Ranking in Google AI Overviews: The 2026 Playbook for UK Websites

A plumber in Worthing searches Google for 'emergency boiler repair near me' and the first thing they see is not a list of blue links. It is a paragraph generated by Google, summarising what emergency boiler repair costs, how quickly an engineer should arrive, and what to check before calling. Three businesses are cited inside that paragraph. Everyone else is invisible. That is the everyday reality of Google search in 2026, and it is happening right now for trades, service businesses, and e-commerce stores across the UK.

Understanding why this shift happened matters less than understanding what to do about it. The sections below cover both, starting from the ground up.

Why AI Overviews Have Rewritten the Rules of Google Visibility

Google AI Overviews now appear for a significant and growing share of UK search queries. For informational searches, how-to questions, and local service queries, the AI-generated summary box sits at the very top of the results page, above every organic listing including whatever sits at position one. If your site is not cited inside that box, a user can read a full answer to their question without ever seeing your name.

The goal has shifted fundamentally. Traditional SEO was about ranking a URL high enough that someone would click it. AI Overview optimisation is about earning a citation inside a generated summary, which is a different target entirely. You are no longer just competing for position on a list. You are competing to be the source that Google's AI trusts enough to quote.

The traffic difference is real, if not dramatic. Semrush research across 7,800+ queries found that sites appearing as sources inside AI Overviews see click-through rates rise from roughly 0.6% to 1.08%. That uplift sounds modest, but across thousands of monthly impressions it compounds into a meaningful stream of visitors, and the brand association of being cited by Google carries weight that pure click numbers do not capture.

What has not changed is the basic requirement for good content. What has changed is what 'good' means in practice. Keyword density is not the main currency for AI citations. Stuffing a page with the phrase 'boiler repair Worthing' fifteen times does nothing to earn a citation. Context, trust, and content structure are what AI models weigh most heavily when deciding which pages to pull from.

UK small businesses in trades, services, and e-commerce are particularly exposed to this shift. Local queries and transactional queries, precisely the type that most small businesses rely on, are among the query types most likely to trigger an AI Overview. A search for 'best local electrician in Brighton', 'how much does a kitchen refit cost in the UK', or 'affordable custom furniture online UK' will very often produce an AI-generated answer. If your content is not structured to be cited, your competitors who have adapted will take that visibility.

How Google Actually Decides Which Sites to Cite

The first thing to understand is that AI Overviews are not pulling from a separate, mysterious database. Google's AI draws on the same crawl index that powers standard organic results. That means a page must first be indexable, crawlable, and capable of ranking before it can ever be considered for a citation. If your site has indexing errors, noindex tags applied by mistake, or thin pages that Google has chosen to ignore, those pages will never appear inside an AI summary regardless of how well-written they are. Technical foundations come before everything else.

Once a page is indexable, Google applies E-E-A-T signals as a filter. E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness, and Google's AI weights these heavily when assessing whether a source is worth citing. For a trades business, this means practical things: a named person or team on the about page, specific credentials or trade body memberships mentioned, real case studies with before-and-after outcomes, and customer testimonials that reference specific jobs rather than generic praise. An anonymous page saying 'we are experts in plumbing' signals nothing. A page saying 'our engineers are Gas Safe registered with 12 years of experience in Sussex, as evidenced by our 94 five-star reviews on Google' signals quite a lot.

Topical authority is the next filter. Google is assessing whether a website genuinely covers a subject in depth, or whether it has one generic services page that vaguely mentions a broad topic. A plumbing company with 15 interlinked pages covering boilers, radiator bleeding, leak detection, emergency call-outs, and bathroom installation is far more likely to be cited on any one of those topics than a site with a single page that mentions all of them in passing. The AI is looking for a domain that has demonstrated it understands a subject from multiple angles.

This is not about volume for its own sake. Ten thin pages are worse than five genuinely useful ones. The goal is a content structure where each page answers a specific question completely, and where the pages link to each other in a way that shows Google how the topics relate. A boiler installation page should link to the boiler maintenance page, which should link to the emergency boiler repair page. Each page is a standalone answer; together they build a signal of expertise.

Semantic relevance matters more than keyword repetition. The AI is not counting how many times a phrase appears. It is assessing whether the content fully addresses the intent behind a query. A user asking 'how long does a boiler installation take' wants to know the time involved, what affects it, how to prepare, and what happens if something goes wrong. A page that answers all four of those sub-questions is semantically relevant. A page that says 'boiler installation time varies' and then lists services is not.

