Across May 21 and 22, 2026, OpenAI shipped the controls ChatGPT Ads was missing: daily and lifetime budgets, geo-targeting by state, designated market area, and ZIP code, list-view reporting totals, and the first call-to-action buttons inside the ad unit. Most of these are broadly available in the Ads Manager Beta. The CTA buttons are an early test on a small subset of ads.
Each of the last three updates expanded a different dimension of the platform. Access opened on May 5. Audience targeting arrived on May 14. This week brought the budgeting, precision, reporting, and direct-response levers that turn a discovery channel into a channel you can actually run on a performance budget. Here is what shipped, what is still missing, and what to do first.
- Budgets: choose a daily or lifetime budget on new campaigns
- Geo-targeting: U.S. state, DMA, or ZIP code (was country level only)
- Reporting: list-view totals for impressions, clicks, and spend
- CTA buttons: Shop now, Book now, Sign up, Learn more (auto-selected, early test)
- Rolling out or in development: cost-per-action bidding (early access June 5, 2026), lookalike audiences, manual CTA control
What OpenAI just shipped
OpenAI made two announcements in two days. On May 21, 2026, Digiday reported a new ad experience: a larger image and an optional, customizable call-to-action button, plus a dedicated e-commerce format. On May 22, 2026, Search Engine Land reported three Ads Manager updates: daily and lifetime budgets, granular geo-targeting, and list-view totals. Taken together, they are the most substantial single week of ChatGPT Ads product change since the platform opened to self-serve.
The pattern across all six is consistent. ChatGPT Ads launched as a brand-and-discovery surface: you wrote a context hint, set a budget, and your message appeared in relevant conversations. The levers that performance marketers expect, pacing control, geographic precision, in-unit response, and roll-up reporting, were the obvious gaps. This week closed four of them at once.
The new budget controls: daily vs lifetime
ChatGPT Ads now lets you choose a daily budget or a lifetime budget when you create a campaign. Before this update, budgeting was a single campaign-level cap, and daily budget controls had been described as in development for weeks. The choice now mirrors the budgeting model on Google Ads and Meta.
The decision is straightforward once you map it to the campaign type:
- Daily budget suits always-on programs: lead generation, evergreen retargeting, category awareness. You set a steady daily ceiling and the campaign paces against it day to day.
- Lifetime budget suits fixed flights: a product launch, a conference push, a quarter-end campaign with a fixed test budget. You set the total, the platform distributes pacing across the run, and spend stops when the budget is exhausted.
One constraint to note: the budget-type selection currently applies to newly created campaigns. If you want to move an existing campaign from a single cap to a lifetime budget, the cleaner path today is to rebuild it. For the wider economics of how much to put behind a campaign in the first place, the ChatGPT Ads ROI framework walks through the click-to-pipeline math for B2B.
Geo-targeting goes from country to ZIP code
ChatGPT Ads now targets by U.S. state, designated market area (DMA), or ZIP code. This is the single biggest practical change for B2B advertisers in the update. Until this week, the campaign wizard offered country-level targeting as the only geographic lever, which made any regional strategy impossible.
Granular geo unlocks three workflows that were previously off the table:
- Account-based concentration. Layer ZIP-code targeting around the metros where your target accounts are headquartered, so spend concentrates on the regions that matter to your pipeline instead of spreading nationally.
- Multi-location and regional offers. Run different creative or offers by state or DMA when your product, pricing, or service area varies by region.
- Field and event support. Spike geo-concentrated spend in a DMA around a conference, a regional launch, or a field-sales push, then pull it back when the event ends.
You can set geo at campaign creation or adjust it inside an existing campaign, so this is one lever you can apply to live campaigns immediately. The Ads Manager walkthrough covers where the targeting controls sit in the campaign flow.
Country-level targeting made ChatGPT Ads a national billboard. State, DMA, and ZIP targeting make it an account-based tool.
CTA buttons: the first direct-response element
OpenAI began testing call-to-action buttons inside the ChatGPT ad unit, the first direct-response element built into the ad itself. The four options in the test are Shop now, Book now, Sign up, and Learn more, paired with a larger image than the previous unit. For B2B advertisers, Sign up and Learn more are the relevant pair.
