Who can get my business recommended by ChatGPT?

Why can nobody just get you into ChatGPT?

Because there is no door. There is no submission form, no paid inclusion for the organic recommendation, no partner program that puts your clinic in the answer. OpenAI's own help documentation is blunt about it: "Ranking in ChatGPT Search is based on a number of factors designed to help users find reliable, relevant information. There is no way to guarantee top placement." Google says the same thing about its own results, and has said it for years: "No one can guarantee a #1 ranking on Google."

So the honest version of the question is not who can put you in the answer. It is who can change the inputs the answer is built from. There are only two, and they need different work.

The first input is whether you are even considered. When a question needs local or current facts, ChatGPT runs a live web search before it writes. OpenAI operates a dedicated crawler for that, OAI-SearchBot, and states that sites opted out of it "will not be shown in ChatGPT search answers". The publisher FAQ adds the flip side: "Any public website can appear in ChatGPT search." Google is equally boring about its AI features: there are "no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary". Ordinary indexing is the entry ticket. Nothing exotic.

The second input is whether you get named once you are in the pool. That is a corroboration problem. When several independent sources describe your business the same way, naming it is safe for the model. When your name, address, services or hours contradict each other across your site, your Google Business Profile and three directories, the model quietly names a competitor it can verify. This is the part the label GEO covers, and the part that responds to deliberate work. The academic paper that coined the term (Aggarwal et al., KDD 2024) measured that "GEO can boost visibility by up to 40% in generative engine responses" after making source pages more quotable. Academic benchmark, older GPT-3.5-class engines, so read it as direction and not as a promise you can bank.

Everything below is about who does that work. None of it is secret, most of it is unglamorous, and the differences between providers are mostly differences in what they will and will not touch.

What are the four options, side by side?

Here is the whole decision on one screen. Costs are market figures with sources, not our quotes for you, except the row where we name our own public prices.

Tool prices verified on the vendors' own pricing pages on July 24, 2026, and they change often. We are not affiliated with either and get nothing if you buy them.

Option 1: can you do it yourself?

Most of it, yes. This is the part agencies do not enjoy saying out loud, so here it is with the boundary drawn honestly.

The work that moves the needle first is unglamorous data hygiene. One exact business name, one address, one phone number, identical character for character on your site, your Google Business Profile, and every directory that lists you. A Business Profile filled to 100%, with every service listed and hours that are actually current. Service pages that answer the money question in the first sixty words instead of burying it under a paragraph about your commitment to excellence. Robots.txt that does not block OAI-SearchBot. An owner or office manager working with a competent web person can do all of that, and the ceiling on quality is skill, not access.

Where do-it-yourself usually breaks is not knowledge. It is the two jobs that need a rhythm nobody keeps unprompted:

• Third-party mentions. Reviews with real words in them, local press, community threads, legitimate roundups. This compounds slowly, cannot be rushed, and is the first thing to slide when a busy week hits.
• Monthly measurement. Fifteen prompts, three runs each, across the engines that matter, logged the same way every month. It is two hours of clicking, and it is the only way to know whether any of the rest worked.

If you want the full do-it-yourself checklist rather than the summary, the step-by-step is in our guide on getting cited by ChatGPT, and the wider hire-or-not calculation is in should you hire an SEO agency or do it yourself.

Option 2: can the SEO agency you already pay do it?

Sometimes, and it is worth checking before you add a second vendor. A good SEO agency already owns the entry ticket. They keep you indexed, your local signals clean, your site fast enough to crawl and your content in decent shape. Retrieval runs on search indexes, so that work has not stopped mattering. If your SEO is solid you are further along than you think.

The gap is usually not effort, it is scope. Classic SEO reports on rankings, sessions and conversions. None of those tell you whether ChatGPT names you when somebody asks for the best clinic in your city. If nobody is running prompts, nobody knows.

So ask three questions and read the answers literally:

1. Which prompts do you track for me, and can I see the list with the date it was frozen? A list written after the fact, or edited between reports, is not a measurement. It is a highlight reel.
2. How many runs per prompt, per engine, per month? One run is a coin flip. Three is the floor.
3. Show me the last report including the prompts where I do not appear. The misses are the useful half. An agency that only shows wins is managing your feelings, not your visibility.

If the answers are real, keep them and ask them to extend the scope. If the answers are vague, you have learned something cheap. What actually differs between the two kinds of shop is broken down in GEO agency vs SEO agency.

Option 3: what does a specialist add?

Three things, and it is fair to be skeptical about all of them until a specialist proves each one.

A measurement system, not a screenshot

The core of the job is a frozen panel of prompts real buyers type, run multiple times per engine, logged month after month against a baseline captured before anything changed. That is the only artifact that can prove work is landing. It is also the thing that makes a specialist accountable, because a frozen panel cannot be quietly edited to flatter the numbers. Our exact method, including the fields we log and the mistakes that invalidate a run, is public in how to track whether ChatGPT recommends your business.

Entity work most SEO plans never touch

Schema that resolves your business to one unambiguous entity, sameAs links tying your profiles together, contradictions hunted down across directories, service pages rewritten so a machine can lift a passage without losing the meaning. Worth being precise here rather than mystical: Google states you do not need "new machine readable files, AI text files, or markup" for its AI features, and there is no special schema that unlocks anything. The value of this work is disambiguation, not a cheat code.

Outreach aimed at the sources the answers actually cite

Measurement produces a list of which pages the engines quoted when they answered your buying prompts. That list is the outreach plan: the local roundups, the community threads, the directories that keep showing up. Doing this without the measurement is guessing at which websites matter.

