What are AI search optimisation services?
AI search optimisation services are the work of making a website easy for AI-powered search tools to find, understand, trust and quote. You will also see it called generative engine optimisation (GEO) or answer engine optimisation (AEO). The target is not ten blue links but the short written answer an AI system gives, and the sources it names alongside that answer.
For a UK business, the practical surfaces today are Google's AI Overviews and AI Mode, ChatGPT when it searches the web, and Perplexity. Each of them reads the open web, decides which pages answer a question well, and cites a handful. If your pages are clear, current and consistent about who you are, you have a real chance of being one of them.
What AI search optimisation is not: a trick, a secret file, or a way to jump over businesses with better answers. Google says plainly that there are no additional requirements to appear in AI Overviews or AI Mode, and no special optimisations necessary. The work is mostly doing the fundamentals of search very well, then adding a few specific steps for AI crawlers and answer extraction.
We are three freelance developers in India. Another of us leads technical SEO and data work, one of us makes the changes in your CMS or code, and the third of us runs the monthly plan and reporting. Because we build websites as well as optimise them, the recommendations in this guide are ones we implement, not just list.
How does a UK business get cited in Google AI Overviews?
Be indexed, be eligible for a snippet, and have the clearest answer on the page to the question being asked. Google's guidance on AI features says a page only needs to be indexed and eligible to be shown in Google Search with a snippet; there is no separate submission, and you do not need new machine-readable files, AI text files or markup to appear.
That makes classic technical health the gate. If a service page is excluded in Search Console, blocked by a noindex tag, or buried behind JavaScript that renders slowly, it cannot be used in an AI Overview however good the writing is. Our technical SEO audit services check exactly those blockers before any AI-specific work.
Beyond the gate, pages that tend to get cited share a few traits: a direct answer in the first sentence or two, specific facts such as prices, timings and areas covered, clear headings phrased as the questions people ask, and a business that is easy to identify as real. A plumbing firm's page that says “We fix boiler faults in Leicester, usually within two days, with call-out charges quoted up front” gives an AI system something it can quote.
Controls work the other way too. Google notes that you can limit what appears from your pages using nosnippet, data-nosnippet, max-snippet or noindex. We sometimes use data-nosnippet on sections such as internal pricing notes that you do not want summarised, while leaving the rest of the page available.
How do ChatGPT and Perplexity decide which sites to cite?
They run their own searches, fetch pages with their own crawlers, and cite pages that answer the question clearly. That makes crawler access the first check. OpenAI's documentation says sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers, while its separate GPTBot crawler is about model training, not search visibility.
Perplexity has a similar split. Its documentation describes PerplexityBot as the crawler that surfaces and links websites in Perplexity search results and says it is not used for model training. Perplexity-User visits pages when a person asks a question and, being user-initiated, generally ignores robots.txt rules.
Once access is sorted, the signals that matter look familiar. Pages that are easy to parse, specific, well structured and consistent with what other sources say about you get used more. Mentions of your business on reputable third-party sites, such as trade bodies, local news, supplier directories and review platforms, help the model confirm that you exist and do what you say.
The difference from Google is that these assistants lean heavily on a few sources per answer, so being the clearest source for a narrow question often matters more than general authority. A Leeds accountancy practice with one excellent page on “making tax digital for landlords in Yorkshire” may be cited where a national firm with a vague page is not.
What does GEO work involve beyond normal SEO?
Most of AI search optimisation is SEO done carefully; the extra parts are about access, identity and extractability. Here is what we add on top of a normal SEO programme for UK clients, in roughly the order we tackle it.
- AI crawler policy. A robots.txt that treats search crawlers (OAI-SearchBot, PerplexityBot) and training crawlers (GPTBot, Google-Extended) as separate decisions.
- Entity consolidation. One consistent name, description, address, service list and founder information across your site and main listings.
- Schema that states facts. Organization or LocalBusiness, Service and Person markup, linked with sameAs to profiles you control.
- Answer-first rewrites. Service and FAQ pages restructured so each section opens with a direct, self-contained answer.
- Evidence on the page. Named sources for claims, clear dates, pricing ranges where you are comfortable, and author details.
- Prompt testing. A fixed list of buyer questions checked monthly across assistants, with citations recorded.
- AI referral measurement. A GA4 channel for AI assistants so you can see visits and conversions, not just impressions.
Notice what is missing: keyword-stuffed pages written for machines, invisible text aimed at AI models, and mass-produced articles. Those tactics are more likely to damage trust with both people and search systems than to earn citations.
