ChatGPT Ads Manager Is Live: How to Advertise on ChatGPT in 2026
On July 22, 2026, OpenAI opened "Advertise in ChatGPT" to the public — a self-serve Ads Manager where any business can build a campaign, set a budget, and place sponsored recommendations inside ChatGPT's answers. The pilot that started as a cautious February experiment now has its own campaign console, complete with CPC bidding and reporting, and brands like Best Buy, Lowe's, and VistaPrint were already running on it at launch. For marketers, the abstract question of "will ChatGPT ever sell ads" is over. The concrete question is now operational: how do you actually run a campaign inside a chatbot, and is it worth a slice of your budget?
This guide walks through what launched, how ChatGPT's conversational-context targeting differs from keyword search ads, how to set up your first campaign, and the traps to avoid before you commit spend to an inventory type that did not exist eighteen months ago.
What actually launched on July 22
The July release turned a limited pilot into a general-availability advertising product. A closed test of the OpenAI Ads Manager began on April 7, 2026, went publicly live with CPC bidding for US businesses on May 5, and the July 22 launch expanded access and refined the targeting engine. The headline pieces:
- A self-serve console. Advertisers create campaigns, set budgets and bids, and manage creative themselves — no sales rep or minimum-spend gatekeeper required, the same shift that made Google and Meta ads accessible to small businesses.
- Conversational-context targeting. Instead of bidding on two-word keywords, campaigns are matched to the context of a live conversation — what the user has described they are trying to do — which OpenAI positions as higher-intent than a query string.
- Geographic expansion. Alongside the US, the ads pilot expanded into the UK, Mexico, Brazil, Japan, and South Korea, and OpenAI began testing multi-advertiser units that group several relevant ads into a single placement.
- Named early adopters. Best Buy, Lowe's, and VistaPrint ran campaigns at launch; ad-tech partner Criteo reported more than 2,000 brands now running on ChatGPT through its platform.
This is the paid half of a story we covered when the split first emerged. Our earlier breakdown of AI search advertising in 2026 explained why ChatGPT went all-in on ads while Perplexity walked away entirely — worth reading first if you want the strategic backdrop before the tactical steps below.
Why conversational-context targeting changes the math
Traditional search advertising infers intent from a handful of typed words. A ChatGPT conversation carries paragraphs of it. By the time someone asks "which of these CRMs works for a 10-person agency that already uses Slack," the model knows the team size, an existing tool in the stack, and the shortlist under consideration. A sponsored answer landing at that moment is not interrupting research — it is the research.
Best Buy's VP of Media framed the appeal as meeting customers "in those moments" of decision-making that increasingly happen inside a chat window rather than a search bar. That is the real reason budgets are moving: the ad reaches a user who is not scrolling to pass time but deliberating toward a purchase. It also explains a metric OpenAI has started sharing — the ad-dismissal rate has fallen by roughly half since ads were first integrated, which the company reads as a sign that contextual relevance keeps the placements from feeling like spam.
ChatGPT ads vs. traditional search ads
If you plan media the way a search marketer does, here is how the new channel maps against what you already run.
| Dimension | ChatGPT Ads Manager | Traditional search ads |
|---|---|---|
| Targeting signal | Full conversational context | Keyword match on a short query |
| Ad format | Sponsored recommendation woven into the answer, labeled | Text/shopping unit above or beside results |
| Buying model | Self-serve, CPC bidding | Self-serve, CPC/CPM bidding |
| Funnel stage | Discovery and evaluation (mid-funnel) | Often bottom-funnel, high commercial intent |
| Inventory maturity | New, relatively cheap, thin measurement | Mature, competitive, deep measurement |
| Creative that wins | Answers a described problem usefully | Matches the query and a strong offer |
The practical read: ChatGPT ads behave more like a paid discovery channel than a demand-harvesting one. They enter the shortlist while preferences are still forming — which is exactly the stage where, as our look at the 2026 buying journey showed, AI tools now dominate.
