Guide 20 min read
Introduction: Why AI search is stealing your clicks
Driving traffic and enquiries from Google has got a lot harder for small businesses.
A year or two ago, your prospective customers would have opened Google and started scrolling down the page, looking for some relevant-looking links to click. If you’d done a decent job of your SEO, your website would appear in these results, and you would be rewarded with traffic which would hopefully convert to paying customers.
These days, not only are a growing number of people turning to ChatGPT and other Large Language Models (LLMs) to answer their questions, those who remain on Google are increasingly presented with an AI Overview (AIO) before they can scroll down to the search listings to find the link to your website.
Even before AIOs, 60% of Google users ended their journey without clicking a single link, meaning all that work you did to rank for more keywords might not actually bring you more traffic.
As you’d expect, the zero-click situation gets even worse when AIOs are present. Just take a look at this analysis from Semrush, showing the increase in zero-click for keywords when an AI Overview appears above the search results:
The stark truth is that if your business isn’t appearing in these new AI-powered search features, it’s harder than ever to drive traffic to your website.
Ok, so what do we do about it?
There’s only one answer: you need a brand presence across AI Overviews and LLMs like ChatGPT and Gemini.
AI search optimisation – sometimes called Generative Engine Optimisation (GEO) or Answer Engine Optimisation (AEO) – is about making your business visible and recommendable within AI tools. You don’t need a separate version of your site - a clean HTML website is perfectly accessible. There may be some changes you need to make to the way your content is organised: not because AI can’t access it, but because good attention to formatting, structure and detail make your content more attractive as a citation source.
1. What’s the big deal about AI search?
First, a bit of a stage-setter: Estimates suggest that somewhere between 2% and 10% of search-like queries are now going through LLMs like ChatGPT and Perplexity, rather than through Google or Bing.
AI search is growing fast, particularly for the kind of queries where someone wants a recommendation or a considered answer rather than just a list of links. “What are the best accounting software options for a sole trader?” is a conversation opener, and LLMs are precisely designed to engage users in conversations.
Google’s AI Overviews appeared for around 40% of searches by the end of 2025, but this is expected to increase sharply to 75% by 2028, according to McKinsey. And, as you may have noticed yourself when you click on an AIO, you’re then invited to go even further down the rabbit hole by switching your search into AI Mode.
What is AI mode?
This is Google’s own version of an LLM, where you’re invited to continue refining your AI Overview response through a chat interface, pulling you even deeper into the AI interface and even further away from the ‘blue links’ that may or may not feature your website.
That 60% zero-click rate for Google I mentioned earlier shoots up to a whopping 93% when people enter AI Mode! The only way to earn visibility for your business within AI Mode is to be mentioned as part of the conversation.
SEO vs GEO: What's actually different
The relationship between traditional SEO and GEO/ AEO can be confusing.
The best way to think of it is to consider traditional SEO as the foundation, and GEO as a layer that sits on top of that foundation. The first thing to say is that there is no shiny box of magic GEO tricks.
Instead, it’s about revising your strategy to include activities that you might not have prioritised in the past.
In traditional SEO, the website is the centre of the universe. You would do certain things outside of the website, largely to build links, but your efforts would focus on building a technically sound, content-rich site. With GEO/ AEO, you’re optimising everywhere in the sense that LLMs and AI Overviews look for signals from around the internet that your brand is a trusted source of information. The digital PR piece, while not new, becomes much more important for GEO.
2. How does AI search actually work?
You don’t need a computer science degree for this bit, but a basic understanding of what’s happening behind the scenes will make the practical advice later in this guide much easier to grasp and implement.
Let’s explore the inner workings of both LLMs and AI Overviews…
How LLMs generate their answers
LLMs are trained on enormous volumes of text – web pages, articles, books, forums, social media posts – scraped from the internet over a period of time. Through that training, they develop a model of the world: what things are, how they relate to each other, which sources are credible, which businesses exist and what they do. When a user asks a question, the LLM draws on that model to generate a response.