Entity and brand signals: helping AI systems know exactly who you are
An entity is a thing a search system can identify without doubt: your business, your founder, your location, your services. AI answers are built from confident facts, so a business whose name, address and description differ across the web is harder to cite than one that tells the same story everywhere.
The work starts on your own site. An about page that states when and where you trade, who runs the business and what you specialise in. Consistent footer details. Team or author pages for the people whose expertise appears in your content. Then we check the outside picture: Google Business Profile, Bing Places, trade body listings, Yell and sector directories, social profiles and any Wikipedia or Wikidata presence if you genuinely qualify for one.
Consistency
Same trading name, same phone number, same address format and same one-line description. Small differences like “Ltd” versus no “Ltd” matter less than contradictory service lists or old addresses.
Corroboration
Genuine mentions on independent sites: a trade association member directory, a local newspaper piece, a supplier's case study page. We never buy fake mentions or reviews.
Connections
Schema sameAs links from your site to the profiles you control, so the pieces join up in one place.
For professional services this overlaps with reputation. Our work for firms that rely on trust, such as solicitors and accountants, leans heavily on author and practitioner signals for exactly this reason.
Which schema markup helps AI search optimisation?
Schema helps by stating facts in a structured way that matches what the page already says; it does not force a citation. Google says no special markup is needed for AI features, so we add schema to make your business easier to understand, not because it unlocks a hidden placement.
The types we use most for UK clients are Organization or the relevant LocalBusiness subtype, Service for each main service, Person for founders and authors, Article for guides, BreadcrumbList for structure and, on product pages, Product with offers. FAQ content we write for people first; whether FAQ markup produces any visible result in Google changes over time, so we treat it as optional.
- Keep schema identical in meaning to the visible page. Markup that claims things the page does not say is a risk, not a boost.
- Use one Organization entity with a stable @id, referenced from every page, rather than dozens of slightly different copies.
- Add areaServed for service-area businesses, using the real towns and counties you cover.
- Validate with the Schema.org validator and Google's Rich Results Test after every template change.
- Review schema whenever prices, services or staff change, as part of monthly maintenance.
On WordPress we usually extend an existing SEO plugin's output rather than stacking a second plugin; on custom and Next.js sites we generate JSON-LD from the same data that renders the page, so the two cannot drift apart.
robots.txt and AI crawlers: what should a UK business allow?
Allow the crawlers that power AI search answers, and make a separate, deliberate choice about the ones used for model training. Blocking every AI-related bot in one go is a common mistake that quietly removes a site from ChatGPT search and Perplexity results.
The key distinction is between search crawlers and training crawlers. OpenAI's OAI-SearchBot surfaces sites in ChatGPT search; GPTBot collects training data. PerplexityBot surfaces sites in Perplexity. Google-Extended is a product token that, according to Google, controls whether content Google crawls may be used for training future Gemini models and for grounding in Gemini apps, and Google states it does not affect inclusion or ranking in Google Search.
Many UK businesses decide to allow search crawlers and are relaxed about training; publishers and firms with proprietary content often block training while allowing search. Both are reasonable. What is not reasonable is not knowing which you have chosen. The crawler table further down this page sets out each bot, its purpose and what blocking it does.
We also check the layers above robots.txt. Some hosting firewalls and bot-protection services block AI crawlers by default, which overrides whatever robots.txt says. We look at server logs to confirm the crawlers you allow can actually fetch pages, and that they are not hitting error or challenge pages.
Is llms.txt worth adding as part of AI search optimisation?
It is cheap to add and harmless, but do not expect it to move Google. llms.txt is a proposal first published by Jeremy Howard in September 2024: a Markdown file at the root of a site with a heading, a short summary and curated links to the pages that best explain the site. The idea is to give AI tools a clean map without navigation and page clutter.
Google's own AI features guidance says you do not need to create new machine-readable files or AI text files to appear in AI Overviews or AI Mode. So for Google, llms.txt is not a lever. Other tools and agents may read it, and for businesses with documentation, a product catalogue or detailed service information, a tidy llms.txt can help those tools find the right pages.
Our approach: we add an llms.txt when it takes less than an hour and the site has a clear set of key pages, keep it in sync with the sitemap, and never treat it as a substitute for crawlable, well-written pages. If a provider sells llms.txt as the core of AI search optimisation services, treat that as a warning sign.
