Businesses in ChatGPT: What It Changes for Your Business

On 10 November 2026 OpenAI launched Businesses in ChatGPT in beta. It lets a company claim its profile inside ChatGPT, keep its product and policy information current, see how often it is mentioned, and, if eligible, publish a Business Agent that answers customer questions in a labelled side conversation. Business Hub is the control panel. Business Agents are limited at launch to eligible ecommerce and online service businesses in the United States.

OpenAI puts the audience at 1.2 billion people using ChatGPT each week. That is the reason this matters. It is also the reason the launch is being read more optimistically than OpenAI's own documentation supports.

What Businesses in ChatGPT actually includes

Three distinct things ship under one name, and conflating them is the first mistake.

Business Hub is where you claim and verify the business, then update key details. Retailers can upload product feeds carrying pricing and availability. A claimed profile gets a blue check. Claiming is initially available to eligible businesses with a website.

Insights report the number of mentions and website clicks a business receives in ChatGPT, plus discovery topics, which OpenAI defines as common themes across conversations where your website is cited. These are aggregated and appear only once there are enough conversations to meet a threshold. OpenAI states it does not share users' conversations or Business Agent chats with businesses.

Business Agents are the agentic layer. When your business appears in a model response, a user can open your profile and start a separate conversation with your agent, which answers from your product and policy information under guidelines you set. OpenAI is explicit that this conversation is clearly labelled, separate from the user's original chat, and distinct from ChatGPT's own independent answers. Businesses can sponsor their agent and use it as the destination for ChatGPT Ads, with sponsored placements labelled.

Who can claim, and who cannot yet

Eligibility is the part most coverage has glossed. OpenAI's own FAQ sets it out: eligibility depends on factors including industry, location, and verification and integrity checks. You must own the business or be authorised to represent it.

So there are two different gates, and they are frequently reported as one:

1. Claiming is open to any eligible business located where ChatGPT is available, with a website.
2. Business Agents are available only to eligible ecommerce and online service businesses in the United States.

A business outside the United States can therefore be claimable but agent-ineligible. Access is rolling out, and reports from launch day indicate some claims outside the US were placed in a queue rather than completed immediately, though coverage disagrees on how general that is. If you are outside the US, treat agent access as unscheduled: OpenAI has committed to expanding markets and categories but has published no timeline.

The line almost everyone is skipping

Buried in the product FAQ is the sentence that should frame every decision you make about this:

ChatGPT continues to make independent recommendations and comparisons. Claiming does not guarantee higher placement or sales.

Read that against how the launch is being sold elsewhere and the gap is obvious. Claiming is a data correction and measurement layer, not a ranking lever. It lets you fix what ChatGPT says about you, and it lets you see what is happening. It does not buy you a better position in an answer.

Whether you get recommended in the first place is still decided by the retrieval and citation behaviour that governs every AI engine: whether your content answers the question cleanly, whether your entity is unambiguous, whether your claims are corroborated off your own domain. That is the same work as generative engine optimisation and answer engine optimisation. Nothing in this launch replaces it. If anything, a claimed profile raises the cost of having thin or contradictory content, because now the gap between what you publish and what you assert in your profile is legible to the platform.

The real unlock is measurement, not ranking

Here is the part worth your attention, and it is not the agent.

Until now, measuring AI visibility meant building your own instrumentation: pick a question set, run it against each engine on a schedule, parse the answers, count mentions and citations. We do this for clients, and it is genuinely useful, but it is a sampled proxy. You are inferring platform behaviour from outside the platform.

Business Hub reports mentions and website clicks from inside. That is a first-party number where previously there was only an estimate. Discovery topics go further: themes across the conversations where your site is cited, which is close to query data for a surface that has never had any.

This matters more than it sounds, because of how concentrated AI citations currently are. In our AEO and GEO work with Keyper, the brand reached mentions in 90% of a 73-prompt tracked question set, with 271 citation events in a single month across 69 distinct URLs. Of those citations, 86% came from ChatGPT.

We flagged that concentration as a risk at the time, because resting visibility on one provider's retrieval behaviour is a single point of failure. The same concentration is what makes native ChatGPT reporting disproportionately valuable: if the majority of your AI citations come from one engine, first-party numbers from that engine are not one data source among many. They are most of your picture.

That is the honest case for claiming today, even with no agent and no ranking benefit. You get measurement you could not otherwise buy, and you get to correct your own facts.

How this compares to Google Business Profile

The closest mental model is not a search ranking product. It is Google Business Profile: a free, claimed, verified record that controls the facts a platform repeats about you, carries a trust badge, and reports aggregated interaction data, without buying position in organic results.

The precedent is instructive in one more way. Businesses that treated Google Business Profile as a box to tick got a tidy record and little else. Businesses that treated it as a distribution surface, kept it current, and fed it real inventory and real answers, took share from competitors who did not. The same split is about to happen here, with a smaller field and a much earlier start.

What the competitors are doing

ChatGPT is not the only engine building a merchant-facing layer, and the direction is consistent across all of them: structured product data in, agentic answers out.

Google reaches AI surfaces through Merchant Center and the Shopping Graph, with structured data on your own site feeding what it can assert. Google also has the deepest claimed-profile infrastructure of anyone through Business Profile, which is precisely the pattern OpenAI is now reproducing.

Perplexity runs a merchant programme that lets stores get products into its shopping results, using Google Shopping feed format. Reporting on its current terms, fees and eligibility is inconsistent across secondary sources, so check Perplexity's own documentation rather than a summary, including ours.

Shopify is the quiet winner here. OpenAI states that eligible products from Shopify merchants are already integrated into ChatGPT through Shopify Catalog, and that Shopify merchants can connect from Business Hub or the Shopify admin to streamline verification. If you are on Shopify, the on-ramp is materially shorter than everyone else's.

We have deliberately not published a feature-by-feature fee comparison across these platforms. The figures circulating this week come from vendor analyses that contradict each other on basic points including audience size and transaction fees, and we are not willing to put numbers in front of you that we cannot trace to a primary source.

What to do this week

The contrarian read

The strategic value of this launch is inverted from how it is being marketed.

The agent is the headline and the least important part for most businesses right now, because most businesses cannot have one. The insights are the footnote and the most important part, because they convert AI visibility from something you estimate into something you measure, on the engine that in our own client data produced 86% of citations.

There is also a second-order effect worth naming. Once a platform offers a claimed profile, ads against it, and an agent you can sponsor, the commercial logic points one way. Sponsored placements are labelled and separate today. The organic surface is still independent today. Plan for a world where the free surface gets more competitive, and make sure the thing that earns you a mention, which is genuinely citable content and a clean entity graph, is strong enough to survive that.

The businesses that will do well here are the ones already doing the unglamorous work: unambiguous entity signals, answers structured so a model can lift them cleanly, claims corroborated somewhere other than their own marketing site. Claiming your profile makes that work visible. It does not substitute for it.

If you want to know where you currently stand before you claim anything, that is the measurement problem we solve, and we do it as a fixed-scope engagement rather than a retainer. For ecommerce teams where the product feed is the real asset, the feed and storefront work is usually the higher-leverage half.