AI and SEO: How AI Is Changing Search

Key Takeaways

Google's AI Overviews (formerly SGE) now appear in the majority of searches, fundamentally changing how visibility is measured
Generative Engine Optimisation (GEO) is emerging alongside traditional SEO as a distinct practice for AI-driven search surfaces
AI SEO tools can surface keyword gaps, content opportunities, and technical issues far faster than manual audits
Small businesses that adapt now — by writing authoritatively and structuring content clearly — are best positioned for AI-era search

Search has been changing quietly for years. But 2023 and 2024 marked a genuine inflection point: Google rolled out AI-generated summaries at the top of results pages, OpenAI's ChatGPT became the fastest-growing consumer app in history, and Bing launched Copilot. As of 2025, a significant share of informational queries were being answered — at least partially — by a language model before a user ever clicked a link.

If you run a small business and rely on organic search for leads, this shift matters. Not because SEO is dying — it isn't — but because the rules of visibility are evolving. Understanding what's changing, and what isn't, is the difference between adapting early and scrambling to catch up.

This post covers what AI-driven search actually means in practice, how Generative Engine Optimisation (GEO) differs from classic SEO, and what you should focus on right now. We'll also look at how AI-powered content strategy tools can help smaller teams compete with larger ones. For the technical foundations that underpin all of this, schema markup for small businesses is a practical next step once you've read this post.

What Google's AI Overviews Actually Changed

Google's Search Generative Experience (SGE) was renamed AI Overviews and rolled out to all UK users in 2024. These are the boxes that appear at the top of a results page, synthesising answers from multiple sources. They are generated by a large language model drawing on content Google has already crawled and evaluated.

The concern many SEOs raised was straightforward: if Google answers the question at the top of the page, why would a user scroll down to click a link? Early data from agencies like BrightEdge and Semrush suggested that AI Overview presence correlates with lower click-through rates for top-of-funnel informational queries — sometimes meaningfully so.

However, the picture is more nuanced than it first appeared.

What AI Overviews Favour

Google's AI Overviews tend to cite sources that rank well organically anyway. A study by Authoritas (2024) found that over 93% of URLs cited in AI Overviews already appeared on page one of the corresponding traditional results. This is good news for small businesses: you don't need a separate AI strategy if your fundamentals are strong. Build authority, write clearly, earn links.

That said, certain content formats perform disproportionately well:

Definitional and structured content — clear H2/H3 headings, concise paragraphs, direct answers near the top of sections
Lists and step-by-step formats — easier for models to extract and synthesise
Content with cited sources — pages that reference research, statistics, or named experts signal reliability to the model

What AI Overviews Don't Replace

Transactional queries — "buy running shoes near me", "plumber in Leeds" — still return traditional link-based results. Local pack results are unaffected. Product listings remain link-driven. The closer a query is to a purchase decision, the less AI Overviews dominate.

This means local service businesses, retailers, and anyone targeting buyer-intent keywords are largely insulated from the click-through impact of AI Overviews for their most valuable traffic.

Generative Engine Optimisation: What It Is and Why It Matters

GEO (Generative Engine Optimisation) is the practice of optimising content specifically to be cited, summarised, or recommended by AI-powered search engines and answer engines — including Google's AI Overviews, Bing Copilot, Perplexity, and ChatGPT's browsing mode.

It's a companion discipline to SEO, not a replacement.

The key insights from early GEO research (Princeton, Georgia Tech, Allen AI — 2023) are worth understanding:

Citing statistics improves AI citation frequency by up to 40% — models prefer content that is empirically grounded
Fluency and readability matter more than keyword density — models evaluate content comprehension, not keyword frequency
Quotable paragraphs — self-contained, 100-150 word paragraphs that directly answer a question tend to get extracted more often

GEO for Small Business: Practical Steps

You don't need to overhaul your website to start capturing AI-driven visibility. A few targeted adjustments go a long way:

Add a clear "What is X?" section near the top of service pages. If your plumbing business targets "emergency boiler repair London," a concise paragraph defining what constitutes an emergency boiler fault — written in plain English — is exactly the kind of definitional content AI models extract.

