Two very different things now sit on a ChatGPT answer screen. Inside the answer is the organic recommendation: the brands the model chose to name, in the order it chose. Below the answer, for some users, is a Sponsored card: paid placement that is labeled as such and, according to the reporting we rely on, is shown only to a subset of free-tier US users. The first is earned and not for sale. The second is bought and reaches a fraction of the audience. Understanding that split is the whole point of this page.
What the 2026 ads rollout actually looks like
We rely here on third-party reporting from the cloro.dev ads tracker, with data to July 2026. We have not seen OpenAI’s internal numbers and nothing below should be read as official. According to that tracker:
- OpenAI announced ads on January 16, 2026 for logged-in US adults on the Free and Go tiers.
- The first paid ads appeared on February 9, 2026.
- Plus, Pro, Business, Enterprise and Edu subscribers are excluded.
- The format is a “Sponsored” card shown under the answer’s sources.
- There was no self-serve ads manager as of July 2026; inventory was sold by OpenAI’s sales team.
- Pricing moved from a CPM model (reported at roughly $60) to CPC in April 2026.
Each of those details shapes the comparison with organic recommendations, so it is worth taking them one at a time.
Who sees an ad, and who sees only the answer
The audience restriction is the first thing a marketing director should notice. The reported eligibility is logged-in, US, adult, on a free tier. Every other combination sees no ad card: a logged-out user in Chicago, a Plus subscriber in Austin, anyone in London or Sydney or Singapore.
Meanwhile the organic recommendation appears for every one of those users. When a buyer in any country, on any tier, asks “best accounting software for a two-person consultancy”, the names inside the answer are the names that reach them. ChatGPT reached 900 million weekly users as of February 2026; the ad-eligible population is a subset of the US, free-tier, logged-in portion of that.
For a brand whose buyers are international, or skew toward professionals who pay for Plus or Business accounts, this alone can decide the question. The ad channel does not reach them at all.
Position: inside the answer versus below the sources
The second difference is placement. The reported ad format puts the Sponsored card under the answer’s sources, which means under the answer text, under the list of recommended brands, and under the citations. The buyer reads the answer first. If the answer names three competitors and your brand appears only in a card below the sources, you have been introduced after the shortlist was formed.
We have written at length about why being named first, second or third in the answer is the outcome that matters commercially in why position one to three in a ChatGPT answer is the whole game. The short version: a buyer asking ChatGPT for a recommendation is asking for a shortlist, and the shortlist is the answer text. Anything outside it is context.
This is not a criticism of the ad format. It is a labeled, clearly separated placement, which is the right design for user trust. It simply means the ad is not competing for the same slot as the organic recommendation.
Trust: labeled paid placement versus an unpaid answer
Buyers already treat AI recommendations with caution. In a 2026 Idea Grove survey of 1,000 US consumers, only 2% said they would buy from an unfamiliar brand on an AI recommendation alone, 45% said they immediately Google the brand, and 18% go to review sites (citybiz report). The recommendation starts a verification process rather than ending it.
An organic mention inside the answer enters that process with the credibility of something the assistant chose to say. A Sponsored card enters it with the label “Sponsored”. Buyers are used to that label from Google and social feeds, and they apply the same discount to it. Neither is wrong; they are different starting points for the trust the buyer then goes and checks.
Side by side
| Dimension | Organic recommendation (inside the answer) | ChatGPT ad (Sponsored card, per cloro.dev reporting) |
|---|---|---|
| Who sees it | Every user, every tier, every market where ChatGPT operates | Logged-in US adults on Free and Go tiers only |
| Where it appears | In the answer text, as one of the named brands | Below the answer’s sources, as a labeled card |
| How it is obtained | Earned through sources ChatGPT retrieves and what the model learned | Bought through OpenAI’s sales team; no self-serve as of July 2026 |
| Pricing model | No media cost; investment is in legibility, coverage and measurement | Reported move from roughly $60 CPM to CPC in April 2026 |
| What happens when you stop | Visibility decays slowly as sources rotate | Placement ends immediately |
| Buyer perception | Unpaid answer, then verified via Google and reviews | Labeled sponsored, then verified via Google and reviews |
| Control over message | Indirect: you shape sources, the model writes the sentence | Direct: you write the card |
| Measurability | Requires sampling answers at scale (see our method) | Standard ad reporting |
The control row deserves emphasis because it is the one honest advantage of the ad. You write the card. With organic recommendations, you influence what the model says by shaping what the web says about you, and the model composes the sentence. That indirectness is the price of credibility.
Why the organic recommendation is not for sale
This matters because the question we hear most from founders is some version of “can we just pay to be in the answer?” Based on the reporting, no. The ad product is a separate card. The names inside the answer come from two places: pages retrieved live for search-routed prompts, and knowledge the model absorbed in training. We explain both routes in how ChatGPT decides which brands to recommend.
