2026-04-17

AI systems need clear, consistent evidence before they recommend a business. This explains the website, review, schema, service-page, and external proof work that gives them something reliable to understand.

By Tyler Stocks · Stocks Local

AI search does not work like a normal list of Google results.

When someone asks an assistant for a recommendation, the answer is compressed. The model is not showing every possible business. It is trying to name a few options it can understand and justify.

That means the job is not to trick the model. The job is to make the business easier to understand: what you sell, where you work, who you serve, what proof exists, and why a customer should trust you.

For local businesses, product brands, and service companies, the businesses that show up tend to have the same foundations in place. Their Google Business Profile is complete, their pages answer real questions, their reviews are recent and specific, their website uses clear structure, and their details match across the web.

Here is the practical work worth doing first. Most of it is not glamorous. That is why so many businesses skip it.

1. They have a complete Google Business Profile

This is the single highest-impact thing you can do.

AI models treat Google Business Profile as a primary source of truth for local queries. Not your website alone. Your Google profile matters. If your hours are missing, your services are vague, or your category is wrong, the model has less reliable evidence before it even considers your content.

The cited businesses share a few traits. Their category is specific and accurate. "Plumbing service" beats "Home improvement." Every service is listed individually, not bundled. Photos are real and recent, not stock. Posts are updated at least monthly. Reviews are responded to publicly.

Action this week: log into your Google Business Profile dashboard. Set a 30 minute timer. Fill in every blank field. Pick the most specific category available.

2. They answer questions directly

AI extracts content paragraph by paragraph. If the first 60 to 100 words of a section answer a specific question clearly, that paragraph becomes citable.

Compare these two openings for a service page.

"Welcome to our website. We are passionate about delivering excellence in plumbing services to customers across the region."

"We repair boilers, clear blocked drains, and install bathrooms. Most emergency callouts are booked within 2 hours. Boiler services start at £75."

The first one will never be cited. The second one could be.

The rule is simple. If someone reads the paragraph in isolation, does it tell them something useful? If yes, AI can use it. If no, AI skips it.

3. They have structured data

Structured data is code that tells AI exactly what your business is. Name, location, phone, hours, services, reviews. It sits in the background of your website and machines read it directly.

The schema types that matter for local businesses:

- LocalBusiness, or a more specific type like Plumber, Dentist, Solicitor - Service, for each service you offer - FAQPage, for any Q&A content - Review and AggregateRating, if you collect reviews on your own site

Schema is invisible to visitors. It does nothing for how your site looks. But it is the difference between being understood by AI and being guessed at.

Most websites do not have it. That is part of why so few sites get cited.

Action this week: ask whoever built your site to check if LocalBusiness schema is in place. If not, get it added. A competent developer should quote one to two hours for a basic implementation.

4. They let AI crawlers access their site

A surprising share of business websites accidentally block the bots that AI engines use to read content. The main ones are ChatGPT-User, OAI-SearchBot, PerplexityBot, and ClaudeBot.

If your robots.txt disallows these, AI cannot read your content. If AI cannot read your content, AI cannot cite you. Simple.

Cloudflare and some content platforms block AI crawlers by default unless you explicitly allow them. Worth checking.

Action this week: type your-domain.com/robots.txt into a browser. Look for any "Disallow" lines that mention these bot names. If you see them, have your developer remove them.

5. They have consistent information across the web

Your business name, address, phone, and website must match exactly across every place they appear. Google Business Profile, Yell, Yelp, Facebook, industry directories, Companies House, old listings you forgot about.

AI models build a profile of your business by cross-referencing these sources. If your address is "12 High Street, Dorchester" on Google Business Profile but "12 High St, Dorchester" on Yell, AI sees conflicting evidence. Confusion reduces trust.

Consistent data across 10 to 15 sources strengthens the signal significantly.

Action this week: start a spreadsheet. Column A: every place your business is listed online. Column B: the exact name, address, phone, and website as listed on that source. Find inconsistencies. Fix them.