Answer-first page structure: writing pages that AI engines quote
Put the answer in the first one or two sentences of every section, then give the detail. AI systems extract passages, and a passage that answers the question on its own is far easier to quote than one that builds up to the point after three paragraphs of introduction.
On a UK service business site, the pages worth restructuring first are the ones people ask assistants about: pricing, “how long does it take”, “do I need”, comparisons between options, and area coverage. We rewrite headings as real questions, open each section with a self-contained answer, add a short table where a comparison exists, and keep specific numbers visible in text rather than hidden in images.
Before
“At Smith Roofing we pride ourselves on quality and have served the community for many years. Get in touch to discuss your needs.” Nothing here can be quoted as an answer.
After
“A flat roof replacement on a typical terraced house in Sheffield takes two to three days. We quote after a free survey and explain whether EPDM or GRP suits your roof.” Specific, local, quotable.
Structure
Question heading, direct answer, supporting detail, a list or table where it helps, then a link to the next relevant page or your contact route.
This style also helps human readers, which is the point: Google's guidance keeps returning to helpful, people-first content. For more on how content structure fits search generally, see our UK SEO packages page.
How do you measure AI search referrals in GA4?
Create a custom channel group in GA4 that pulls AI assistant referrals into their own channel, then watch visits, engagement and conversions from that channel alongside organic search. Without it, AI referrals sit scattered inside the general Referral channel and are easy to miss.
Google's Analytics help explains that custom channel groups are rule-based, can match on Source using regular expressions, and can be applied to reports retroactively. Standard GA4 properties allow two custom channel groups, each with up to 50 channels. We build one that copies the default grouping and adds an “AI assistants” channel matching sources such as chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com, placed above Referral so it catches those visits first.
Google AI features are measured differently. Google says sites appearing in AI Overviews and AI Mode are included in the overall search traffic in Search Console, in the Performance report under the Web search type, rather than broken out separately. So for Google we track impressions and clicks for the queries and pages we are targeting, and pair that with prompt testing.
- GA4: AI assistants channel for visits, key events and landing pages.
- Search Console: impressions and clicks for target queries, which include AI feature appearances.
- Monthly prompt log: which assistants cited you, for which questions, and which page they linked.
- Enquiry forms and call tracking: ask “how did you hear about us?” with an AI assistant option.
Be realistic about the limits. Some assistants do not always pass a referrer, and many people read an AI answer, remember your name and search for you directly. Branded search growth in Search Console is often the clearest indirect signal.
AI search optimisation for UK local service businesses
For a plumber, dentist, solicitor or salon, local SEO and AI search optimisation are the same project viewed from two sides. People ask assistants questions like “who can fix a leaking radiator in Bath this week” or “best family solicitor near Chelmsford”, and the answers draw on the same signals that drive Google's local results: a complete Google Business Profile, consistent listings, reviews and clear location pages.
Our work for local businesses starts with Google Business Profile: correct categories, services with descriptions, service areas that match reality, opening hours, photos and regular updates. Then we check that your website's location and service pages say the same things, in plain answer-first language, with the towns and postcodes you really cover.
Reviews matter to both people and AI summaries. We never write, buy or incentivise reviews; we can set up a simple follow-up message that asks happy customers to leave one, and we reply templates for you to personalise. For multi-location businesses, each branch needs its own profile and page with genuinely local detail, not copies with the town name swapped.
If you want the local side covered in depth, our local SEO services page explains it, and trade-specific pages such as web design for plumbers show how the website foundations fit together.
How much do AI search optimisation services cost in the UK?
Quotes vary widely, because “AI search optimisation” can mean a one-page report or months of implemented changes. With us, AI search optimisation services start from US$150/mo a month, usually as part of an SEO retainer, with the first month weighted towards audit and set-up.
What moves the price is the amount of implementation, not the label. A ten-page local business site with sound foundations needs mostly schema, entity tidying, a robots.txt review, GA4 set-up and a few rewrites. A 400-page ecommerce or professional services site with duplicate content, slow templates and inconsistent listings needs far more.
- Size of the site and number of templates that need schema.
- State of technical health: indexing problems must be fixed first.
- Number of locations and listings to align.
- How many pages need answer-first rewrites, and who writes the copy.
- Whether GA4 and Search Console are already set up correctly.
- Platform: WordPress, Shopify, Webflow or custom code change the effort.
Retail sites have extra product and category work; our ecommerce SEO page covers that side. You receive an itemised quote in about two working days, and nothing is billed until you approve it in writing.
How long before AI search optimisation shows results?