What a specialist cannot add is certainty. If a pitch includes a specific answer on a specific date, the extra thing being sold is confidence, not capability. If you are comparing named firms, we keep an honest comparison of GEO agencies for local businesses where every claim is dated and self-reported numbers are labeled as such.

Option 4: what can a tracking tool do, and what can it never do?

Tracking tools run your prompts on a schedule and tell you what the engines said. They are genuinely useful and priced within reach: Otterly.AI publishes plans at $29, $189 and $489 a month, and Profound publishes $99 and $399 a month billed yearly (both checked on their pricing pages on July 24, 2026). For a single-location business that just wants to know where it stands, the entry tiers cost less than an hour of agency time.

What they do not do is anything. No tool edits your site, corrects a wrong phone number on a directory, publishes a service page, replies to reviews or gets you into a local roundup. A tool converts an unknown into a number. Somebody still has to move the number.

The failure mode is predictable: a business buys a dashboard, watches its share of voice sit at zero for four months, and concludes AI search does not work. The dashboard was fine. Nothing was ever changed. If nobody in your business owns the follow-up work, the tool is a subscription to bad news.

What can nobody promise, and how do you spot who promises it?

This is the section to read twice, because the sales pitches in this category got aggressive faster than the results did. Five claims, each one either impossible or a rebrand of something ordinary.

• A guaranteed answer or a guaranteed spot. Both platforms have written it down. OpenAI: "There is no way to guarantee top placement." Google: "No one can guarantee a #1 ranking on Google." Google's own guidance goes further and tells you to "Beware of SEOs that claim to guarantee rankings, allege a 'special relationship' with Google, or advertise a 'priority submit' to Google." That warning is fifteen years old and transfers to AI search without a single edit.
• A partnership with OpenAI or Google. The documented way to be eligible in ChatGPT search is to let OAI-SearchBot crawl you. That is available to everyone with a website and costs nothing. Anyone selling access to a private channel is selling you a robots.txt line with a markup.
• Getting your brand into the model. Training data is a different bot with a different job. OpenAI documents GPTBot as the crawler used for content that may train foundation models, and separately notes that ChatGPT-User "is not used to determine whether content may appear in Search". Nobody can insert your business into a model's weights on request, and if they could, retraining cycles do not run on your marketing calendar.
• A special file or schema that unlocks AI Overviews. Google's documentation closes this one: "You don't need to create new machine readable files, AI text files, or markup to appear in these features. There's also no special schema.org structured data that you need to add." Structured data is still worth doing for clarity. It is not a lever anybody controls.
• Paying for the recommendation itself. Ads on ChatGPT are real and OpenAI documents a dedicated bot, OAI-AdsBot, that "validates safety of web pages submitted as ads on ChatGPT". An ad is a purchased placement sitting next to the answer. It is not the sentence where the model says you are the best option in town. Those are different products and honest vendors say which one they are selling.

Two more tells that cost nothing to check. First, a single screenshot offered as proof: ask how many runs produced it and on what date. Second, a guarantee with no measurement attached. A guarantee is only meaningful if the thing being measured was frozen before the work started, which is why ours is written against a prompt panel captured on day one and why we say out loud that we guarantee the work and the measurement, never a particular sentence from a particular model.

Why does the answer change every time you ask?

Because these systems are probabilistic, and the same question can produce different names ten minutes apart. This is not a technicality. It decides how you should judge anyone's claim, including ours.

In our own tracking in July 2026, we ran a panel of 15 buying prompts across dental, gym and med spa searches, three runs each across engines, 90 runs in total. Google showed an AI Overview in 42 of 45 runs. For one query in that same panel, it fired in zero of three. Same panel, same day, same method. If we had measured that one query once and stopped, we would have concluded AI Overviews had switched off for the category. If we had measured a different query once, we would have concluded they always fire.

That is the whole argument for a fixed panel with multiple runs. Three consequences follow, and they apply whoever does the work:

• One run proves nothing, in either direction. Not your appearance, not your absence.
• The panel has to be frozen before the work starts. Prompts added later, after you know where you show up, turn a measurement into marketing.
• The number that matters is a share, not an anecdote. In what fraction of runs are you named, who else is named, and which sources got cited.

The full method, including what to log and what invalidates a run, is in our tracking guide. You can run it yourself for free. It takes about two hours a month.

What we do, and what we do not

Since this page is on our site, the honest disclosure belongs here rather than in the footer.

We are a founder-led shop doing GEO and SEO for dentists, gyms and beauty and aesthetics studios in the US and Australia. We freeze a panel of prompts before we touch anything, measure it multiple times per engine, clean up your entity data and Business Profile, rewrite the pages that need to be quotable, and chase the sources the measurement says the engines are actually citing. We report monthly with the misses included, and how the score is calculated is public on our GEO score methodologyBest gym and fitness studio marketing agencies in 2026 page.

Prices are on the pricing page and not behind a call: $1,500 for the deep audit, credited 100% toward month one, then $1,500, $2,500 or $5,000 a month depending on scope. On Growth and above we carry the risk: cited in 90 days against your frozen panel, or we keep working free. We take one business per city, because an AI answer names one or two names and selling that seat twice in the same market would be dishonest.

What we do not do: promise you a specific answer, claim a relationship with OpenAI or Google, sell you a file that unlocks anything, or count a lucky screenshot as a result. If your entity data is already clean and somebody in-house will run the prompts, a $29 tracking tool and this guide may be all you need, and we would rather you knew that before booking anything.