6. They get recent, specific reviews

Cited businesses average 4.3 stars and above, with at least 5 to 10 recent reviews. Age matters. A review from last month beats a 5-star review from 2022.

Specific reviews beat generic ones. "Stuart fixed our boiler the same day we called, priced fairly, left the kitchen tidy" carries more weight with AI than "Great service, highly recommend."

The reason is that AI looks for signals of authenticity and specificity. Generic reviews read like they could apply to any business. Specific reviews tie to real outcomes a customer can picture.

Action this week: build review-asking into your workflow. After every completed job, ask. Once a month is not enough. Weekly is better.

7. They publish content that answers real questions

Not marketing fluff. Actual answers to the questions customers type into search.

"How much does emergency boiler repair cost?" is a real query people search. A page that answers that question with specific, current numbers will get cited. A page that says "Our pricing is transparent and competitive" will not.

The content formula for AI is short. Specific question as the heading. Direct answer in the first 60 to 100 words. Supporting detail underneath. Numbers, dates, and named examples throughout.

Most business websites talk about themselves. Cited businesses answer questions.

How long until it works

None of this is instant. But it is faster than traditional local SEO.

Schema and robots.txt fixes typically reflect in AI responses within 1 to 2 weeks after the engines re-crawl. Google Business Profile improvements show up in AI within 2 to 4 weeks. Content changes take 4 to 8 weeks for new or updated pages to be cited. Review and consistency work compounds over 3 to 6 months.

The businesses that start now will establish positions that become harder to displace later. AI models train on historical data. The longer your business has been consistently present, structured, and cited, the stronger your position becomes.

Where to start this week

If you do nothing else, do these three things in the next seven days.

First, open ChatGPT or Perplexity. Ask it to recommend a business in your category and location. See if you appear. Screenshot the result.

Second, audit your Google Business Profile. Fill every empty field. Update your category to the most specific option. Add three new photos taken in the last month.

Third, check your robots.txt. Make sure the AI crawlers listed above are not blocked.

That is enough to start moving. The other items can follow as time allows.

The gap is the opportunity

Most business websites were not written for this kind of discovery. They have vague service pages, missing proof, thin FAQs, inconsistent directory data, or product copy that never answers the practical buying questions.

That is the opportunity. You do not need to pretend AI visibility is guaranteed. You do need to build a stronger evidence trail than competitors who have not looked at their website structure in years.

The work is unglamorous. Most of it is administrative. But it is the work that gets done.

If you want a direct read on your Shopify store or website, request a free written teardown. I will tell you what looks unclear, what is blocking trust, and what should be fixed first.

Questions

Asked and answered.

  • How do I check whether AI recommends my business?

    Open ChatGPT or Perplexity and type a query your ideal customer would use. For example: 'best plumber in Dorchester' or 'accountant near Bournemouth'. If your business appears in the response, note why it was recommended. If not, compare your website, reviews, service pages, and Google Business Profile with the businesses that were named.

  • Is AI search optimisation the same as SEO?

    No. SEO optimises your position in a list of ten blue links. AI search optimisation decides whether you get recommended as the single answer. Some signals overlap, like content quality and domain authority. Some do not, like schema markup and AI crawler access. Most businesses need both.

  • How long until I move from invisible to cited by AI?

    There is no guaranteed timeline because each platform crawls and refreshes differently. Basic fixes like schema, clearer pages, and robots.txt access can be completed quickly, while reviews, content depth, and external proof compound over time.

  • Do I need a new website to appear in AI search?

    Not necessarily. Most of the improvements in this article can be applied to any existing website. Sites built with AI visibility in mind from day one tend to perform noticeably better because the foundational work is already clean.

  • Which AI engines matter most for UK businesses?

    ChatGPT accounts for around 77% of AI-driven referral traffic globally, including the UK. Perplexity and Google AI Overviews are growing quickly. Optimising for ChatGPT captures most of the volume, but the same underlying signals tend to help across every major AI engine.

Want a clear second opinion on the site?

Get a free Shopify teardown