The important detail for now is that the button label is selected automatically based on the ad creative and the destination experience. Advertisers cannot choose the CTA themselves yet, though OpenAI has said manual control may come later. In OpenAI's framing, the buttons are a “lightweight feature designed to help users better understand and engage with ads in ChatGPT.” The test is running on a small subset of ads shown to users on the ad-supported tiers.
A direct-response button changes the job of your destination page. A discovery ad earns a curious click; a Sign up or Learn more button earns a click from someone the ad has already nudged toward a decision. The page has to meet that intent immediately. This is the moment to revisit landing pages built for ChatGPT Ads traffic, which arrives with more conversational context than a Google search click.
A click is only as good as the page it lands on. The CTA button moves the decision earlier in the conversation, so the landing page has to meet it there.
Alongside the buttons, OpenAI introduced a dedicated e-commerce format that displays in portrait or landscape and pulls in shopping data such as price and customer reviews, with the portrait version able to stack three or four units into a carousel-style placement. This is more relevant to retail than to B2B SaaS, but it signals where the ad unit is heading: richer, more visual, and more conversion-oriented than the text-led format ChatGPT Ads launched with.
Why these changes matter together
Individually, each update is incremental. Together, they move ChatGPT Ads across a threshold from an awareness channel to a performance channel. A channel you can run on a performance budget needs four things the platform was missing a week ago: pacing control, geographic precision, in-unit response, and roll-up reporting. All four arrived at once.
The sequence since access opened in early May tells the story. Self-serve opened access to every U.S. advertiser on May 5. Custom audiences added first-party targeting on May 14. This week added the operating controls. Access, then audiences, then performance levers, in roughly the order Meta and Google matured a decade earlier, compressed into weeks.
Access made ChatGPT Ads buyable. Budgets, geo-targeting, and CTAs make it a channel you can run on a performance budget. The roadmap that was “coming soon” for months arrived in a single week.
For B2B marketers, the practical implication is that ChatGPT Ads has graduated from an experiment you watch to a line item you can plan. The targeting is precise enough to concentrate on real accounts, the budgeting is controlled enough to cap risk, and the unit can now ask for a response. What it still lacks is the automated optimization layer.
What's still missing
The week closed several gaps but left others open. Knowing which levers are not yet available keeps your plan honest.
- Cost-per-action bidding and outcome optimization. Cost-per-click bidding is live, and CPA bidding reaches early access on June 5, 2026 through conversion-optimized campaigns (set up the pixel or Conversions API by June 1 to qualify). You can also measure cost per action through the conversion pixel and optimize toward it manually until your account has access.
- Lookalike audiences. Custom audiences shipped on May 14, but the algorithmic expansion of a seed list to similar users has not. You can target and suppress your own customer lists today; you cannot yet scale them into a lookalike pool.
- Manual CTA selection. The new buttons are auto-selected. Until advertiser control ships, the way to influence the CTA is through the creative and destination signals the system reads.
- Independent measurement. Third-party measurement and verification are still maturing. For now, the most reliable read comes from your own analytics: the pixel, plus disciplined UTM tagging on every destination URL.
None of these is a reason to wait. They are reasons to plan against the levers that are live, not the ones that might land.
What to do first with the new controls
The new levers are worth roughly half a day of setup on an active account. The order below puts the broadly available controls to work first and prepares for the direct-response shift the CTA buttons signal.
Set budgets to match each campaign's job
Go through your active campaigns and decide, for each, whether it is an always-on program or a fixed flight. Rebuild always-on campaigns on a daily budget and fixed-flight campaigns on a lifetime budget. This is the cheapest way to stop overspending on campaigns that should taper and underspending on campaigns that should run flat out.
Tighten geo before you touch anything else
Country-level targeting was leaking budget into regions that will never convert for most B2B advertisers. Narrow each campaign to the states, DMAs, or ZIP codes where your buyers are. If you run account-based marketing, build a ZIP list around your target accounts' headquarters and apply it as a geographic layer on your prospecting campaigns.
Get the landing page ready for a decided click
The CTA buttons will reach your account on their own timeline, but the destination work is yours to do now. Make sure each campaign's landing page answers the question the ad raises within the first screen, and that the primary action on the page matches the likely button (a Sign up button should not land on a page whose main call is “read the blog”).