Depending on the prompt, LLMs will sometimes deploy Retrieval Augmented Generation (RAG), or live web search. They do this to cross-reference what they already know with up-to-date information. So, if you ask ChatGPT for the top five sights to visit in Paris, it probably knows how to answer that purely from the training data it absorbed. If, however, your prompt included a request to share details of upcoming events in Paris, the LLM wouldn’t be able to answer that accurately without RAG. Which brings us to the next important point …
There’s more than one way to show up in LLM outputs – in fact, there are three ways:
1) A mention
Pretty much as it sounds - you get mentioned within the generated text but you don’t get a link to your website.
2) A citation
This is where a clickable link to your site is provided as a reference point for the generated text.
3) A contribution
This is where your content was synthesised to create the answer, but you weren’t mentioned or cited.
There’s no real way to know is #3 is happening, but my advice to all small business owners is to assume that it is, or that it could, and prioritise content improvements (check out section 5!)
How Google's AI Overviews generate their answers
Google’s AI Overviews work similarly, combining Google’s existing index with generative AI to produce summaries at the top of search results.
What this means for your business is that AI tools build a picture of you from many places simultaneously. Your website is one input. But so are your Google reviews, your LinkedIn page, any press coverage you’ve received, mentions on industry sites, your entries in local business directories, and conversations about you on forums and social media. The AI is assembling a composite portrait of your business, and the richness of that portrait determines whether you get recommended.
Think of it like a reputation in a small town. If ten different people all independently mention the same local builder as reliable and fairly priced, that builder has a strong reputation, even if they’ve never run a single ad. AI works on a similar principle, except instead of ten neighbours, it’s synthesising signals from across the entire internet.
The practical implication is significant: a business with a beautifully optimised website but no presence anywhere else is, from an AI’s perspective, much less credible than a business with an OK website and strong mentions across a variety of third-party sources.
3. How can I get my website AI ready?
Your website is still the starting point, and getting it in order is the prerequisite for everything else in this guide. AI tools will crawl your website to find out about you and to verify third party information. Several things about how your site is built affect how well AI bots can understand and represent what you do.
Say clearly what you do, for whom, and where
This sounds obvious but a lot of SME websites fall down here. Homepage copy that’s heavy on aspirational language (“we’re passionate about delivering transformative solutions for your business”) and light on specific information (“we provide HR consultancy for Scottish SMEs with between 5 and 50 staff”) is hard for AI to interpret accurately. AI tools reward clarity. If your website requires a human reader to read between the lines, the AI won’t bother: it’ll just move on to a competitor whose site is more explicit.
If you’re a plumber, don’t just talk about your “quality plumbing service”. Instead, talk about your 24/7 emergency response and your £60 call-out charge. These are the kind of specific details that an AI is looking for when it’s building an answer to a prompt about choosing a plumber.
Go through your main pages and ask yourself: if someone had never heard of my business, would they know within ten seconds exactly what I do, who I do it for, and where I’m based? If the answer’s no, that’s your first job.
Build content around questions
This is probably the single most impactful change most small businesses can make to improve their AI visibility. LLMs are trained on natural language, and they’re used through natural language queries. Content structured around the questions your customers actually ask – “how do I choose a financial adviser?”, “what’s included in a standard web design package?”, “do I need an employment contract for part-time staff?” – feeds directly into AI responses.
FAQs are undervalued by most small businesses. A well-written FAQ page that genuinely addresses the questions your customers ask you every day is exactly the kind of content AI tools draw from when constructing answers. Similarly, “how to” guides, explainer articles, and “what to expect” content all tend to perform well in AI search because they match the intent behind conversational queries.
Keyword-stuffed content – writing that’s been engineered to include a specific phrase seventeen times – doesn’t translate well into the AI era. Your content needs to be helpful and specific.
Demonstrate expertise with real evidence
The E-E-A-T framework (Experience, Expertise, Authoritativeness, Trustworthiness) was originally developed by Google to evaluate content quality, but it maps directly onto what AI tools are also looking for. Author bios that explain your qualifications and background, case studies with real outcomes, specific examples from client work, your own opinions and analysis based on experience: all of these signal real expertise in a way that generic content doesn’t.
