What Google and AI Search Both Check Before Recommending Your Business

When a potential customer asks Google or an AI assistant "best restaurant in Sheffield city centre" or "hair salon open Sunday in Birmingham", a decision is made in fractions of a second about which businesses to recommend. That decision is not random. It reflects a set of signals that both Google and AI assistants have already evaluated.

Some of those signals are on your website. Some are on your Google Business Profile. Some are across every directory, review platform, and data source that has indexed your business name and address.

Understanding what those signals are is the starting point for improving your visibility. Most of them are within a small business owner's control.

Signal 1: Structured data on your website

Structured data is code added to your website that tells search systems exactly what your content represents, rather than leaving them to infer it from text alone.

For a small business, the most valuable structured data types are:

LocalBusiness schema (or a subtype like Restaurant, HairSalon, BeautyBusiness, or GroceryStore). This tells search systems your name, address, phone number, opening hours, geographic area, and business type in a format they can read directly without parsing your page content.

FAQPage schema. This wraps your frequently asked questions in a format that allows search systems and AI assistants to match your answers directly to user questions. A restaurant with FAQPage schema answering "do you take walk-ins" or "is there parking nearby" can have those answers appear in AI assistant responses without the user visiting the website.

Service schema. For businesses offering multiple services, this describes each service, its description, and optionally its price range.

Adding structured data requires a developer if you are not technical, but for a standard small business website it is typically a few hours of work. Plugins exist for most website platforms including WordPress, Squarespace, and Wix that generate structured data automatically. The value is disproportionate to the time required.

Signal 2: Google Business Profile completeness

Your Google Business Profile is the single most direct input into local search results. AI assistants use it for local queries. Google Maps pulls from it. Users who find you via search often make contact through it without ever visiting your website.

Completeness matters. Profiles with every field complete receive more visibility than partial profiles. The fields that carry most weight are:

Primary and secondary business categories. Choose the most specific category that accurately describes your business. A pizza restaurant choosing "Restaurant" when "Pizza Restaurant" is available is missing a targeting opportunity.

Opening hours, including special hours for bank holidays and seasonal closures. Inaccurate hours are one of the fastest ways to generate negative reviews and to be deprioritised in recommendations.

Services or menu. Listing your specific services or menu items gives both Google and AI assistants more detail to match against specific customer queries.

Photos updated regularly. Profile completeness algorithms weight recent photo additions. A profile with ten photos added three years ago carries less signal than one with recent additions.

Signal 3: Review volume, recency, and responses

Reviews are a quality signal for both Google and AI assistants. Volume matters. Recency matters. And the presence of owner responses matters more than most businesses realise.

An AI assistant asked to recommend a business in a specific area will factor in review data as part of its decision. A business with 15 reviews from three years ago competes poorly against one with 80 recent reviews, even if the average rating is similar.

Responding to reviews signals that the business is active and that someone is paying attention. Responding to negative reviews thoughtfully, without defensiveness, signals professionalism. AI assistants have begun to use the tone and content of business responses as a signal alongside the rating itself.

The practical implication is that review management is not optional for visibility in either traditional search or AI search. Asking customers for reviews at the right moment in the customer journey, and responding to every review promptly, are high-return activities for a small business.

Signal 4: Clear, direct answers on your website

AI assistants are optimised to find direct answers to specific questions. If a user asks "does [business] have parking" and your website has a clear sentence answering that question, an AI assistant can surface that answer. If your website has 500 words of general description without addressing common questions, it cannot.

The most effective format for this signal is a dedicated FAQ page or FAQ section on key pages. Each question should be phrased the way a customer would actually ask it. Each answer should be direct and complete in one to three sentences.

Questions that work well include practical information (parking, accessibility, booking process, cancellation policy, payment methods accepted) and questions that address common concerns for your business type.

A FAQ section also provides the content for FAQPage structured data, which compounds the effect: the answer appears on your website, it is wrapped in machine-readable schema, and it is indexed for both Google search and AI assistant responses.

Signal 5: NAP consistency

NAP stands for name, address, phone. When these three pieces of information are consistent across your website, Google Business Profile, Facebook, Yelp, TripAdvisor, your industry directories, and wherever else your business is listed, search systems and AI assistants can confidently associate all that information with a single business.

Inconsistencies create ambiguity. A business listed as "The Crown Restaurant" in one directory, "Crown Restaurant Ltd" in another, and "Crown Restaurant, Sheffield" in a third may be treated as three different entities or given reduced confidence scores as a local authority.

Common sources of inconsistency include an old address from a previous premises, a landline number that has changed, and the business name used in different formats across platforms. Checking your main listings and correcting inconsistencies is a one-time task with lasting benefit.

Signal 6: Page speed and mobile usability

Google's Core Web Vitals are measurements of real-world page experience: how quickly the main content loads (Largest Contentful Paint), how quickly the page responds to interaction (Interaction to Next Paint), and how stable the page layout is during loading (Cumulative Layout Shift).

Poor performance on these metrics affects Google ranking directly. It also affects user behaviour: a page that takes five seconds to load on a mobile connection loses a significant proportion of visitors before they see any content.

Most small business websites built in the past two to three years on modern platforms perform adequately on Core Web Vitals without specific optimisation. Older websites, particularly those with large unoptimised images, third-party scripts, or outdated themes, often perform poorly.

Testing your website using Google's PageSpeed Insights tool (freely available) gives you a current score and specific recommendations. Many of the issues it identifies can be addressed without rebuilding the website.

Where to start

For most small businesses, the highest return on time is:

1. Complete and maintain the Google Business Profile. This affects more searches than any website change.
2. Add LocalBusiness structured data to the website. One implementation, lasting benefit.
3. Create a clear FAQ page answering the ten most common customer questions. Add FAQPage schema.
4. Check NAP consistency across main directories.
5. Set up a process for requesting and responding to reviews.
6. Test page speed on mobile and address any obvious issues.

None of these require a developer for the first four steps. Step 2 benefits from one, and step 6 may require one if there are technical issues to resolve.