Close the measurement loop
Turn on list-view totals so you can read account-level performance at a glance. Confirm the conversion pixel is firing, and tag every destination URL with the canonical UTM convention (utm_source=chatgpt, utm_medium=cpc) so the new clicks classify as paid traffic in your analytics. The free Google Ads to ChatGPT Ads converter drafts context hints and headlines if you are spinning up new campaigns to use the new controls.
Where AI-Advisors fits
The new controls live inside the OpenAI Ads Manager. The work that surrounds them, measuring whether the spend pays back and connecting paid performance to your organic AI visibility, is where AI-Advisors sits. More budgets and a direct-response button mean more clicks; the question is which of those clicks move pipeline.
Three touchpoints connect the new controls to measurement:
- Conversion pixel install. Paste your OpenAI Pixel ID, and AI-Advisors stages the tag inside your Google Tag Manager container for a one-click publish. See the pixel install walkthrough.
- UTM tagging and channel attribution. Consistent UTM parameters on every destination URL keep ChatGPT Ads clicks classified correctly in GA4, so the budgets you set translate into a clean cost-per-pipeline read.
- Paid and organic overlay. The AI Ads platform overlays paid spend with your organic AEO citation data, so you can see where paid ChatGPT Ads and organic AI visibility compound on the same buyers.
New budgets and CTA buttons create more clicks. The AI-Advisors ChatGPT Ads integration installs the OpenAI conversion pixel in two minutes and overlays paid spend with your organic AI citation data, so you can see which spend actually moves pipeline. Free on all plans.
Try the AI-Advisors ChatGPT Ads integration →Paid ChatGPT campaigns convert at higher rates against brands the platform's context model already recognizes from organic citations. The new controls make the paid side easier to run; pairing it with organic AEO is what makes the spend efficient.
Frequently Asked Questions
#What did OpenAI add to ChatGPT Ads in May 2026?
Across May 21 and 22, 2026, OpenAI added five things to the ChatGPT Ads Manager Beta: daily and lifetime budget options on new campaigns, granular geo-targeting by U.S. state, designated market area, or ZIP code, list-view totals for impressions, clicks, and spend, an early test of call-to-action buttons inside the ad unit, and a dedicated e-commerce ad format that pulls price and review data.
#Can you set a daily budget on ChatGPT Ads?
Yes. As of May 22, 2026, advertisers choose between a daily budget or a lifetime budget when creating a campaign in the ChatGPT Ads Manager. A daily budget paces steady spend for always-on programs; a lifetime budget caps total spend across a fixed flight and lets the platform distribute pacing. The option currently applies to newly created campaigns.
#What geographic targeting does ChatGPT Ads support now?
ChatGPT Ads supports targeting by U.S. state, designated market area (DMA), or ZIP code as of May 22, 2026. Before this update, country-level targeting was the only geographic lever. You can set geo at campaign creation or adjust it inside an existing campaign's settings, which makes regional, multi-location, and account-based targeting practical for the first time.
#What are the new ChatGPT Ads CTA buttons?
They are optional call-to-action buttons inside the ad unit: Shop now, Book now, Sign up, and Learn more. OpenAI calls this the first direct-response element built into a ChatGPT ad. During the early test, the button label is selected automatically based on the ad creative and destination, and the unit also features a larger image than the previous format.
#Are ChatGPT Ads CTA buttons available to every advertiser?
Not yet. OpenAI describes the CTA buttons as an early test running on a small subset of ads, shown to users on the ad-supported tiers. The label is auto-selected for now, and manual advertiser control over which CTA appears is something OpenAI says it may explore later. Treat the buttons as a signal of direction, and prepare your landing pages for direct-response clicks.
#Does ChatGPT Ads have CPA bidding or conversion optimization yet?
It is rolling out. Cost-per-click bidding is live, and cost-per-action bidding reaches early access on June 5, 2026 through conversion-optimized campaigns, for accounts that set up the pixel or Conversions API by June 1. The conversion pixel already tracks post-click conversions, so you can measure cost per action through your own analytics in the meantime. Plan against click-based bidding until your account is in the CPA rollout.
#How is this different from the May 5 launch and the May 14 custom audiences update?
Each step expanded a different dimension. May 5 expanded access by removing the minimum spend and opening self-serve to every U.S. advertiser. May 14 expanded targeting depth with custom audiences from your own customer lists. The May 21-22 update expanded the controls and the ad unit itself: budgets, geo precision, reporting, and the first direct-response button. Access, then audiences, then performance controls.