Sort out the technical fundamentals
You don’t need to become a developer for this, but a few technical basics significantly affect how AI tools read your site:
- Structured data (Schema markup) helps AI tools understand what your content is about — identifying that a page is about a specific service, a specific location, a specific person. Most modern website platforms (WordPress, Wix, Squarespace) have plugins or built-in tools that handle this without any manual coding.
- Using the proper H1, H2, and H3 heading styles to organise your content makes it much easier for AI to understand what each section is about. So, if you’ve got any pages where your ‘headings’ comprise normal text in bold, replace these with the proper styles!
- Page speed and accessibility matter too. AI tools that crawl sites can index fast, accessible pages more reliably than slow, clunky ones. Run a Google PageSpeed Insights report to identify any issues that are affecting your website’s performance.
Keep content current
The whole point of AI bots running RAG, or live web search, is to find up-to-date information to supplement what’s already contained in the training data. For that reason, AIs are unlikely to be interested in your blog post from 2019 that hasn’t been updated since.
4. Building your brand footprint beyond your website
AI tools build a composite picture of your business from across the web. If your website is your only significant presence, that picture is very thin – and a thin picture doesn’t inspire confidence.
AI needs to be able to corroborate what your website says about you by finding it confirmed elsewhere. Reviews. Press mentions. Directory listings. Social media activity. Forum contributions. They all add a brushstroke to the portrait.
Online reviews are now a search ranking signal
This is the one that businesses most consistently underinvest in, and it’s increasingly important. Google Reviews feed directly into Google’s AI Overviews. A business with 4.8 stars and 200 reviews is much more likely to be surfaced in AI-generated recommendations than one with 3.9 stars and four reviews, all other things being equal. TrustPilot, industry-specific review platforms like Checkatrade or Clutch and even Facebook reviews all contribute to the picture AI tools build.
If you’re not actively asking happy customers to leave reviews, start doing it. It’s easier to do than you might think: a follow-up email after a completed job, a gentle prompt in your email footer, a QR code on your receipts or invoices. The businesses getting recommended by AI tend to be the ones with consistent, recent, genuinely positive reviews across multiple platforms.
Online PR and press coverage
Getting mentioned in online publications, local news sites, trade press, and industry blogs is one of the most powerful things you can do for your AI visibility. News sites carry high authority in AI training data, and a mention in even a local newspaper’s website carries real weight.
This sounds daunting for a small business without a PR budget, but you don’t need an expensive PR agency to get coverage. In fact, I’ve even seen some anecdotal evidence to suggest that AI actually sometimes prefers to cite niche industry titles over more high profile but less specialised news sources. So it’s not necessary to get featured in the BBC or the Guardian to receive a mention or citation from an LLM. If you sell kitchen appliances, a feature piece in Toasters Daily* might be even better than coverage by the BBC!
(*Not a real publication as far as I know)
And don’t forget your local paper – they’re often on the lookout for expert comment on business topics. If you’re an accountant, they might welcome a quote about the implications of a budget change; if you’re in hospitality, your perspective on retaining staff may be welcomed. If you’re launching something new, or doing anything in your local area, press releases about these activities often get picked up by local press. Trade associations often publish member content. Chambers of Commerce produce newsletters and online features.
Have a look at who’s publishing what in your industry and give some thought to how you can get featured there.
Social media signals
Public social media content is indexed by AI tools, and consistent, engaged social media activity contributes to your overall brand footprint. A LinkedIn company page last updated in 2022 raises questions; one with regular posts, client wins, and consistent engagement does the opposite.
LinkedIn is particularly important for B2B businesses, both for the professional credibility it conveys and because LinkedIn content tends to rank well in both traditional and AI search. And the good news is that you don’t need 50,000 followers or 1000 likes on every post to be cited: research by Semrush suggests that authors posting once a week with a modest following (around 2,000) and consistent engagement are the most likely to be cited by LLMs.
Directory listings and citations
Consistency of your business information across online directories matters more than most businesses realise. Your business name, address, and phone number appearing consistently across Google Business Profile, Yell, Thomson Local, your local council’s business directory, the Federation of Small Businesses, your industry’s trade association directory, and anywhere else you’re listed sends strong signals about your legitimacy and location.