Finally, the relationship between organic ranking and AI citation is real but not absolute. Pages already ranking in the top ten organic positions are cited most often, and aiming to rank in the top ten remains the most reliable path to citation. However, well-structured pages ranking between positions 11 and 20 are cited when they contain clearly extractable answers that higher-ranking pages lack. This is the gap that careful formatting can exploit. A page in position 14 that opens every section with a direct, quotable answer stands a better chance of citation than a page in position 8 that buries its answers in dense paragraphs.

The Inverted Pyramid: Writing Content AI Can Actually Extract

For years, a common SEO content pattern was to open with background context, build through supporting evidence, and land the key point at the end. That approach works for novels. It fails completely for AI citation. Google's AI scans the first 60 words after each heading to find an extractable answer. If those 60 words are scene-setting or hedging, the AI has nothing to cite and moves on to the next page.

The format that works is called the inverted pyramid, or summary-first writing. The answer comes first. The context and detail come after. Every heading introduces a question, the first paragraph answers it directly, and the rest of the section expands on why and how.

A Worked Example

Take a page on a web design site covering the question 'how long does a website redesign take'. The wrong way to open the section is: 'Website redesigns are complex projects that depend on many different variables, including the size of the site, the design requirements, and the availability of content from the client.' That sentence says nothing the user can act on.

The right way is: 'Most small business website redesigns take four to eight weeks, depending on the number of pages and how quickly content is supplied.' That sentence is directly quotable. The AI can pull it into a summary. The user gets an immediate answer. The rest of the section can then cover what extends the timeline, what speeds it up, and what a realistic project schedule looks like week by week.

This principle applies to every H2 section on every page. Each section should function as a self-contained answer unit: a question in the heading, a direct answer in the opening paragraph, and supporting detail below. A user who only reads the first paragraph of each section should come away with the key facts. That is also exactly what an AI model does when it scans your page.

Sentence and Paragraph Length

Sentences should average under 20 words. Paragraphs should be no more than three or four lines on screen. This is not about writing for people who struggle to read. It is about making content parseable, both by AI models processing the text and by business owners who are reading on a phone between jobs. Dense walls of text hide information. Short, clear units surface it.

Watch for hedging openers. Phrases like 'It is important to consider that...', 'There are many factors that influence...', or 'In order to fully understand this topic, it is first necessary to...' consume the extractable window at the top of a section without giving the AI anything to work with. Every word of that opening paragraph is competing for citation space. Use those words to answer the question, not to warm up to it.

When to Use Lists

AI models are trained to extract structured lists into summaries more reliably than prose paragraphs. Use numbered lists for processes where order matters. Use bullet lists for features, comparisons, or groups of related items where order does not matter.

For example, if you are explaining what affects the cost of a website redesign, do not write: 'The cost is affected by the number of pages, the complexity of design, whether e-commerce functionality is required, and whether copywriting is included.' Write it as a list instead:

• Number of pages and the amount of existing content to migrate
• Level of custom design versus a template-based approach
• Whether the site needs e-commerce or booking functionality
• Whether copywriting is included or supplied by the client

The list version is easier for a reader to scan, and it is far easier for an AI to extract and reformat into a summary. Both benefits matter. According to research from Averi.ai, sites with structured data and formatted content see up to 30% higher visibility in AI Overviews. Proper use of lists is one of the most immediately actionable formatting choices you can make, and it costs nothing to apply to pages you have already written.

One practical exercise: open your three most important service or product pages and read only the first paragraph after each H2 heading. If those paragraphs do not contain a direct, factual answer to the implied question in the heading, rewrite them before doing anything else. That single change will do more for your AI citation potential than any amount of keyword research on pages that have not yet been updated.

Structured Data: The Markup That Gets You Noticed

Schema markup is one of those tasks that feels technical but delivers results that are hard to argue with. Sites with structured data see up to 30% higher visibility in AI Overviews, according to research cited by Averi.ai. That is a significant return on what is, for most WordPress sites, a few hours of setup work.

The principle behind it is straightforward. Schema markup translates your content into a language Google's AI systems can parse without guesswork. Instead of inferring that a block of text is about your business hours, the markup states it explicitly. That removes ambiguity, and AI models strongly prefer unambiguous sources when constructing their Overviews.

