How to Audit Your Naples Business for AI Search Visibility

Start by Asking the Engines, Not by Reading Your Config

Run the outcome test first. Ask ChatGPT, Perplexity and Google AI Mode for the best provider in your category in Naples, in the plain language a customer would use, and write down whether you are named, which competitors are, and what the answer cites. Configuration checks tell you what is possible. Only the answer tells you what is happening.

Open each engine you care about and ask the question your customer would actually ask. Not your brand name, which proves nothing, but the category question: the best plumber in Naples FL, a pediatric dentist near Old Naples, who does pool service in Collier County.

Record three things for each engine. Whether your business is named at all. Which businesses are named instead. And what the answer cites as its source, because the cited pages tell you what kind of page the engine trusts for that question, which is usually more instructive than the ranking itself.

Repeat the category question two or three different ways. Answer engines are far more sensitive to phrasing than the classic ten blue links, so a single query is a single sample rather than a verdict. If you are named for one phrasing and absent for two others, that is a content coverage problem, not an access problem, and the rest of this audit will come back clean.

Do this before touching anything technical. Everything below explains why an engine cannot name you. If it already names you consistently, your problem is somewhere else entirely and you will have saved yourself the afternoon.

The Crawler You Blocked Is Probably Not the One That Matters

OpenAI documents GPTBot as the crawler for training and OAI-SearchBot as the one used to surface websites in ChatGPT search features (OpenAI, 2026). Google documents Google-Extended as covering Gemini training and grounding, and states it does not impact a site inclusion in Google Search nor is it used as a ranking signal (Google, 2026). Blocking the training crawler is a licensing decision. Blocking the search crawler is a visibility decision.

This is the single most common mistake in the whole audit, and it is entirely understandable, because the crawlers have similar names and the difference is not obvious from the outside.

OpenAI publishes distinct user agents for distinct jobs. GPTBot is described as the crawler used to crawl content that may be used in training generative AI foundation models. OAI-SearchBot is described as the one used to surface websites in search results in ChatGPT search features, and ChatGPT-User covers user initiated actions rather than automatic crawling (OpenAI, 2026). A business that adds a Disallow for GPTBot has opted out of training. It has not removed itself from ChatGPT search, and a business that blocks OAI-SearchBot has done the opposite of what most owners intend.

Google draws the same line in different words. Google-Extended governs whether crawled content may be used for training future Gemini models and for grounding at prompt time, and the documentation states plainly that Google-Extended does not impact a site inclusion in Google Search nor is it used as a ranking signal in Google Search (Google, 2026). Blocking it is a decision about Gemini, not a decision about whether you can rank.

So the useful question is not whether you block AI crawlers. It is which ones, and whether the choice you made is the choice you meant to make. Plenty of Naples businesses inherited these rules from a template, a plugin default or a developer who added a blanket block in 2024, and have never once looked at what it actually does.

Read Your robots.txt the Way a Crawler Reads It

Open your domain followed by /robots.txt and read the served file rather than the one in your repository. Crawlers obey the single most specific group that matches their user agent, so a rule under one agent does not apply to another, and a named group can quietly cancel the protections you thought applied to everyone.

Type your domain followed by /robots.txt into a browser and read what comes back. Read the served file, not the copy in your project, because platforms and CDNs sometimes prepend or rewrite rules and the difference is invisible until you look.

The rule that trips people up is group matching. A crawler follows the one group that most specifically matches its own user agent, and it ignores every other group in the file. A Disallow sitting under User-agent: * does not apply to a crawler that has its own named group higher up. That means adding a named group for a crawler, with the intention of giving it one extra permission, can silently exempt it from every restriction you thought was site wide.

Go through the file agent by agent and write down what each named group actually permits. Pay particular attention to any group naming a search-time crawler such as OAI-SearchBot, since that is the one governing whether ChatGPT search can fetch you at all (OpenAI, 2026). Then check for the accidental blanket rules: a Disallow of a slash on its own, a stray wildcard, or a staging rule that shipped to production.

Two more files are worth checking while you are there. Confirm your sitemap is referenced, check that you are not blocking the crawlers you actually want, and check whether you serve llms.txt, which some engines look for as a plain text map of what matters on the site.

Check What Survives With JavaScript Turned Off

Load your key pages with JavaScript disabled and see what content remains. When we measured 80 Naples businesses in July 2026, 12 of the 69 we could assess served content that depended on JavaScript and 4 blocked AI crawlers outright, so 16 of 69, about one in four, were effectively unreadable or off limits to AI search engines (SEO Elite Agency, 2026).

Access is only half the problem. A crawler that is allowed in still has to find words on the page, and many AI crawlers do not execute JavaScript the way a browser does.

The check needs no tools. Turn JavaScript off in your browser settings and reload your homepage and your two most important service pages, or view the page source and search it for a full sentence you can see on the screen. If the sentence is not in the source, an engine that does not render JavaScript is not reading it either.

This is where the local numbers get uncomfortable. Across 80 independent Naples businesses sampled in July 2026, 16 of the 69 we could successfully measure, about one in four, were effectively unreadable or blocked to AI search engines, and JavaScript dependence was the larger of the two causes (SEO Elite Agency, 2026).

The reason it stays hidden is that classic Google renders JavaScript. Rankings hold, traffic holds, and nothing anywhere warns the owner. The site looks healthy by every measure they are used to checking, while the assistants their customers increasingly ask are receiving a nearly blank page.

Make Sure the Engine Can Tell Which Naples Business You Are

An engine that can reach and read your site still has to resolve who you are before it will name you. Check that your name, address and phone match exactly across your site, your Google Business Profile and the directories that list you, that your pages carry LocalBusiness structured data, and that something other than your own website says the same facts.

Answer engines are cautious about naming a specific business, because naming the wrong one is a visible failure. What resolves that caution is corroboration, and corroboration is mostly not on your website.

Start with identity consistency. Your business name, street address and phone number should match character for character across your own site, your Google Business Profile and every directory that lists you. Suite numbers, abbreviations and a second phone number that appears in one place and not another are the usual culprits. If two sources disagree, an engine has to guess which is right, and the safe guess is to name a competitor it is more certain about.

Then check the machine layer. Your pages should carry LocalBusiness structured data that states the same address and phone as the visible page, links your Google Business Profile, and does not contradict itself between pages. This is the part an engine reads first and the part almost nobody checks after launch.

Finally, look off your own site. Count the places outside your domain that confirm what you do and where you are: your Google Business Profile, reviews with recent dates, local directories, a chamber listing, any local press. This is the weakest column for most Naples businesses and it is the one that decides borderline cases, because an engine weighs what others say about you more heavily than what you say about yourself.

If the audit ends here with everything technically clean and you are still not being named, that is a genuine and useful result. It means the constraint is corroboration rather than access, and the work is earning mentions rather than editing code.