OpenAI is building a way to claim your business in ChatGPT, and the first version reportedly gives verified businesses a daily chart of how often they appear in ChatGPT answers and how many clicks those appearances send to their site. My advice is to claim it the day it shows up, and then go in knowing it will tell you how often you showed up without telling you which questions you showed up for, what ChatGPT said about you, or who it recommended instead.
None of this is launched. OpenAI declined to comment to Digiday, and everything below comes from screenshots and reporting, so treat the details as provisional.
The short version
OpenAI is testing a "Business Hub" where verified businesses can track mentions, keep product and business details current, and set up an agent that answers customer questions (Search Engine Roundtable).
The reported dashboard shows daily impressions and clicks. It doesn't show prompts, user data, or how you rank against competitors.
When OpenAI has shared this kind of data before, the click rate was tiny, well under 1%, so expect a small click line and don't panic about it.
You'll still need your own prompt checks to see what ChatGPT says about you, which is the part that decides whether a buyer picks you.
What was found and who found it
Juozas Kaziukėnas, who founded Marketplace Pulse, posted a screenshot on LinkedIn on November 6 showing a banner for the feature, and he said he'd first seen traces of it back in August (Digiday). According to his post, the tool "measures both impressions and clicks" and plots them by day (Search Engine Roundtable). It sits inside a bigger Business Hub that only opens to businesses that prove they can represent the company, and connecting a Shopify store is one of the ways to verify.
There were earlier signs too. A tipster found the string "Claim {name} on ChatGPT" in a desktop build in October, which The Tech Outlook reported with the warning that app strings aren't a product announcement. Kaziukėnas thinks it's close to launch because he saw a banner for it, and he also pointed out the limit you'd expect, which is that the chart won't explain why your mentions or clicks go up or down.
From what's been reported, this is aimed at organic answers and looks separate from ChatGPT ads. If you're weighing the two, I wrote about what an ad can and can't buy you in ChatGPT a couple of weeks ago.
How it stacks up against Bing and Google's AI reports
So ChatGPT would be the third big AI surface to hand you first-party numbers about yourself, and each of the three answers a different set of questions.
Bing went first. Its AI Performance report opened in public preview on February 10 and shows total citations, average cited pages, a sample of grounding queries, and page-level citation counts across Copilot and Bing's AI summaries (Bing Webmaster Blog). Google followed with a generative AI tab in Search Console, rolled out to every site by August 31, which shows impressions by page, country, device and date, with no click data and no query data (Search Engine Roundtable).
Of the three, Bing is the only one that tells you anything about the questions, and that's why I keep pushing people toward its newer topic view (I went through it in how to measure AI visibility by topic). The ChatGPT dashboard, as described, would be the only one of the three with clicks.
Click through the questions below to see which report answers what.
Interactive · Pick a question
Which AI report answers what
Three first-party AI reports, plus the prompt checks you run yourself. ChatGPT's details are from reporting on an unreleased tool and could change.
Sources: Search Engine Roundtable and Digiday (ChatGPT, Nov 2026); Bing Webmaster Blog (Feb 2026) and Topics Portfolio beta (Oct 2026); Search Engine Roundtable (Search Console, Sept 2026).
Why the click line is going to look small
We have a pretty good idea of what ChatGPT click data looks like, because some of it leaked last year. Vincent Terrasi shared an OpenAI file on LinkedIn showing one top page with 610,775 link impressions and 4,238 clicks, a click rate of 0.69%, and most other pages sat at 0.1% or lower (MarTech). That's one publisher's report from 2025, so I wouldn't treat it as a benchmark, but it tells you what shape to expect.
Now put that next to what Similarweb found when it followed people after ChatGPT recommended a brand. Those users were 2.5 times more likely to visit that brand's site within seven days, and the visits mostly came later, through search, with close to 56% of AI-influenced visits arriving that way compared with about 40% of ordinary visits (Similarweb). So a lot of the value of being recommended shows up as a branded search later on, which ChatGPT's click chart won't ever see.
This matters more now than it did a month ago, because ChatGPT links stopped passing a clean referrer in October and a lot of that traffic is landing in Direct (I covered that in how to track ChatGPT traffic). A first-party click count from OpenAI would help fill that hole. I'd just warn your CFO before the first chart comes out that the click number is going to look like a rounding error, and that it's supposed to.
Interactive · Toggle the view
What a ChatGPT click chart can and can't see
0.69%
Click rate on one top page in a leaked 2025 OpenAI publisher report. Most other pages in the file were at 0.1% or lower.