Which schema types matter most for UK businesses

For UK service businesses, four schema types deliver the most immediate value:

• LocalBusiness tells Google exactly who you are, where you operate, and when you are open
• Service describes individual services you offer, which helps when AI answers queries like "who provides X near Y"
• FAQPage turns a question-and-answer block into a set of discrete, extractable citation candidates
• HowTo marks up step-by-step guides so Google can pull individual steps into an AI Overview answer

E-commerce sites should add Product and Review markup to the list. Product schema carries price, availability, and SKU data. Review schema surfaces aggregate ratings, which AI Overviews regularly include when summarising product or service options.

Getting LocalBusiness schema right

The fields that matter most in LocalBusiness schema are the ones that establish geographic and operational specifics. At minimum, include the following:

• name: your exact trading name as it appears on your Google Business Profile
• address: full UK postal address with PostalCode populated
• telephone: a local area code number, not just a mobile
• openingHoursSpecification: days and hours per service, especially if you offer emergency call-outs
• areaServed: list specific towns or postcodes rather than a vague region like "the South East"
• priceRange: even a rough indicator such as "££" adds context AI models can use

Setting areaServed to named UK towns (Worthing, Brighton, Horsham) rather than leaving it blank or using a county name alone gives the AI a much cleaner geographic anchor to work with.

FAQPage and HowTo schema in practice

FAQPage schema is particularly powerful because each Question and acceptedAnswer pair functions as an independent citation unit. A page with eight well-written FAQ pairs is effectively eight separate opportunities for Google's AI to extract a relevant answer. Write each acceptedAnswer as a complete, self-contained response rather than one that relies on the surrounding page content to make sense.

HowTo schema is genuinely underused by UK trade and service businesses. A guide titled "How to choose a web design agency" or "How to apply for a UK business grant" with proper HowTo markup gives Google a structured sequence it can pull directly into an AI Overview. Each step should have a name and text property. If your step involves a tool or supply, mark those up too.

Implementation and checking your work

JSON-LD format placed in the page <head> is Google's preferred method. In WordPress, the easiest path is through Rank Math or Yoast SEO Premium, both of which generate clean JSON-LD without you writing raw code. Once implemented, run every page through Google's Rich Results Test and the Schema Markup Validator at validator.schema.org. Schema errors are extremely common, especially mismatched property names and missing required fields, and they fail silently. The markup looks present in your source code but Google simply ignores the broken sections.

Building Topical Authority With Content Clusters

A single well-optimised page can earn an AI Overview citation. But earning citations consistently, across multiple queries, on an ongoing basis, requires something deeper. Google's AI systems are designed to reward sites that demonstrate genuine expertise across an entire topic area, not just isolated knowledge on one page. One strong article is a data point. A cluster of well-researched, interlinked pages is a pattern of authority.

How a content cluster is structured

The model is simple. You build one authoritative pillar page that covers a broad topic comprehensively but not exhaustively. Around it, you publish five to ten supporting pages that each go deep on a specific subtopic. Every supporting page links back to the pillar, and the pillar links out to each supporting page. The internal link network tells Google that these pages belong together and that the site has genuine depth on the subject.

The AI systems that construct Overviews appear to favour this kind of interconnected depth. When multiple pages on a site consistently answer related questions within the same topic area, the site accumulates a form of topical trust that makes it a more reliable citation source overall.

A worked example for a Worthing-based web design agency

Take a web design agency serving small businesses in the South East. The pillar page covers website design for small businesses broadly: what it involves, what it costs, what good looks like, how long it takes. The supporting cluster pages each handle one specific subtopic properly:

• How to set up an e-commerce store on a small business budget
• Website speed optimisation: what affects it and how to fix it
• Landing page design principles that improve conversion rates
• Mobile-first design: why it matters and what it changes
• Website redesign costs in the UK: a realistic breakdown

Together, these six pages cover the questions a small business owner asks during the research phase of hiring a web designer. When Google's AI is constructing an Overview for any of those queries, this cluster is a strong candidate for citation because the site has demonstrably covered the territory.

Anchor text and internal linking

Internal links only deliver their full value when the anchor text is descriptive. Linking the phrase "website redesign costs in the UK" to your pricing breakdown page tells both Google and your readers exactly what they will find. Linking the word "here" tells nobody anything useful. This is a small change that makes a real difference to how Google's systems understand the relationship between your pages.

