ChatGPT and Gemini choose local businesses by pulling together signals from across the web — your listings, reviews, website content, and third-party mentions — and then deciding which businesses best match what the person is asking for. They are not simply reading a single directory or ranking page; they are synthesising information from many sources at once. Understanding which signals carry weight is the first step to making sure your business gets named.

What "recommendation" really means (and why it varies by assistant)

When someone asks ChatGPT or Gemini "Where's a good plumber near Galway?" or "Best hair salon in Dublin 4?", the assistant generates a response rather than displaying a ranked list of links. This is fundamentally different from a Google search result. The assistant is not pointing to a page — it is composing an answer, and it decides which businesses to name based on what it has learned from its training data and, in some cases, live web browsing.

GEO (Generative Engine Optimisation) is the practice of shaping your online presence so that AI-generated answers include your business — as distinct from traditional SEO, which focuses on ranking in blue-link search results. If you want to go deeper on how these two approaches differ for Irish businesses, GEO vs AEO for local businesses in Ireland: what to focus on covers that in detail.

Because ChatGPT, Gemini, and Perplexity each use different underlying models and data pipelines, they do not always reach the same conclusion. One assistant might name you; another might not. That inconsistency is not random — it reflects real differences in what each platform has indexed or been trained on.

Core signals: location, relevance, reputation, and recency

AI assistants weight several broad categories of signal when deciding which local business to surface.

Location accuracy is foundational. If your address, Eircode, and service area are stated clearly and consistently across your website, Google Business Profile (GBP — the free listing that powers your presence on Google Maps and Search), and third-party directories, the assistant can confidently associate you with a specific place. Inconsistencies — a different phone number on your website versus your GBP, or an outdated address on a directory — create ambiguity that works against you.

Relevance means how clearly your content describes what you actually do. An assistant trying to answer "emergency electrician Cork" needs to find explicit, unambiguous text that says you are an electrician in Cork who handles emergencies. Vague taglines and generic homepage copy make this harder.

Reputation covers the volume, recency, and sentiment of your reviews across platforms. This includes Google reviews, but also Tripadvisor for hospitality businesses, Treatwell or Booksy for beauty and wellness, and Golden Pages for trades. The assistant draws on all of these, not just Google.

Recency matters because AI assistants — particularly those with live browsing capabilities — give some weight to how recently your business has been mentioned or updated. A business that has not published anything new, responded to a review, or updated its listing in a year looks less active than one that has.

How the assistant uses your web and listing data

It helps to think of an AI assistant as a very thorough reader who has processed an enormous amount of text. When it encounters a query about a local business, it is drawing on:

  • Your website content — particularly your About page, service pages, and any location-specific pages you have created
  • Your Google Business Profile — including your business category, description, services listed, Q&A section, and photos
  • Third-party directories and review platforms — Tripadvisor, Yelp, Trustpilot, and sector-specific sites like TheFork for restaurants or Booksy for salons
  • Editorial mentions — local news coverage, blog posts, or guides that name your business
  • Social media profiles — particularly public pages that describe what you do and where you are based

The assistant does not visit each of these sources in real time for every query (unless it has browsing enabled). Much of what it knows comes from training data compiled at a particular point in time. This is why keeping your information accurate and up to date across all of these touchpoints matters — you are not just optimising for today's query, but for the next time the model's data is refreshed.

A useful mental model: if a journalist who had never heard of you tried to write two accurate sentences about your business using only what they could find online, could they do it? If the answer is no — because your website is thin, your GBP is incomplete, or your listings are inconsistent — the assistant faces the same problem.

This is one of the most common sources of confusion for business owners. You might test ChatGPT and find it names you, then try Gemini and find it does not. Or you might be recommended in one city but not another, even though you serve both.

Several factors explain this:

Factor What it means in practice
Training data cut-off Each model was trained on data up to a certain date; newer listings may not be included
Browsing capability Some assistants can search the live web; others rely entirely on training data
Data source weighting Gemini integrates tightly with Google's own index; ChatGPT draws on a broader but different corpus
Query phrasing The same business may appear for "best café Limerick" but not "coffee shop near the Hunt Museum"

This is also why myths about AI recommendations can be misleading — being well-known locally or having strong Google rankings does not automatically translate into being named by an AI assistant.

What you can actually do about it

Knowing the signals is only useful if it leads to action. Here is a practical starting point:

  1. Audit your Google Business Profile. Make sure your category is precise, your description uses natural language that describes your services, and your opening hours are current. Fill in every optional field — services, products, attributes, Q&A.

  2. Check your NAP consistency. NAP stands for Name, Address, Phone number. These three details should be identical across your website, GBP, Tripadvisor, Golden Pages, and any other directory where you appear. Even small differences (e.g. "St." versus "Street") can create confusion.

  3. Build out your service pages. Each core service you offer should have its own page that clearly states what it is, where you offer it, and who it is for. Thin, single-page websites give AI assistants very little to work with.

  4. Encourage and respond to reviews. Ask satisfied customers to leave a review on Google or whichever platform is most relevant to your sector. Respond to reviews — both positive and negative — as this signals an active, engaged business.

  5. Get mentioned in local editorial content. A mention in a local news article, a round-up blog post, or a community guide carries weight because it is a third party vouching for your existence and quality.

  6. Update your listings regularly. Even small updates — a new photo, a seasonal offer, a response to a question — signal to crawlers and models that your business is active.

How suggestedby.ai helps you diagnose and act

Before you invest time in any of the above, it is worth knowing where you currently stand. suggestedby.ai's free check runs your business against real queries across the main AI assistants and tells you whether you are being named — and if not, gives you a clearer picture of why. It is a practical starting point rather than a theoretical exercise, and it means you can focus your effort on the gaps that are actually costing you visibility.

The signals that matter to AI assistants are, for the most part, things you already have some control over. Your listings, your website, your reviews, your mentions — these are all levers you can pull. The key is understanding which ones are currently weak and addressing them in order of impact.