Your Business Has No Website. That's Now an AI Problem.

Here is a thing that happened to roughly nobody in 2024 and happens constantly now. Someone stands on a street corner with their thumbs out, types best place for brunch near me that isn't packed, and reads three sentences instead of ten blue links. They do not scroll. They do not compare. They walk to one of the three names in the paragraph.

So the obvious question, the one every owner asks the moment they hear that 45% figure: am I in there? Almost certainly not, and it is worth understanding exactly why before deciding it does not matter.

Being named is rarer than being ranked

The old local pack was a lottery with decent odds. Google surfaced about a third of businesses for a given local query. The assistants are a doorman with a very short list: ChatGPT names 1.2% of locations, Perplexity 7.4%, Gemini 11%.

Where ten blue links gave thirty businesses a shot at a click, three sentences give three. Which means the honest baseline is that you are not named, nobody told you, and there is no notification for it. It is the quietest possible way to lose business. No bad review, no angry customer, no dip you can point at. Just a conversation you were never part of.

Where the answers actually come from

This is the part that surprised me, and it is the part that makes the problem fixable. The assistants are not making it up, and they are not ranking you from some secret reputation score. They are reading something. Yext commissioned Researchscape International to analyse 6.8 million AI citations, the actual sources models reached for when answering location-intent questions.

86% of what an assistant reads to answer a local question is something the business itself controls. Not journalists. Not Reddit. Not luck.

You are not badly positioned in the 44%. You are not in it.

A Google profile is a listing, not a website

Here is where a lot of owners get quietly caught out. If your entire web presence is a Google Business Profile, and for an enormous number of independents it is, then you are competing in the 42% bucket and completely absent from the 44% one.

There is no page for a model to read, no sentence for it to quote, no structured data telling it what you are and where. Half the citation pool for the question your customer just asked does not know you exist.

And unlike almost every other marketing problem, this one has no gradient. There is no partial credit for being on Facebook. Either there is a page with your name on it that a crawler can read, or there is not.

Five things that make a page citable

The good news is that the bar is technical, not creative. Nobody is judging your photography.

1. 01

   It is crawlable and indexable

   No login wall, no JavaScript-only rendering, no noindex left on by a developer three years ago.

2. 02

   Its NAP matches everywhere else

   Name, address, phone, character for character, across Google, Yelp, TripAdvisor and your page. "St." on one and "Street" on another is how a model decides it is looking at two businesses and trusts neither.

3. 03

   It emits LocalBusiness structured data

   Ideally with AggregateRating. This is the difference between a model inferring what you are and being told.

4. 04

   It loads fast

   A crawler on a budget does not wait around.

5. 05

   It says something concrete

   "Family-owned since 1998, gluten-free menu, open Sundays" is quotable. "Welcome to our website" is not. Models lift specifics and skip atmosphere.

That is the whole list. It is achievable in an afternoon, and it is achievable with one page. You do not need a site. You need a page that exists and behaves.

The cheapest version of this

This is the part where I tell you we built the thing, so treat it with the suspicion that deserves. The five conditions above work anywhere: on a Squarespace page, a Carrd one-pager, or a hand-written HTML file on a five-dollar host.

What we built is narrower on purpose. Ansview generates a public page at ansview.app/p/your-business out of data you already have: your name, category, city, rating, review count, recent positive reviews, services. It is server-rendered, indexable, emits LocalBusiness and AggregateRating automatically, and takes a slug you choose. One toggle in your establishment settings, on the free plan.

It is deliberately not a website builder. No blocks to drag, no themes, no blog. If you want to control every pixel, this is the wrong tool and a real site is the right one. What it is good at is going from no page at all to a page that satisfies all five conditions in about ninety seconds, which, if you have been putting off building a site for six years, is the gap that actually matters.

Do this even if you never touch our product

Three things, in order of return.

Audit your NAP across every platform you are listed on. Open Google, Yelp, TripAdvisor and Facebook side by side and make the address string identical. This is free, boring, and the single highest-leverage hour in local search.

Get one indexable page with your name on it. Ours, someone else's, your own. The bucket matters more than the vendor.

Go and ask. Open ChatGPT, type the question a customer in your city would type, and read what comes back. Note who is named. That list is your real competitive set, and it is frequently not the one you assumed. It costs nothing but the sting, and most owners have never once run the query their customers are running.

The uncomfortable framing

For a decade, the answer to "do I need a website?" was honestly: probably not, if Google Maps is working for you. That answer was correct, and it stopped being correct somewhere around the point 6% became 45%.

What changed is not that customers want to read your About page. Nobody wants to read your About page. What changed is that a machine now reads it on their behalf, forms a one-paragraph opinion, and passes it on as advice. 63% of the people receiving that advice say they trust it.

The page is not for your customers any more. It is for the thing your customers ask before they decide.

Questions people actually ask