AI Search for Service Businesses: The New Front Door

Reading time: 8 minutes

Last verified: 11 August 2026

The buyers who used to find you on Google are now asking ChatGPT, and if it does not know your name, you do not exist in the answer.

The May 2026 Google core update launched on 21 May 2026 and finished rolling out on 2 June 2026, according to Google's Search Status Dashboard. It moved the ground under every service business that earns its leads from organic search. Search for almost anything now and the top of the page is an AI-generated answer. Perplexity cites its sources. Google summarises before you click. OpenAI confirmed in August 2026 that ChatGPT had passed 1 billion weekly active users, and many of those people are asking it the questions they used to type into a search box.

→ The path your buyers take to find you has been rewritten, and most owners are still optimising for the path that no longer leads to them.

This article explains what AI search actually is, what the evidence does and does not say about it, and the first three things to fix so AI tools start naming you instead of the firm down the road. By the end you will know whether your business is in the game or watching from the car park.

What is AI search, and why should a service business owner care?

Quick Answer: AI search is the layer of discovery sitting on top of Google, where tools like ChatGPT, Perplexity and Google's AI Overviews answer a buyer's question directly and cite a handful of businesses. If you are not one of the cited businesses, you are not in that conversation, however good your website looks.

A plumber in Leeds used to compete for a spot in the top three Google results. Now there is a layer above that. A homeowner asks an AI tool for the best emergency plumber near them and gets back two or three named businesses with a short reason for each. Every other plumber in town is invisible at the moment the buyer decides.

The same shift is running through interior design, beauty, retail with a service arm, accountancy, and most categories where a buyer is choosing someone to trust with money.

The prevalence numbers are worth quoting precisely. BrightEdge reported in February 2026 that AI Overviews trigger on approximately 48% of tracked queries, up from around 30% a year earlier. That is tracked queries, not all searches, and the difference matters. Even so, roughly half the queries BrightEdge follows put a generated answer in front of the buyer before a single blue link.

Then there is what happens after the answer appears. SparkToro and Similarweb published research in June 2026 finding that 68.01% of US Google searches between January and April 2026 ended without a click on anything. Zero-click behaviour is not new, and AI answers are not its only cause, but the direction is consistent: more questions are resolved on the results page itself.

Why are AI tools citing my competitors and not me?

Quick Answer: Because the things that make a page easy to quote are missing, thin or unreadable on your site. It is rarely about how attractive the site is. It is about whether a stranger, human or machine, can lift a clean fact off the page and see who stands behind it.

Nobody outside the labs knows what a large language model checks before it names a business. Anyone who says they have that list is guessing. What you can do is study the pages AI answers actually cite and build a scoring framework from the editorial patterns they share.

That is what our 19 trust signals are: the Ambitions scoring framework, built from editorial patterns observed in sources that AI answers do cite. It is not a leaked ranking algorithm and we do not present it as one.

A few of the signals, so you can see the shape of it:

Named authors. Every service page names the human responsible, with a real biography and real qualifications.
Service and LocalBusiness schema. Structured data telling a machine what you do, where you do it, and for whom.
A live review feed. Not just the lifetime total, but whether anything has landed recently.
Direct answers on the page. Questions your buyer would type, answered in the first two sentences under the heading.
Third-party mentions. How often other credible sites refer to you by name, in context.

Ahrefs analysed 1.9 million citations across 1 million Google AI Overviews in July 2025 and found that 76.1% of cited pages already rank in Google's top 10. That covers Google AI Overviews specifically, not ChatGPT or Perplexity. This is not a separate internet. It is the same content, judged with a different question in mind.

Most service-business websites were built before any of this mattered. The fix is usually editorial rather than technical, and that is the part that catches owners off guard.

Is this just SEO wearing a new hat?

Quick Answer: No. SEO earns you a position in the blue-link results. Answer engine optimisation earns you a mention inside the generated answer. They overlap heavily, as the Ahrefs data shows, but the work that makes you quotable is not the work that makes you rank.

Strong SEO still matters. What has changed is what a ranking is worth on queries where an answer sits above it.

Seer Interactive looked at 3,119 informational queries and 25.1 million organic impressions between June 2024 and September 2025. On queries that showed an AI Overview, organic click-through rate fell from 1.76% to 0.61% over that period. Read the caveat, because most articles leave it out: Seer does not claim AI Overviews alone caused the fall, because click-through on queries without an AI Overview dropped 41% over the same window. Something broader is happening to informational clicks.