Source: data shared by Vincent Terrasi, reported by MarTech, Nov 2025. One publisher's report, not a benchmark.
2.5×
People ChatGPT recommended a brand to were 2.5 times more likely to visit that brand's site within seven days, and more of those visits came through search.
Visit within 7 days (relative)
Share of visits via search
Source: Similarweb, June 2026, user journeys in finance, travel and beauty. A click chart inside ChatGPT won't count the later search visits.
What the dashboard can't tell you
An impression, as described, means you appeared in an answer. It doesn't tell you whether ChatGPT said "go with them" or "I'd skip them for a team your size," and those are very different outcomes that would both count as one impression. I don't know how OpenAI will define an impression yet (a brand name in the text, a link, a product card?), and I'd want to know before I put the number in a board deck.
The reason I care about this is the AI Verdict Study I ran in September. Across 576 buyer conversations on four engines, category leaders got advised against 35% of the time and challengers 16% of the time, and in 21 of 72 brand-engine pairings the same buyer got both a yes and a no across runs (AI Verdict Study). A brand in one of those unstable pairings could see a steady impression line while ChatGPT flips its verdict underneath it, and nothing in a daily chart would tell you.
There's also the question of who ChatGPT goes looking for in the first place. A geoSurge study reported by Search Engine Land found models ran searches for brands they already knew 3.2 times more often than for brands outside their top 10. If you're not on that shortlist, a low impression count is the symptom and the dashboard won't show you the cause. I wrote more about how engines build that picture of you in where AI engines go to decide about your brand.
The IAB said in August that only 16% of brands track AI visibility in any systematic way (PPC Land summary of the IAB framework). My guess is a free OpenAI dashboard pushes that number up a lot, and a lot of teams will decide the dashboard is the whole job, which I think is the mistake to avoid.
How to get ready to claim your business in ChatGPT
Before the banner shows up in your account, this is what I'd do.
Number one, figure out who can verify. If verification runs through Shopify for some brands, your ecommerce team probably holds the keys, and if you're not on Shopify you'll want to know who owns your OpenAI relationship (if anyone does). While you're at it, look at OpenAI's merchant page. Product feeds are already open, Shopify and Etsy sellers are integrated, and shopping is live for U.S. users.
Two, get a baseline now. Run your buying questions through ChatGPT several times each and write down whether you're mentioned, whether you're linked, and what it says about you, so when the dashboard arrives you have a "before" that includes the verdict and not only the count. The free Quick Check is built for exactly this first pass.
Three, decide what you'll put next to the impression line. I'd pair it with branded search volume and Direct visits to your AI-cited pages, since that's where Similarweb's data says the follow-up visits go, and I'd check those monthly instead of reacting to a single day.
Then last, pick an owner for the business info and the agent. If the hub lets you edit your product and company details and stand up an agent that answers customer questions, somebody has to keep pricing, policies and claims accurate in there, and in a lot of companies that falls between marketing, ecommerce and support until something goes wrong.
Self-check · 6 questions
Are you ready to claim your business in ChatGPT?
Answer for where your team stands today. You'll get a score and the fixes to start with.
What I don't know yet
A few things I'd love answers to, and I'll update this when OpenAI says more. Does an impression count when ChatGPT names you to warn a buyer off? Will it count mentions without a link? Is it U.S. only at launch, like shopping? Will agencies get access, or only verified owners?
I'm also curious whether OpenAI eventually adds the prompts. Bing gave up a sample of grounding queries, Google gave up nothing on queries, and OpenAI holds a far more detailed record of what buyers ask than either of them. My hunch is they keep that back for a long time, and maybe tie it to ads, but I could be wrong.
The bigger idea underneath all of this, that the engine is forming an opinion about your brand and the count of appearances only tells you part of the story, is a big part of what I wrote about in Explainable. If you want the wider numbers, The Click That Didn't Happen walks through what the 2026 data says about AI visibility.
If you get access to the ChatGPT business dashboard before I do, I'd love to hear what an impression turns out to mean and what your click line looks like. Send me a note, especially if it disagrees with what I've guessed here.
Jarred Smith is the author of Explainable: Why AI Recommends Some Brands & Ignores Others, an Amazon bestseller on AEO, GEO, and SEO. He's a marketing leader with nearly 20 years of experience across healthcare, public media, retail, and environmental services. Find him at jarredsmith.com.