ChatGPT Ads Get a June Refresh — Here's What Just Changed

When OpenAI flipped on advertising inside ChatGPT back in February, the kindest thing you could say about the platform was that it was promising. A single ad format — a favicon and some text at the bottom of a response — a six-figure minimum spend, and a buying interface that mostly meant emailing your rep. Four months later, that picture is almost unrecognizable. This week's round of product updates is the clearest signal yet that OpenAI isn't dabbling in ads; it's compressing into months the platform maturation that took Google and Meta the better part of a decade.

Here's what landed, and more importantly, what it means if you're the one spending the money.

The Ads Manager beta is becoming a real ads platform

Most of the updates are aimed at one thing: giving advertisers the operational control they expect from a mature channel. Taken individually they read like routine release notes. Taken together they tell a story.

The headline changes in the Ads Manager beta:

• Switch existing campaigns from lifetime to daily budgets. With daily budgets now rolled out, you can convert campaigns instead of rebuilding them.


• One-click clone to convert CPM campaigns to CPC. Move an impression-based campaign onto cost-per-click bidding with a single button.


• Custom CPM max bids. Impression campaigns used to run on a fixed max bid; now you can set your own.


• Bulk edits directly in the UI. CSV bulk editing shipped recently; now you can make bulk changes to campaigns, ad groups, and ads inside the interface itself.


• Flexible (average) daily budgets. Daily budgets are now enforced as average daily spend with a weekly allocation — the same pacing model Google and Meta use, where a high-opportunity day can spend over budget and a slow day spends under.


• More countries coming. Targeting is live in the US, Canada, Australia, and New Zealand, with the UK, Japan, South Korea, Brazil, and Mexico arriving in the coming weeks.

If you've ever managed a Google or Meta account, none of this is novel — and that's exactly the point. CPC bidding, average-pacing daily budgets, in-UI bulk edits, and a self-serve manager are the table stakes of a performance channel. OpenAI is checking those boxes one by one, in public, at a pace that should make every other ad platform a little uncomfortable.

A couple of these deserve a second look from anyone actually running spend:

Flexible budgets change how you watch your accounts. Moving from hard daily caps to average daily pacing means your spend can run hot on any given day and true up over the week. That's good for performance — the system can lean into your best days — but if you're used to hard caps as a guardrail, you'll want to monitor at the weekly level, not panic at a single day's overspend. Set your average where you'd want the week to land.

The CPM-to-CPC conversion is the tell. OpenAI nudging advertisers off pure impression buys and toward click-based bidding is how a brand-awareness experiment turns into a performance channel. Combined with the conversion pixel and Conversions API that arrived in May — and conversion-optimized campaigns now rolling out — the platform is assembling a genuine direct-response stack. The infrastructure to optimize toward purchases and leads, not just impressions, is now real enough to plan around.

The bigger story: ChatGPT's first real auction

The Ads Manager updates are the maintenance work. The more significant news is buried at the bottom of the announcement: OpenAI is testing multi-advertiser placements on a small slice of ads, showing several relevant ads together in a single spot, priced through a second-price auction.

That's a bigger deal than it sounds. Up to now, a ChatGPT ad placement was essentially a single sponsored result. Putting multiple advertisers into one placement and pricing them through a second-price auction is the moment the system starts behaving like an actual ad marketplace — the same mechanism that underpins Google Search and most of programmatic. It introduces real competition for placement, it gives OpenAI a lever to increase revenue per query, and it quietly tells you where this is headed: more ad load, more competition, and eventually, higher prices.

For advertisers, the read-through is simple. The cheap, low-competition window is the early one. Right now ChatGPT ads sit in a rare spot — a brand-new, high-intent channel before the auction gets crowded and expensive. Multi-advertiser placements are the first crack in that. The brands that learn the channel now, while clicks are inexpensive and competitors are scarce, will have a structural head start when the auction matures and the CPCs climb.

What this means for B2B and considered-purchase advertisers

The case for ChatGPT ads has always rested on intent depth. Google captures intent at the keyword level; ChatGPT captures it at the conversation level — what someone is comparing, what they've already ruled out, where they are in a decision. For considered B2B and high-ticket purchases, where buyers genuinely research before they commit, that's a richer signal than almost anything in the current paid landscape.

But honesty matters more than hype here, so a few caveats worth keeping in front of you:

• Ads currently appear to free and lower-tier users. Plenty of B2B decision-makers live on paid ChatGPT tiers and won't see your ads at all. Size your expectations accordingly.


• Measurement is still maturing. The first-party pixel and Conversions API exist, but third-party measurement isn't fully there yet. Wire up the pixel and Conversions API from day one if you test, and judge results against a clear, modest KPI.


• Creative is constrained on purpose. Favicon-plus-text, no rich media. This channel rewards a specific, intent-matched offer far more than a polished brand campaign.

Our take: treat ChatGPT ads as an experimental line in the media plan, not a core channel — but treat it as a funded, instrumented experiment, not a tire-kick. Scope a small budget, set up measurement properly, write ad copy that answers the exact question a user is asking, and learn the channel while it's still cheap to learn. The platform is being built in public and the rules are still in motion, which is precisely why a deliberate early test beats waiting for the dust to settle.

If you want help deciding whether ChatGPT ads belong in your mix — and how to structure a test that produces real signal instead of burned budget — that's the kind of question we like. Get in touch.