So the levers for the organic recommendation are the ones that affect those two inputs: consistent brand facts across the web, recent third-party coverage that names you in category context, review signals on the platforms the model reads, and structured facts on your own site. None of these can be bought from OpenAI, and none of them should be bought in the form of fake reviews or purchased mentions, which carry their own risks and, in our experience, underperform.
Durability: what each channel looks like in month six
Paid placement has a known shape. Spend, appear; stop, disappear. That is true of Google Ads and will be true of ChatGPT ads.
Organic recommendations have a different shape, and it is not simply “permanent”. Cited sources rotate heavily: a Digital Authority Partners study found only 10.6% of cited URLs persisted across 28 days, with 40 to 60% of cited sources rotating monthly (research roundup). An organic mention built on one article is fragile. An organic mention built on broad, consistent, refreshed coverage degrades slowly and can be maintained with far less than the original effort.
That is why the fair comparison is not “free versus paid” but “a compounding asset that needs maintenance versus a rented placement”. For the economics in detail, including how the two compare with search advertising, see ChatGPT recommendations versus Google Ads.
Where ads genuinely make sense
We are not arguing against ChatGPT ads. There are situations where, if the product is available to you, they are a sensible supplement:
- You already have strong organic visibility and want to add a controlled message (a launch, an offer) to the free-tier US audience.
- Your category is dominated by incumbents in the organic answer and you need interim presence while you build consensus coverage.
- You want a creative-testing surface to learn which framing resonates with ChatGPT users before you shape your organic footprint around it.
In each case the ad is doing a job the organic recommendation does not do: carrying your exact words to a defined audience, now. What it does not do is put you in the shortlist.
Where ads cannot help at all
Two groups are structurally excluded from the ad channel as reported. The first is any brand whose buyers are outside the United States or on paid tiers; the audience simply is not there. The second is any brand in a category that advertising platforms restrict or refuse. If Meta and Google decline your ads, there is no reason to expect a new ad network to be more permissive, and the organic recommendation becomes the only ChatGPT channel open to you. We cover that situation in ChatGPT visibility for brands that cannot run ads.
How to decide for your brand
A simple rule works for most teams. Measure organic first: find out whether ChatGPT names you today across the prompts your buyers use, and where you sit when it does. Our audit does this with about 100 real answers per prompt per market, classified for mention, position and link, repeated monthly; the full approach is on our methodology page. You can get a first impression of what ChatGPT says about your brand at theblueoceangpt.com.
If organic mention rate is near zero, that is the gap to close, because it affects every user. If organic visibility is already solid, ads are a reasonable experiment for the slice of users who see them. Funding ads while the organic answer names three competitors and not you is paying to appear below a shortlist you are not on.
For the buyer behavior that underpins all of this, including why so few ChatGPT referrals show up as clicks and why brand search is the second signal to watch, read the ChatGPT buyer journey versus the Google buyer journey.
What to do next
- Pull your organic baseline across 20 to 40 buyer prompts before you talk to anyone about ad budget.
- Treat ChatGPT ads, where available, as a controlled-message supplement for the free-tier US audience, not as a path into the answer.
- Re-check the reported ad details periodically; the product changed several times in its first six months and will keep changing.
Frequently asked questions
Can you buy a recommendation in ChatGPT?
No. Based on third-party reporting from the cloro.dev tracker, ads in ChatGPT appear as a Sponsored card beneath the answer's sources, separate from the answer text. The brands named inside the answer are chosen by the model from retrieved sources and training knowledge, not sold. Paying for an ad does not change which brands the answer recommends.
Who actually sees ChatGPT ads?
According to the cloro.dev tracker, ads were announced for logged-in US adults on the Free and Go tiers, with Plus, Pro, Business, Enterprise and Edu users excluded. That means logged-out users, non-US users and every paying subscriber see no ad card at all, while all of them still see the organic recommendation inside the answer.
Is there a self-serve ChatGPT ads platform?
Not as of the most recent reporting we rely on. The cloro.dev tracker reports that there was no self-serve ads manager as of July 2026 and that inventory was sold by OpenAI's sales team, with pricing moving from a CPM model of roughly $60 to CPC in April 2026. Treat this as third-party reporting that may have changed since.
Should I run ChatGPT ads or work on organic recommendations first?
For most brands, organic first. The organic recommendation reaches every tier, every market where ChatGPT operates and every session, carries the credibility of an unpaid answer, and compounds. Ads reach a slice of free-tier US users, stop when spending stops, and sit below the answer the buyer actually reads. Ads make more sense as a supplement once organic visibility is measured and climbing.
Do ChatGPT ads help my brand get mentioned organically later?
There is no evidence we can cite that they do, and the reported ad format keeps the Sponsored card separate from the answer. Organic mentions depend on the sources ChatGPT retrieves and what the model learned in training. If an ad campaign drives press coverage or reviews, that secondary effect could help, but the ad itself is not a ranking input.
Find out what ChatGPT says about your brand before you spend anything.
Apply in about two minutes. We check your keywords, then run your real buyer prompts live on a 30-minute call. If we cannot create the result for your keywords, we tell you on the call, not after an invoice.
See if my keywords qualify →