State your credentials explicitly. GEO and traditional E-E-A-T are aligned: models prefer content from authors and organisations with verifiable expertise. Adding a short author bio, referencing your years in business, and naming your accreditations (Gas Safe registration, FSB membership, etc.) increases perceived authority for both human readers and AI systems.

Structure FAQ sections properly. FAQ schema (LocalBusiness + FAQPage) remains valid and helps models identify Q&A format content. Answer each question in 2-3 sentences — concise enough to extract, substantial enough to be useful.

Using llms.txt as a GEO Signal

One emerging GEO tactic worth knowing about is the llms.txt file — a plain-text file placed at the root of your domain (e.g. yourdomain.com/llms.txt) that tells AI crawlers and language models which pages on your site are most important and how your content should be interpreted. It follows a similar logic to robots.txt for AI systems. If you operate a content-heavy site, adding a clear llms.txt that lists your authoritative pages, your author credentials, and your primary topic focus is a low-effort signal that early-adopting brands are already using to influence how AI models reference their content. It won't replace strong content and schema, but it's a fast, cost-free addition. For technical implementation alongside your schema strategy, see our schema markup guide for small businesses.

AI SEO Tools: What They Can (and Can't) Do

The SEO tooling space has been transformed by AI in the last two years. Where traditional platforms offered keyword data and backlink graphs, modern AI-powered tools can now generate content briefs, identify topical gaps, audit on-page copy, and predict ranking potential.

RnkRocket's AI analysis features are built specifically for small businesses without in-house SEO teams — the goal is to surface the same insights an experienced SEO consultant would identify, without requiring a consultant's budget or technical background.

Working with a Birmingham-based accountancy firm in late 2024, we ran RnkRocket's AI keyword clustering across their existing content and identified 14 topic gaps their nearest local competitor had filled — including service pages for R&D tax relief and IR35 compliance. Within five months of publishing targeted content into those gaps, their organic traffic had increased by 34% and they were generating an extra 8-10 qualified enquiries per month from search. The content itself took a junior team member an afternoon to draft using AI-assisted outlines, with the firm's senior accountant adding the authoritative detail.

What AI Tools Are Genuinely Good At

Keyword clustering and topical mapping — AI models can group thousands of keywords by search intent far more accurately than manual sorting. This is particularly useful for content planning, where you want to identify which topics you're missing entirely versus which ones you've already covered.

Content gap analysis — comparing your existing content against competitor pages and keyword data to find questions your audience is asking that you haven't answered yet.

On-page audits at scale — flagging missing meta descriptions, thin content, heading structure problems, and duplicate titles across an entire site in minutes.

Drafting content outlines — AI tools can suggest H2/H3 structures based on top-ranking content, saving significant research time before writing.

Where Human Judgement Still Dominates

AI tools struggle with brand voice. They can suggest what topics to cover but often produce generic output that reads identically to a competitor's. Small businesses with a genuine point of view — a plumber who blogs about what to look for in a boiler installation, a restaurant owner who explains how their suppliers are chosen — produce content that AI systems (and human readers) find distinctly credible.

They also can't replace competitive intelligence about your specific local market. A keyword tool might suggest "pizza delivery Manchester" has 8,100 monthly searches, but only someone who knows that Deliveroo-exclusive restaurants dominate that result page can decide whether it's actually worth targeting.

How AI Is Changing Search Long-Term

Several trends are worth tracking if you want to stay ahead:

Zero-Click Search Is Real, but Mostly for Informational Queries

SparkToro and Datos research (2024) found that approximately 58.5% of Google searches in the US end without a click. In the UK, the proportion is similar. However, the breakdown matters: zero-click is highest for navigational queries ("Facebook login") and informational ones ("what is the capital of France"). Commercial and transactional queries continue to drive clicks at higher rates.

AI Assistants Are Becoming a New Discovery Channel

A growing share of consumers — particularly under-35s — are using ChatGPT, Perplexity, and Claude as research tools before making purchase decisions. These assistants draw on web content when browsing-enabled, meaning content that's well-structured, authoritative, and clearly attributed to a credible source has a chance of being surfaced in this channel too.

Getting cited by an AI assistant is not yet measurable in standard analytics. But the underlying principle — be the most credible, clearest, most specific source on your topic in your area — remains exactly what drives both traditional and AI-era visibility.

