ChatGPT Ads Product Updates: 9 New Features Announced in September 2026

All 9 Updates at a Glance

Conversion Optimization with Impression Billing (Now Generally Available)

This is the headline update. A new campaign type, now generally available, optimizes for conversions from both ad clicks and ad views — rather than click conversions only — while billing stays impression-based. Delivery is tuned toward your conversion goal, and OpenAI calls it the recommended model for maximizing performance.

What changed in practice: previously, view-through conversions were tracked but excluded from optimization and from the Conversions column. Now a campaign can learn from people who saw an ad and converted without clicking, then their impressions become billable events. The trade-off is worth understanding before you switch — you pay for impressions on delivery, including impressions that never convert, and the algorithm decides which placements and hints extract the most conversions per impression.

For teams running a solid conversion tracking setup with the pixel and Conversions API, this is usually the right default. If your conversion volume is thin, the learning engine has less signal to optimize against — give it a full conversion window before judging.

Flexible Attribution Windows

When customizing columns in Ads Manager, you can now choose a 7-, 14-, or 30-day click window and a 0- or 1-day view window for conversion reporting. The preference applies to all conversion columns you display, so every metric in your view — Conversions, CPA, ROAS — uses the same window.

This matters because window choice changes what “performance” means. A 7-day window suits impulse research products; a 30-day window captures the long, considered B2B sale. Pick windows that match your buyer's research cycle, not your competitor's. Note that this changes reportingattribution — bidding still optimizes on the model's learned value.

Granular Platform Targeting

You can now target Android app, Android web, desktop web, iOS app, and iOS web separately in Ads Manager and the Advertiser API. Performance insights can also be segmented by platform.

If your conversion rate is dramatically different between mobile and desktop, this lets you shift budget instead of guesswork — bid more to the platform that converts, or create platform-specific ad copy. It also gives app developers a native path: iOS app and Android app are now first-class placements rather than afterthoughts.

Shopify & HubSpot Integrations

AI-Assisted Ad Creation

When creating an ad, advertisers can now receive suggested copy and imagery based on their landing page and campaign objective, then review and edit the suggestions before adding them to a campaign. The feature is currently in testing, with broader availability expected in the coming weeks.

The key detail is the review step: nothing gets published automatically. Advertisers stay in the loop, which keeps creative control where it belongs. If you want to see the alternative to a hand-rolled copy pipeline, our ChatGPT Ads copy generator does the same job standalone today.

Text Customization & Translation

Advertisers can opt into text customizationthat adapts existing headlines and descriptions to better fit the context of a conversation, and can translate ad copy into a user's preferred language. Advertisers remain in control of whether the setting is enabled.

This is an opt-in, not a default — and that control matters for brand voice. Two practical notes: if your ads already run in multiple languages through manual variants, translation may duplicate or conflict with them, so audit before enabling; and because dynamic translation changes the rendered text, any analysis of “winning copy” becomes fuzzier. Keep your core message tight so that adaptation and translation have less to break.

Event Quality Score

Event Quality Score is now available in Ads Manager under Tools > Conversions. It gives eligible Pixel and Conversions API data sources a 1-10 setup quality score, assesses user data coverage, and provides actionable recommendations to improve conversion event implementation.

A low score is a warning light: malformed payloads, missing currency, or conversion events wired to the wrong page all reduce data quality without breaking reporting loudly. If Score is low for you, check your events against OpenAI's rules — a pixel event debugger catches the exact failures Ads Manager tolerates silently, and the ChatGPT Pixel Helper extension surfaces them live in the browser.

Sponsored Agents (Testing)

OpenAI is testing Sponsored Agents with select advertisers in the United States. After clicking an ad in ChatGPT, users may be able to start a clearly labeled conversation with a business-sponsored agentto ask follow-up questions, explore options, and learn more about the advertiser's offerings.

The ad unit pictured above is the conversation itself — the sponsored agent becomes the ad experience, not a landing page. Two implications: user intent signals get richer (the questions people ask an agent are far more revealing than a page view), and your measurement changes because each conversation is a deep-engagement event worth tracking with a custom event, not just a landing visit.

What This Means for Advertisers

• Move to impression-based conversion campaigns once your conversion tracking is verified — OpenAI states it is the recommended model for performance.
• Match attribution windows to your buyer cycle in Ads Manager before you read any cross-campaign comparison.
• Split platform targeting to expose mobile vs desktop performance instead of averaging it away.
• Treat low Event Quality Scores as the first task, not an afterthought — clean data is what the new optimization model learns from.
• Watch for Sponsored Agents and international Shopify rollout; both change what a “conversion” can mean.

Before any of the optimization updates can work, your conversion events must actually fire — check any URL for the pixel in seconds, then review our full setup guide for the complete picture.

Frequently Asked Questions