That nuance sharpens the case for doing the work. If informational clicks are getting harder to win across the board, being named inside the answer stops being a bonus and starts being the point.

So you add a layer on top of SEO and write in a way that makes your pages easy to quote. That is exactly what this article is doing. You are reading the mechanic work.

What is the one thing I can do this week that actually moves the needle?

Quick Answer: Open your three best-performing service pages and ask whether a real, named human expert appears on each one. If not, add a named author with a short biography, their qualifications and a photograph by Friday. It is the cheapest fix on the list.

Across the service-business websites we audit, author identity is the thing we most often find missing. I will give you the reasoning as reasoning rather than dress it up as research: a page that names nobody gives an answer nobody to credit. If a claim on your site cannot be attached to a person, it is a claim without an owner.

Google's E-E-A-T framework is worth reading, with one caveat: it is not a ranking factor, and I will not claim AI engines apply it. It is the framework Google's human quality raters use to grade search quality. It still tells you what Google believes good looks like.

The order to fix in is simple.

Author. Name a real person on every service page, with a bio and credentials.
Schema. Add Service and LocalBusiness structured data for each location, then check it with Google's free Rich Results Test.
Reviews. Switch on a flow that asks for a review the day after the job ends.

Three jobs, most of them a morning each. Broken schema is more common than owners expect, so test rather than assume.

How long until any of this pays off?

Quick Answer: Nobody can honestly give you a date, and I will not invent one. No engine publishes when it will next crawl your site or how it decides what to cite, so any agency quoting you a specific window is selling you a guess.

Timing claims are where this industry loses credibility. What I can tell you is what we see across our own clients: consistency beats speed. Owners who make one fix, wait a week, then panic-buy ads tend to give up shortly before the work starts showing.

A US plumbing company we worked with, twelve service pages, moved from 7 of 19 on the trust signals to 14 of 19. Those scores come from our own scorecard records. That is one business's result, it is not typical, and the owner did the operational, sales and delivery work while we did the marketing layer.

The businesses that win in AI search are not the biggest. They are the clearest. A focused plumber in one town with a clean trust-signal layer can be named ahead of a national franchise that never did the editorial work.

Liyana's insight

I have this conversation every week. The owner has heard about AI search, nodded politely, and quietly moved it to the "next year" pile. That pile is where good businesses go to lose ground. What frustrates me is that the first three fixes are small: a name on a page, a block of structured data, a review request that goes out automatically. None of it needs a big budget. The owners who start this quarter are the ones I will be congratulating in twelve months. The ones who wait are the ones I will be rebuilding in eighteen, while their competitors sit inside the answer.

So where does that leave your business right now?

You have three choices. Wait and see whether AI search settles down. Spend more on ads and hope the leak does not widen. Or do the editorial work this quarter, while your competitors argue about whether any of it is a fad.

Start with the trust-signal layer. "The 19 Trust Signals That Decide Whether AI Search Cites You" walks through every signal and how to score your own site. Once you have a score, "The Five-Pillar Blueprint for AI Search Visibility" shows the order the pieces go in, from the buyer-question habit through the website, the citation layer, the AI sales and service engine, and reviews automation.

Comment SIGNAL on any Ambitions AI post on Instagram or LinkedIn and ManyChat sends the 19 Trust Signals briefing to your DMs.

FAQ

Do I need to abandon Google SEO to do well in AI search?

No. Ahrefs found that 76.1% of pages cited in Google AI Overviews already rank in Google's top 10. Answer engine work sits on top of SEO rather than replacing it.

Are the client results you publish verified?

They are reported by the client and are not independently audited. We publish named results only with that client's agreement, and we say clearly when a figure comes from the owner rather than a third party. Any single result is one business's outcome and should not be read as typical.

Is the list of 19 signals something Google or OpenAI published?

No. It is our own scoring framework, built from the editorial patterns we observe in the sources AI answers cite. Nobody outside those companies knows what a model checks internally, and you should be sceptical of anyone who claims otherwise.

Can I do this myself, or do I need an agency?

You can do the first three fixes yourself over a long weekend: named authors, schema, and an automated review request. The ongoing work, citation building and keeping a review feed alive, is what gets hard to sustain while running the business.

Why will you not tell me how long it takes to get cited?

Because I would be making it up. The engines do not publish their crawl or citation schedules, and the honest answer is that we measure the trust-signal score, which we control, rather than promising a date we do not.