A website built so an assistant can read it, understand it, and describe your business correctly.

An assistant answering "who should I call for this?" is not browsing your site the way a customer does. It is reading text, and it has to decide two things: what this business is, and whether the page can be trusted enough to repeat. Most sites fail on the first question before the second one is ever asked — a single scrolling page of images and slogans says almost nothing a machine can restate in a sentence.

So the foundation is structure. Schema markup is a block of labelled data (JSON-LD) sitting in the page's code that states, in a fixed vocabulary, that this is an organization, that it is a local business with these hours and this phone number, that these are services it offers, and that these are questions with these answers. A heading structure that goes in order, one clear subject per page, and a title and description written for that page rather than copied across the site do the same job in ordinary prose. An XML sitemap lists every page so nothing has to be stumbled upon, and internal links between your service pages, your area pages, and your answer pages show how they relate.

An llms.txt file is a plain text file at the root of your site that summarizes what the business is and points to the pages worth reading. It is a young convention rather than a settled standard, which is exactly why it is cheap to add and worth having: it costs one file, and it removes any ambiguity about what you want quoted.

Consistency is the part owners underestimate. If the site says one set of hours, the map listing says another, and an old directory has a phone number you stopped using two years ago, a machine weighing those sources has no reason to prefer yours. Making the name, address, and phone number identical everywhere, and claiming the Google Business Profile so the map entry is yours to correct, is unglamorous work that decides whether the rest of it is believed.

The chatbot add-on is the same idea pointed at your visitors. A knowledge base is simply your own information written down in a form the assistant can search — services, areas covered, hours, pricing rules you are willing to state, and the questions you answer on the phone every week. The assistant answers from that and nothing else, which is the point: it is constrained so it cannot invent a price or a policy. It qualifies the visitor, captures a name and a number, and books where booking makes sense. Anything warm, complex, or outside what it was given gets handed to you with the conversation attached, rather than guessed at.

Everything is built in your name. The domain, the hosting account, and the profiles are registered to you, and handover is written down, so the site is an asset you own rather than something rented from whoever built it.