AI search readiness for local businesses
When somebody asks an AI assistant for a provider in their town, the answer names a handful of businesses. Here is what decides whether yours is one of them.
Updated September 3, 2026
A local business is asked for by place, not by name. Nobody types your company into ChatGPT before they know you exist; they ask who does this work near them. That makes two things decisive: whether a machine can tell what you do, and whether it can tell where you do it. Most local sites are clear to a human on both counts and ambiguous to a machine on at least one.
The questions the engine generates
These are the shapes of question the audit builds from your own business profile, then checks your pages against. Yours are generated from your site, in your site’s language.
- Which providers are there for childcare in Ulm?
- Which childminders in Ulm are recommended?
- How do you recognise a trustworthy childcare provider?
What the audit reads to recognise this
- LocalBusiness schema and its subtypes, including the specific ones like ChildCare, Dentist or Plumber
- The postal address, phone number and opening hours, and whether they agree across pages
- The service named in your page titles, with the town and the brand affix stripped off
What usually goes wrong
The town appears only in the footer
A place name in small print under the copyright is not a statement that you serve that area. The service and the place need to appear together, in a sentence, on the page about that service.
The schema says Organization, not what you actually are
Generic Organization markup tells a machine you exist. A LocalBusiness subtype tells it what kind of business you are, which is what a local question is matched against.
The offer lives in the logo or a photo
If the only place your service is named is inside an image, it does not exist for a crawler. The audit reads text; so does the model.
Contact details disagree between pages
A phone number that differs between the footer, the contact page and the schema is a trust problem before it is a data problem. Machines resolve the conflict by trusting none of them.
Other site types
- AI search readiness for online shops. Shoppers now ask an assistant what to buy before they ever reach a shop. Whether your products are in that answer depends on what your product pages state in machine-readable form.
- AI search readiness for comparison and finder sites. A comparison site is not a seller, and it should not be tested as one. Here is what answer engines actually ask a finder, and what decides whether it is the source they use.
- AI search readiness for SaaS and software companies. Software buyers ask assistants for shortlists and alternatives. Getting onto those lists depends on facts your marketing site usually states everywhere except in text.
- AI search readiness for service businesses and agencies. Service buyers ask assistants who does this work, who is any good, and what it costs. Most agency sites answer none of the three in a form a machine can use.
- AI search readiness for one-page sites and landing pages. A single page has one chance to say what this is, who it is for and why it is credible. Most one-pagers spend it on a slogan.
The seven signals behind the score are explained in how scoring works, and each one has a guide of its own.
Frequently asked questions
- Do I need a Google Business Profile for AI search?
- It helps, because it is a corroborating source, but it does not replace your own site. Answer engines read your pages directly, and a profile cannot state facts your website contradicts.
- We serve several towns. Do we need a page for each?
- You need one page per town only if you can write something genuinely different on each. Ten near-identical pages with the place name swapped is the classic doorway pattern and reads as thin content to both search engines and models.