Expect technical fixes to take effect within weeks of recrawling, and citation changes to build over a few months. There is no fixed timetable, because it depends on how often crawlers revisit your pages, how competitive the questions are, and how much of the foundation needed fixing.
A typical first quarter with us looks like this. Month one: audit, robots.txt and firewall checks, schema, GA4 channel, entity clean-up and the first rewrites of your most important pages. Month two: more rewrites, listings alignment, internal linking and the first full round of prompt testing against the baseline. Month three: review what is being cited, strengthen pages that are close, and plan new content for gaps.
We report honestly when something is not working. If an assistant keeps citing a competitor for a question, we look at what their page answers that yours does not, rather than blaming the algorithm. And we will never tell you a citation is guaranteed, because no one controls those systems except the companies that run them.
Red flags when buying AI search optimisation services
Be wary of anyone who promises guaranteed placement in ChatGPT or AI Overviews, or who sells a single file or plugin as the whole solution. AI search optimisation services are only as good as the changes they make to your site and the accuracy of the picture they build of your business.
- Guarantees of citations, positions or specific traffic numbers.
- Hidden text or instructions aimed at AI models; this is manipulative and risky.
- Mass-produced AI articles published without expert review.
- Fake reviews, paid reviews or invented testimonials.
- A report with no implementation, when your site needs changes made.
- No measurement plan, or measurement only in the provider's own tool.
- Blocking or allowing crawlers without explaining the trade-off.
A good provider explains what they will change, why, how you will measure it and what they cannot control. That is the standard we hold ourselves to.
Working with an AI search optimisation team in India from the UK
Our working day overlaps the UK business day from late morning, so a call at 11:00 UK time sits comfortably in our afternoon. Changes agreed in the afternoon are often live by your next morning, which suits SEO work, where most tasks are made and then left to be recrawled.
Most clients speak to us once a month on a video call to review the report, with WhatsApp for quick questions in between. We work in English; the copy we write for your pages is in UK English, and you approve every change to wording before it goes live.
Access
Delegated roles only: Search Console, GA4, Google Business Profile manager access, and a separate CMS login. You stay owner of everything and can remove us at any time.
Payments and contracts
Quotes are in USD; you pay in USD or GBP by Wise, bank wire or PayPal. Scope sits in the approved quote and our published terms. Invoices come from India, and your accountant can advise on your side.
Your first two weeks
Days one to five: access set up, baseline prompts tested and recorded, crawl and robots.txt review. Days six to ten: audit findings shared with priorities. By day fourteen: GA4 channel live and the first schema and page changes made.
Worked example: a hypothetical Nottingham heating engineer
This is an invented scenario to show how AI search optimisation services play out for a small UK business. Say a family-run heating engineer in Nottingham has a ten-page WordPress site, a Google Business Profile and steady local rankings, but when the owner asks ChatGPT and Perplexity about boiler servicing in Nottingham, two larger competitors are named and he is not.
The audit finds three things. The site's security plugin blocks unfamiliar bots, including OAI-SearchBot and PerplexityBot. The service pages open with a paragraph about the company's values, and prices and response times appear only in a PDF. The business name is written three different ways across the site, the profile and two directories.
The changes: allow search crawlers in both robots.txt and the plugin, keep training crawlers blocked at the owner's request, rewrite the five service pages answer-first with visible price ranges and areas covered, add LocalBusiness and Service schema, fix the name everywhere, and set up an AI assistants channel in GA4.
What happens next is not guaranteed, and we would not promise it. What the owner does have is a site that AI crawlers can reach, pages that answer the questions customers actually ask, and a monthly log showing which assistants cite him and which pages they send visitors to.
AI search optimisation checklist for UK websites
Use this as a quick self-check. If you can tick every line, your foundations for AI search are sound; if not, these are the places to start.
- Key pages are indexed in Search Console and not blocked by noindex or robots.txt.
- robots.txt makes separate, deliberate choices for OAI-SearchBot, PerplexityBot, GPTBot and Google-Extended.
- Your firewall or bot-protection service is not silently blocking crawlers you allow.
- Business name, address, phone and description match across site and listings.
- Organization or LocalBusiness schema is valid and matches the visible page.
- Service pages answer the main question in the first two sentences.
- Prices, timings and areas covered appear as text, not only in images or PDFs.
- GA4 has a channel that separates AI assistant referrals.
- You test a fixed list of buyer questions across assistants every month.
Want us to run through it for you? Send your domain on WhatsApp or through the contact page, and you will get a scoped quote in about two working days.