Inconsistencies, on the other hand, create confusion for AI tools trying to build an accurate picture of your business. Do an audit of where you’re listed and make sure everything matches.
Communities and forums
Quora, Reddit, and niche industry forums are extensively represented in AI training data, and genuine, helpful contributions to conversations on these platforms can improve your brand’s perceived authority. The key word is genuine: AI tools are increasingly good at identifying spammy or promotional content, and that can do more harm than good. But if you’re in a field where customers are asking questions in online communities, being the person with the useful answer is a form of visibility that carries real weight.
The diagnostic test
Before you build a strategy, run some searches. Ask ChatGPT, Copilot, and Perplexity the kinds of questions your customers might ask. Note which businesses get recommended and, where the tools cite sources, look at what those sources are. That gives you a map of where you need to show up. If all the recommendations come from businesses with strong TrustPilot profiles, that’s your priority. If the sources are all trade press, that’s your PR angle. Let the data tell you where to focus.
5. Content strategy for AI visibility
Besides adding more specific details, what changes should businesses make to their content strategies to gain exposure in AI search?
This is probably the area with the most crossover with SEO, because when AI bots run live web searches, what are the pages they’re most likely to find? That’s right, the ones with good SEO!
These are your main considerations for a citation-friendly content strategy:
Publish original insight
This is the content principle that matters most for GEO, and it’s also the hardest to shortcut. Google has historically rewarded content that contains original perspectives and first-hand expertise, and it’s easy to see why. After all, which piece would you rather read: an interesting new study or a rehash of information that you’ve seen before?
It makes sense that a blog post that contains your own opinion or a client story (with permission) has a far higher chance of being cited by an AI tool than one that summarises what everyone else is already saying.
So, think about what you know from doing your job every day that your customers would find useful. That first-hand knowledge is exactly what both traditional and AI search likes to reward.
Write for humans
If you’ve spent the last few years writing content with keywords in mind rather than readers, it’s worth revisiting your main service pages and key blog posts with fresh eyes. Ask yourself: does this actually answer the question my customers are asking? Does it say something useful? Is it easy to read? Use Hemingwayapp.com to score and improve your content’s readability. After all, you want your whole audience to be able to understand what you’ve written, don’t you?
The bottom line is: If your content reads like it’s written for a machine, re-write it! Ironically, the machines will like it more if it sounds human (again, that’s true of both traditional Google search and AI search).
Publishing consistently
Frequency matters less than consistency. A business that publishes one thorough, well-researched article per month is in a stronger position than one that posts frantically for a fortnight and then disappears for three months. Consistent publishing signals that the business is active and the content is being maintained.
Most small businesses struggle to commit to a regular publishing schedule. My advice is to identify five or six topics that your customers ask about regularly, then write one substantial, well-researched piece on each, keep them updated, and let that form the core of your content. Add to it over time as new questions emerge.
Become a source, not a reference
The best content strategy for GEO is to become the business that other people cite, rather than the business that cites others. That means producing original research (even simple surveys of your own clients can count), sharing proprietary data or observations from your work, writing genuinely useful guides on your area of expertise. It takes more effort than publishing roundups of other people’s ideas, but the visibility payoff is substantially higher.
6. Local SEO and AI search for Scottish SMEs
For many small businesses, local visibility is the whole game – being recommended to customers in your town or region matters far more than national reach, especially if you have a physical presence in those locations.
The good news is that local AI search is an area where small businesses can compete with larger organisations, because local specificity is something the big players often lack.
Your Google Business Profile is more important than ever
Google’s AI Overviews and Gemini (Google’s AI assistant) pull heavily from Google Business Profile (GBP) data when responding to local queries. A complete, accurate, actively maintained GBP listing is crucial. That means: correct name, address, phone number and website; accurate business category and description; opening hours kept up to date; regular posts (Google lets you publish updates, offers, and events directly from your GBP); and photos that actually show your business, products, or team.