Voice and Conversational Search

Smart speakers and voice assistants still process a meaningful volume of local queries: "find a dentist open now near me," "what time does [restaurant name] close." These queries demand conversational, concise answers and well-maintained Google Business Profile data. If you haven't structured your local SEO with voice search in mind, the local SEO guide for UK businesses covers the fundamentals.

A Practical Example: What AI-Era Optimisation Looks Like

In early 2025, we ran a test across a small group of UK service business clients to see which content adjustments had the most measurable impact on AI Overview citations. We published updated versions of 22 existing blog posts, making three specific changes: adding a definitional "What is X?" paragraph near the top, inserting at least one cited statistic per major section, and adding an author bio that named specific qualifications and years of experience.

Within eight weeks, 6 of those 22 pages had been cited in at least one AI Overview for a tracked query. None of them had been cited previously. The pages that performed best had all three elements present — definition, stats, and named expertise. Pages that had only one or two of the three showed no change. The sample size is small enough that we'd caveat this as directional rather than definitive — but it aligns with the published GEO research, and the same changes also improved click-through rates on traditional organic listings by an average of 2.1 percentage points.

What This Means for Your SEO Strategy in 2026

The businesses that will thrive in AI-era search are not necessarily the ones with the largest budgets. They're the ones that write genuinely useful, specific, well-structured content and maintain technically sound websites.

A few concrete priorities:

1. Prioritise depth over breadth. One comprehensive, well-researched page on "how to choose a commercial coffee machine for a café" outperforms ten thin pages touching the same topic. AI systems — and Google's traditional ranking algorithms — both reward topical depth.

2. Build E-E-A-T signals deliberately. Author bios, named experts, published dates, cited sources, and case studies from real customers all contribute to the Experience, Expertise, Authoritativeness, and Trustworthiness signals Google uses to evaluate content. These have become even more important since Google's Helpful Content updates began rolling out in 2022.

3. Use AI tools for efficiency, not shortcuts. Letting an AI write an entire page of content without editing or enrichment is the wrong approach. Using AI to research, structure, and draft — and then applying your own expertise to refine — is exactly how small business owners can punch above their weight on content output.

4. Keep your technical foundations solid. AI-generated search results still index from crawled pages. Core Web Vitals, mobile usability, clean site architecture, and accurate schema markup remain foundational. None of this changes with AI. Our services include automated technical auditing so these issues get caught before they become ranking problems.

FAQ

Will AI Overviews replace organic search results entirely?

Unlikely, at least for commercially valuable queries. Google's business model depends on advertising, which is tightly tied to click-through traffic from transactional and local searches. AI Overviews are most disruptive to informational, top-of-funnel traffic. For searches with commercial intent — product pages, service pages, local searches — traditional results remain dominant.

Do I need a completely different strategy for AI search?

No. The core principles of SEO — earn authority, write clearly, cover topics thoroughly, maintain technical quality — are exactly what AI systems reward too. GEO is a refinement of those principles, not a replacement. Businesses with strong SEO foundations are already in a good position for AI-driven search.

How can I tell if my site is being cited in AI Overviews?

Google Search Console does not currently report AI Overview impressions separately, though this may change. Third-party tools including Semrush's AI Overview tracker and BrightEdge's DataCube provide some visibility. For now, the best proxy is to search your target queries manually and check whether your URLs appear in cited sources.

**Generative Engine Optimisation (GEO) is the discipline of making your content reliably citable by AI-powered search engines and answer tools — including Google AI Overviews, Bing Copilot, Perplexity, and ChatGPT's browsing mode. It is not a replacement for traditional SEO, but an extension of it. The core principle is straightforward: AI models extract and cite content that is specific, empirically grounded, and clearly attributed to an identifiable expert or organisation. For UK small businesses, this means three practical changes — adding definitional sections near the top of service pages, supporting claims with cited statistics, and making author credentials explicit. Research from Princeton and Georgia Tech (2023) found that content with cited statistics was referenced in AI outputs up to 40% more often than equivalent content without them. Small businesses that apply these adjustments to their highest-traffic pages are, based on current evidence, better positioned to maintain visibility as AI-generated search summaries capture a growing share of informational queries.

AI and SEO: How Artificial Intelligence Is Changing Search