Each supporting page should also link to at least one or two other cluster pages where relevant, not just back to the pillar. This builds a richer network rather than a simple hub-and-spoke structure.

Finding the gaps your competitors have missed

The most valuable cluster pages are often the ones nobody else has bothered to write properly. Google's People Also Ask boxes give you a live view of what searchers want to know alongside your target topic. AlsoAsked.com maps these out visually, showing clusters of related questions and the branches they generate. Spend twenty minutes on either tool before you plan your cluster and you will almost certainly find subtopics that competitors have skipped entirely.

Resist the temptation to fill those gaps quickly with thin content. Under Google's helpful content guidance, publishing four well-researched, genuinely useful cluster pages carries significantly more weight than publishing twelve pages that each contain a few hundred words of loosely related text. Completeness and usefulness are the metrics that matter here, not volume.

Local Signals That Push UK Businesses Into AI Overview Results

When someone searches for a service in a specific UK town or postcode, Google's AI doesn't just look at your website. It pulls together signals from multiple sources and constructs an answer based on what it finds across all of them. For local businesses, that means your visibility in AI Overviews is only as strong as your weakest signal, whether that is an incomplete Google Business Profile or inconsistent contact details across directories.

Your Google Business Profile is a prerequisite

AI Overviews for local queries draw heavily on Google Business Profile data. A partially completed GBP listing, or one that hasn't been updated in eighteen months, is a handicap you are giving yourself for no reason. Every section of your GBP should be filled in: business category and subcategories, opening hours, service areas, service list with descriptions, and at least a few photos. Mark updates when your hours change seasonally or for bank holidays. An active, complete profile signals to Google that the business is genuinely operating.

NAP consistency across directories

NAP consistency (your business Name, Address, and Phone number) is a trust signal that AI models treat as a proxy for legitimacy. If your address appears as "14 High Street, Worthing" on your website but "14 High St, Worthing, West Sussex" on Yell and something slightly different again on Checkatrade, the AI encounters conflicting data points and its confidence in citing you drops accordingly.

The directories worth prioritising for UK businesses include:

• Yell.com
• Checkatrade (for trades)
• FreeIndex
• Yelp UK
• Thomson Local
• Any industry-specific directories relevant to your sector

When the same NAP and service descriptions appear consistently across dozens of authoritative UK directories, the AI builds confidence that the business is legitimate and active. That confidence translates directly into a higher likelihood of citation.

Location-specific landing pages

A single "areas we cover" page listing twenty town names is not a local signal. It is a list. What actually earns AI citations for local queries is a dedicated landing page for each significant service area, each one containing original content about that location, LocalBusiness schema with the specific town or postcode in the areaServed field, and genuine relevance to the local context rather than a template with the place name swapped in.

For trades and service businesses, embedding a service area map on each location page and listing the specific postcodes covered reinforces geographic relevance at both the indexing level and the AI layer. A plumber covering BN11, BN12, and BN13 should say so explicitly, on the page, in the schema, and in the GBP service area settings.

Reviews that feed the AI's language

Customer reviews on your GBP that mention specific services and locations contribute to the language pool AI Overviews draw from when summarising local businesses. A review that says "great service" helps very little. A review that says "used Samson Web Design for a new e-commerce site for my Worthing-based shop, really happy with how it turned out" contains service, location, and outcome data the AI can actually use.

Asking satisfied customers to mention the specific service they used and where they are based is a completely legitimate approach. Include a brief prompt in your follow-up email: something as simple as "if you have a moment to leave us a Google review, it really helps if you can mention the service we helped you with and your location." Most customers are happy to oblige when the request is that specific.

E-commerce Tactics: Getting Products Cited Rather Than Just Listed

The gap most UK online shops fall into is assuming that having a product page is enough. For the majority of product-category queries, Google's AI Overview pulls from informational content sitting around the shop, not from the shop itself. A buyer searching for the best waterproof walking boots under £80 will trigger an AI Overview that cites a buying guide or comparison article, not a category page stuffed with facet filters. You need both: the transactional page and the informational layer on top of it.

Product schema is where the transactional page earns its seat at the table. When a query is specific enough, such as 'best noise-cancelling headphones under 100 pounds UK', Google will pull individual product data directly. Make sure every product page includes price, availability, aggregateRating, and brand fields inside your Product schema. Missing any of these leaves the page technically ineligible for that kind of citation, regardless of how well the copy reads.

