paid media, so the return on investment of each channel is visible. Brand awareness, social media activity and digital PR feed the same machine: the more trusted sources talk about the business, the more the models do too. None of this is a trick, which is why it keeps working.

Is this a digital marketing agency or an SEO specialist?

A specialist. The marketing services here are search engine optimisation including AI search visibility, Google Ads and Meta Ads management, and the websites and landing pages that make them convert, run by one experienced SEO specialist rather than a digital marketing agency with an SEO team, a design team and an account manager between you and the work. If you need social media management, brand design or video production, you will be pointed to people who do that well rather than sold a bundle.

Will AI search replace traditional search results?

Not in any timeframe a business plan should worry about. Traditional search results still carry the overwhelming majority of commercial clicks, AI Overviews appear on a minority of queries, and the AI features themselves are built on the traditional index. What has changed is that the top of the page is now often an answer rather than a list, so the strategy has to cover both the ranking and the citation. Businesses that treat AI search as an extension of SEO rather than a replacement are the ones getting the mentions.

Does this work for a new business with no domain authority?

Yes, more slowly. A new domain needs technical foundations, a tight topic cluster, consistent entity signals and the first wave of reviews and mentions before AI systems trust it. The advantage a new business has is focus: a narrow, well-covered topic with clear answers and honest pricing gets cited before a large generic site does. The audit tells you honestly how long that is likely to take in your market, and whether paid search should carry the first months.

What tools are used, and do I need to buy any?

Google Search Console, Google Analytics 4, Google Tag Manager and Google Business Profile, all free and all in your name, plus the crawling, rank tracking, keyword research and AI visibility tracking tools that a dedicated SEO practice pays for. You buy nothing extra. If a specialist tool would genuinely help your team, it is recommended with the reasons and the price, and the decision is yours.

How do you handle a website redesign or migration without losing AI visibility?

With a checklist and a calendar. Before launch: a full crawl of the old site, a redirect map for every URL that matters, the schema and content carried over intact, rendering tests on the new build and crawler access confirmed. At launch: sitemaps resubmitted, Search Console watched daily. After launch: rankings, citations and AI referral traffic compared to the baseline for eight weeks, with any drop investigated the day it appears. Migrations are where most businesses quietly lose years of authority, and it is avoidable.

Can AI SEO help with lead generation, not just visibility?

Visibility that does not become enquiries is not the goal. Every page built for AI search is also built to convert: a clear call to action, a visible phone number, a form that works on a phone, pricing where possible and trust signals near the decision. Conversion tracking is set up in month one so that leads, calls, bookings and sales are attributed to organic and AI referral sources, and the weekly report leads with those numbers, not with mention counts.

What does success look like after a year?

A website that ranks for the commercial searches your customers make, a brand that is named in AI answers for the prompts that matter in your market, a Google Business Profile that wins local recommendations, a steady rise in branded search, and a cost per lead from organic and AI search that is lower than any paid channel, all documented week by week in reports you can hand to anyone. And a process your own team understands well enough to keep.