If you’ve claimed your GBP listing and left it to gather dust, put some time into it. It’s one of the highest-return tasks available to a small business for both traditional and AI search.
Consistent local citations
Your business information should be consistent across every online directory where you appear. Start with the big ones – Yell, Bing Places, Apple Maps, TripAdvisor if relevant – then work through industry-specific directories and any local resources.
Each consistent citation strengthens the local signal that AI tools use when constructing location-specific recommendations.
Local content signals
Look for opportunities to create content that’s explicitly rooted in your location. These could include blog posts about local topics, case studies that name the areas you serve, and involvement in local events and community activities. A website that mentions Glasgow in relevant, meaningful contexts is more likely to appear in response to a prompt that mentions Glasgow than one with no location-specific content.
7. Measuring and monitoring AI visibility
One of the most frustrating aspects of GEO right now is that the measurement tools are still catching up with the technology. You can’t log into a dashboard and check your “AI ranking” the way you can check your Google positions in Google Search Console. But there are practical things you can do to track progress and identify where to focus effort.
Manual testing
The most accessible monitoring approach is simply to run regular test queries through ChatGPT, Copilot, Perplexity, and Google’s AI Overviews and record what comes back. Do this monthly. Note which businesses get recommended, what sources are cited, and whether your business appears.
Really try to emulate your prospective customers when you do this. Write prompts the way you think they might write them when searching for answers to their problems.
If you notice certain competitors are getting cited and mentioned and you’re not, it’s time for some reverse-engineering: where are they showing up that you’re not? Who’s talking about them, but not you? Let these insights guide your strategy.
Track AI referral traffic
Google Analytics and other analytics platforms are beginning to surface traffic from AI sources. ChatGPT.com, Perplexity.ai, and Copilot are starting to appear as referral sources for businesses that are being cited in AI responses. Set up a specific segment or report in your analytics to track this traffic, and watch for growth over time. Even if the numbers are small today, the trend is what matters.
Monitor brand mentions
Set up Google Alerts to be notified when your business is mentioned online. This is useful for understanding your brand footprint: where you’re appearing, in what context, and whether the mentions are positive. It also alerts you to any inaccurate information being published about your business, which is worth knowing and addressing quickly.
Also keep an eye on your brand search volume within Google Search Console. Often, if you’re mentioned but not cited in LLMs, people will Google who you are. An increase in traffic from brand keywords often correlates with increased exposure in LLMs.
Traditional SEO metrics as a proxy
Domain authority, the quality and quantity of your backlinks, organic search traffic, and your overall Google search rankings are all meaningful proxy indicators for AI visibility. The signals that tell Google your site is authoritative are, broadly speaking, the same signals that tell AI tools your business is worth recommending. Improving your traditional SEO metrics will tend to improve your GEO/ AEO as a downstream effect.
Set a review cadence
A realistic review schedule for most small businesses is quarterly. Set aside an hour every three months to run test queries, check your analytics for AI referral traffic, review your brand mention alerts, and assess whether your content publishing has stayed on track.
I've finished reading this guide: what next?
This conclusion won’t surprise you, but you need an AI search audit. After all, you can’t improve your discoverability until you have a reasonable idea where it’s at in the first place.
Start with the simple diagnostic test from Section 4: open ChatGPT (or your LLM of choice) and run a prompt as a customer looking for your type of business in your area. What you find will tell you more about where to prioritise your effort than any generic checklist.
If your website content is the weak point, that’s where to start: clear, question-led, specific content that demonstrates real expertise. If your third-party presence is thin, focus on reviews and a simple PR strategy. If your local citations are inconsistent, an audit of your directory listings should pay dividends quickly. Base your initial strategy on plugging the most obvious gaps in your online visibility, and build from there.
The businesses that will be easiest to find in AI search in two years are the ones building the foundations now. You don’t need to be a tech nerd or spend thousands on expensive consultants: you just need to establish a consistent and helpful presence in the most relevant parts of the internet for your business.
Thank you for reading, and good luck!
Written by: Gary Ennis, Business Gateway trainer, and Managing Director of Digital Skills training company NSDesign Ltd.