Review schema carries disproportionate weight for UK e-commerce. AI systems use star ratings and review counts as a shorthand for trustworthiness when they are summarising product recommendations to a user who cannot see the page for themselves. If your reviews live inside a third-party widget that Google cannot parse, you are losing that trust signal entirely. Pull reviews into schema-marked-up structured data on the page itself.

Category pages have a specific job to do. Open each one with a 60 to 80 word paragraph that directly answers the question 'what should I buy in this category and why'. Not a keyword-heavy paragraph about the long history of, say, garden furniture. Something like: 'For most UK gardens, a powder-coated aluminium set with UV-resistant cushions is the best balance of durability and price. Expect to pay between £300 and £600 for a four-seat set that will last five or more seasons without rusting.' That is extractable. The history paragraph is not.

Shipping, returns, and payment details are consistently underestimated as AI Overview fuel. Transactional queries often carry an implicit sub-question: 'will this actually work for me in the UK?' Answering that with clear headings (Free UK Delivery, 30-Day Returns, Buy Now Pay Later Available) and backing it with Offer and ShippingDeliveryTime schema reduces friction for extraction. A buyer asking 'where can I buy X with next-day delivery UK' is not going to be directed to a page that buries delivery information in a PDF.

Comparison tables deserve special attention. Build them in proper HTML table markup, not as screenshots or image files. Google cannot read an image table, which means the AI cannot cite it. Include a column for UK-specific factors: whether prices include VAT, typical delivery timescales in working days, and whether a brand has UK-based customer support. These details are exactly what a UK shopper is trying to confirm before buying, and they are exactly what AI Overviews are designed to surface.

Technical Health Checks You Cannot Skip

All the content strategy in this article sits on a foundation of technical health. A page that loads slowly, returns a 5xx error, or is blocked in your robots.txt file cannot earn an AI Overview citation, full stop. Google cannot cite what it cannot crawl and render. Technical hygiene is not a nice-to-have layer you add later; it is the floor everything else stands on.

The Core Web Vitals targets for 2026 are specific and measurable. You are aiming for Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, and Cumulative Layout Shift under 0.1. Run every key page through PageSpeed Insights and look at the mobile score, not just desktop. A page that passes on desktop but fails on mobile is still a failing page, because Google's crawler and its AI systems index the mobile version of your site first.

HTTPS is not negotiable. An expired SSL certificate or mixed-content warnings (where a secure page loads insecure assets like old images or scripts) undermine the trust signals that AI systems look for when deciding whether to cite a source. If your browser shows a padlock with a warning triangle on any page of your site, fix it before doing anything else. Mixed content is often caused by older images or tracking scripts using HTTP URLs, and a tool like Why No Padlock will identify every culprit in seconds.

For sites with more than around 200 pages, crawl budget becomes a real concern. Pages with no internal links pointing to them are effectively invisible to Google's crawler, because the crawler follows links to find pages and rarely reaches orphaned ones. Excellent content on an orphaned page earns zero citations. Audit your internal linking structure using Screaming Frog's free tier (up to 500 URLs) and make sure every important page is reachable within two or three clicks from your homepage.

Google Search Console's Index Coverage report is something you should be checking weekly, not monthly. CMS updates and plugin conflicts are the two most common causes of pages accidentally being set to noindex or being blocked by a conflicting canonical tag, and both are surprisingly easy to miss. A page that was indexed last Tuesday may not be indexed today because a plugin update rewrote your meta robots settings. The Coverage report will show you the problem before your rankings tell you about it.

Mobile usability is both a ranking factor and an AI Overview eligibility factor. Use Google's Mobile-Friendly Test on every page you are actively optimising for AI citations. Pay particular attention to tap-target sizing (buttons and links that are too close together for a finger to hit accurately) and viewport configuration errors. These issues take an hour to fix and unlock all the content work you are about to do.

Measuring Whether Your AI Overview Strategy Is Working

The honest answer is that measurement here is imprecise, and anyone telling you otherwise is selling something. Google Search Console does not yet have a dedicated AI Overview filter. What you can do is use the Performance report to track impressions and click-through rates for specific question-format queries where you know or suspect an AI Overview is appearing. Filter by query, sort by impressions, and look for queries starting with 'what', 'how', 'best', 'which', and 'is'.

