ChatGPT Ads automated bidding becomes default for new ad groups: what it means for control and costs

OpenAI has rolled out an automated bid strategy called Maximize results for ChatGPT Ads, and it now arrives preselected as the default for eligible new ad groups. The strategy optimizes for click or conversion volume within your budget, but it does not guarantee any cost per action (CPA), cost per click (CPC), or return on ad spend (ROAS) target. Media buyers and product teams should not assume defaults are optimal.

Maximize results is the new default

The bid strategy is selected automatically for new ad groups unless you choose otherwise in the Bid strategy section. According to OpenAI's help center, it is designed to "return as many results as possible from a given budget" by setting and adjusting bids on your behalf toward the campaign goal of clicks or conversions. There is no account-level setting to opt out; the choice resets per ad group. This means advertisers who build multiple ad groups must manually switch each one to manual bidding if they want a cost ceiling.

No cost efficiency guarantee

The most important limitation is explicit: "Maximize results prioritizes result volume and does not guarantee delivery against a specific CPA, CPC, ROAS, or other cost-efficiency target at this time," according to OpenAI's documentation. This separates it from Google Ads Target CPA/ROAS or Meta's Cost Cap, which enforce a numeric target. With Maximize results, the budget is the only constraint, and whatever cost per outcome emerges is the cost. Roadmap language ("at this time") suggests efficiency targets may ship later, but they are absent today.

Manual bidding still available

Manual: Max bid remains available in the ad group settings and is the only documented way to impose a hard ceiling on what you pay per click or impression. Advertisers who need cost control should actively select this option for each new ad group.

Budget and measurement shifts

Two recent changes compound the control loss. Fixed daily budgets were converted to seven-day averages in July without an opt-out, meaning any single day can overshoot or undershoot the daily target. View-through conversion reporting was added in the same rollout, but OpenAI siloed it from the main conversions total, from cost per acquisition, and from bidding. Post-view credits show in a column but do not influence the algorithm. An advertiser reading rising view-through numbers as proof of performance is reading a metric the automation cannot see.

Market expansion and early adopter risks

The pilot reached nine markets on August 17, 2026, including Mexico and Brazil, with European inventory scheduled to open. Minimum spend dropped from $200,000 to zero over six months, and reported CPMs fell from $60 toward $25. The platform is now open to self-serve advertisers who may lack historical benchmarks for ChatGPT Ads. Without conversion tracking or a working pixel, the automation optimizes against weak signals, and automatic advanced matching became default on existing web pixels on August 17 after a short opt-out window.

What builders should watch

For AI product teams using ChatGPT Ads or building ad tools on top of it, the key takeaway is that the platform now defaults to volume-focused automation with no cost guardrails. You need to actively manage bid strategy per ad group, set up robust conversion tracking, and monitor costs manually. The lack of transparency into bidding signals and pacing logic means unexpected delivery patterns are hard to diagnose. As OpenAI continues shipping features every three weeks, expect more automation defaults and measurement changes.

If you need cost predictability, Manual: Max bid is your only option. And if you stay with the default, remember the fine print: volume is the target, not efficiency.