How to set up your first campaign
The workflow will feel familiar to anyone who has used Google Ads or Meta Ads Manager, but the creative logic is different. A sensible first run:
- Start with one high-intent category. Pick a product or service where buyers ask detailed, comparison-style questions — the richer the conversation, the better the context match. Do not spread a small budget across everything.
- Write answer-style copy, not slogans. The unit sits inside a helpful answer, so the creative that earns a click reads like a genuinely useful recommendation to the question being asked. A shouty headline breaks the format and gets dismissed.
- Set a modest CPC budget and a clear test window. Treat it as an experiment with a fixed cap. Early inventory is underpriced, but "cheap" is not a reason to run it open-ended without a read on performance.
- Measure assisted conversions, not last click. Because this is a discovery-stage placement, judge it on whether it introduced qualified users who converted later — not on immediate last-click sales. Wire the landing pages into AI-traffic tracking in GA4 so you can actually see the assist.
- Keep your product data machine-readable. Whether the model surfaces a paid or an organic recommendation, it needs parseable prices, specs, and availability in HTML and
Productschema — not trapped in images or behind forms.
What to watch before you scale
The launch is real, but the channel is young, and three cautions apply. First, measurement is still thin relative to Google and Meta — attribution windows and reporting granularity are evolving, so do not scale spend faster than your ability to prove it works. Second, disclosure and brand-safety rules are still settling across the industry; sponsored answers are labeled, but the norms around AI-mediated recommendations are being written in real time, a maturation visible in Google's parallel AI-ad disclosure rules. Third, paid never replaces earned. A sponsored answer sits alongside the citations the model chooses on merit, and platforms like Perplexity have no ads at all — so the organic path still matters everywhere.
That last point is the one most teams get wrong. The smartest budgets treat the Ads Manager as an accelerant on top of a strong organic foundation, not a shortcut around one. Our guide to optimizing for AI search covers the citation mechanics that decide which brands get quoted for free — the visibility you keep whether or not you are bidding.
Is it worth it right now?
For most advertisers, yes — as a small, disciplined test rather than a budget reallocation. The case for moving early is straightforward: the audience is enormous and high-intent, the targeting signal is richer than anything keyword search can offer, and the inventory is priced for a land grab that will not last. The case for restraint is equally real: reporting is immature, the format punishes lazy creative, and no single platform should ever control your funnel. Run a capped experiment on your best category, hold your organic investment steady, and let assisted-conversion data — not the novelty — decide whether it earns a bigger line next quarter.
Frequently asked questions
When did the ChatGPT Ads Manager launch?
OpenAI opened its self-serve "Advertise in ChatGPT" Ads Manager to the public on July 22, 2026. It followed a closed test that began April 7 and a public CPC launch for US businesses on May 5, 2026; the July release expanded access and refined the targeting engine.
How does ChatGPT ad targeting work?
Instead of bidding on short keywords, campaigns are matched to the conversational context of a live chat — what the user has described they are trying to accomplish. OpenAI positions this as higher-intent than a query string because the model already knows the specifics of the problem before an ad appears.
Which countries can run ChatGPT ads?
At the July 2026 launch the ads pilot covered the United States plus the UK, Mexico, Brazil, Japan, and South Korea. OpenAI also began testing multi-advertiser units that group several relevant ads into a single placement, and availability is expanding.
How much do ChatGPT ads cost?
Campaigns run on a self-serve CPC bidding model, so you set your own budget and bids. Early inventory is generally underpriced relative to mature search channels, but pricing will rise as demand catches up — the standard pattern for a new ad platform's early phase.
Do ChatGPT ads replace SEO or GEO?
No. Paid placements sit alongside the organic citations a model chooses on merit, and ad-free engines like Perplexity can only be won through earned visibility. Generative Engine Optimization remains the common denominator across every AI surface; the Ads Manager is an accelerant to layer on top where it is available.