The signal to watch for is a rising CTR alongside flat or declining impressions on a specific query. That pattern suggests you have earned a citation inside the AI Overview. The total number of people seeing the result has not grown, but a higher proportion of them are clicking through to your site, which points to higher-intent visitors being drawn in by seeing your site named as a source. Research across 7,800+ queries found that being featured as an AI Overview source lifts CTR from around 0.6% to 1.08%, a meaningful jump when applied across hundreds of queries.

Third-party tools close the gap that Search Console leaves open. Semrush's AI Overview tracker and Ahrefs' SERP feature filter both let you see which of your target keywords currently trigger an AI Overview, and whether your domain appears as a cited source. Neither is free, but if you are investing serious time in this strategy, the visibility into what is actually working is worth the cost. Run these checks monthly against your priority keyword list.

Set a monthly review cadence of around 90 minutes. In that session, check which cluster pages have gained or lost AI citations, which schema implementations are throwing validation errors in the Rich Results Test, and which competitor pages are now being cited on queries you are targeting. Competitor citations are the most actionable finding: if a competitor is being cited, their page structure is worth analysing closely.

One more trap to avoid: overall organic traffic is a misleading headline metric when you are running an AI Overview strategy. AI Overviews can flatten total clicks even as your citation presence grows, because some users get their answer from the overview without clicking anywhere. Segment your traffic by query type in both Search Console and GA4 using custom channel groups. Look at the behaviour of users who do arrive via question-format queries, because those visitors tend to have higher engagement and conversion rates than generic organic traffic.

Your First 30 Days: A Prioritised Action Plan

The plans that work are the ones that are specific enough to actually do. Here is a week-by-week breakdown that prioritises impact over comprehensiveness.

Week One: Fix Your Five Most Important Pages

Pick the five pages that currently drive the most organic traffic. Audit each one for inverted-pyramid structure: does the opening paragraph lead with a direct, citable answer, or does it warm up slowly before getting to the point? Rewrite the opening paragraph of each page so that the first 60 words directly answer the primary question the page is targeting. Then add or correct FAQPage schema on each one. These two changes take roughly two to three hours per page and represent the highest-leverage work you can do.

Week Two: Nail Your Local and Directory Presence

Before building any new citations, verify that your existing Name, Address, and Phone number information is consistent across your own website, your Google Business Profile, and the five most authoritative UK directories relevant to your industry. For a plumber in Brighton, that means Checkatrade, Rated People, Yell, TrustATrader, and the relevant trade association directory. A single address formatted differently across these sources (Street versus St, for example) is enough to dilute the trust signal AI systems use to confirm you are a legitimate, active business. Fix discrepancies before adding anything new.

Week Three: Build Into the Gaps

Use the People Also Ask boxes that appear on your target queries, plus the free tool AlsoAsked.com, to identify two or three questions your content cluster does not yet answer. Each gap is a missing page. Commission or write one new cluster page per gap, and build each one to the summary-first format from the start rather than retrofitting it later. A page written correctly from day one takes the same time as one written badly and then rewritten. Aim for 600 to 900 words per page, not 2,000, because depth of topic coverage across a cluster matters more than word count on a single page.

Week Four: Technical Sweep

Run Google Search Console's Coverage report and resolve any noindex errors or canonical conflicts. Run PageSpeed Insights on your five priority pages on mobile, and fix any Core Web Vitals failures before you do anything else. Then validate all structured data using Google's Rich Results Test, paying attention to warnings as well as errors, because warnings often indicate incomplete schema that will not qualify for citation even if it technically validates. By the end of week four, your technical floor is solid and your content foundation is in place.

From Day 30 Onwards

Set a recurring monthly session of around 90 minutes. In it: check Search Console CTR trends on question-format queries, run schema validation on any pages you have updated that month, review which competitor pages have appeared in AI Overviews on your target queries, and publish one new cluster page to maintain topical momentum. That last habit, one new cluster page per month, is what compounds. Twelve pages across a year is a topic cluster with real breadth, and breadth is what earns authority in the eyes of both Google's ranking systems and its AI extraction layer.

The businesses that will own AI Overview citations in 2026 are not the ones with the biggest budgets or the most technically sophisticated teams. They are the ones that do the basics consistently: answer questions directly, mark up their content clearly, keep their technical house in order, and show up every month to check what is working. The knowledge is available to everyone. Consistent execution is the only real differentiator left.