ChatGPT Ads Goes Agentic: Inside OpenAI’s September 16 Ad Stack Overhaul

ChatGPT Ads Shifts From Response Placements to Agentic Conversations

ChatGPT Ads just added three new layers: advertiser-backed agent conversations after a click, prompt-based campaign controls, and native connections to HubSpot and Shopify. The September 16 rollout positions OpenAI’s ad stack as an AI-centered platform rather than a placement network.

The most visible update is the Sponsored Agent. After a user clicks a relevant ad, ChatGPT can now let that user begin a labeled conversation with an advertiser-backed agent inside the same environment.

That conversation is designed to handle the questions that sit between awareness and purchase: whether a table fits a room, how many people it seats, or how its finish should be maintained. OpenAI says the agent exchange is separate from ChatGPT’s independent answers and from the original thread the user started.

What OpenAI Shipped for Advertisers in the September 16 Rollout

As OpenAI announced, the release bundles three distinct feature groups for marketing teams. Each one moves ChatGPT Ads closer to a full-funnel system, not just a response unit.

• Sponsored Agents — a labeled post-click conversation separated from ChatGPT’s own answers.
• Ads Manager plugin — natural-language campaign creation and analysis inside ChatGPT Work.
• AI creative tools — copy and image suggestions, plus opt-in conversational text adaptation and translation.
• HubSpot and Shopify integrations — native ad management and performance tracking in existing systems.

Sponsored Agents: Post-Click Conversations Inside ChatGPT

Sponsored Agents are in testing with select US advertisers. The agent acts as a branded conversational layer after the ad click, using the user’s stated needs to guide product discovery and follow-up questions.

For users, the interaction reduces the need to leave ChatGPT to qualify a purchase. For advertisers, it creates a new post-click engagement path that sits between the impression and the website visit.

Prompt-Based Campaign Management and AI Creative Controls

ChatGPT Work now accepts natural-language commands through the Ads Manager plugin. Advertisers can convert a site or creative brief into a campaign, review performance, and ask for next-step guidance in one place.

Inside Ads Manager, AI assistance can propose ad text and creative based on the advertiser’s landing page and stated objective. Advertisers review and edit those suggestions before anything goes live.

The opt-in text customization layer goes further. It can adjust current headline and description copy to match the flow of a chat and localize the text into a language a user prefers.

HubSpot and Shopify Bring ChatGPT Ads Into Existing Workflows

HubSpot becomes the first CRM partner. Businesses can link their ChatGPT Ads profile, build campaigns, monitor results, and continue lead follow-up from inside HubSpot while using their existing customer data.

Shopify is the first ecommerce partner. Merchants in the US can add the ChatGPT Ads application through Shopify’s app marketplace and launch campaigns from their current product catalog, with international availability beginning September 23 in supported markets.

A $1 Billion Run Rate, a $100 Billion Target, and a Platform Still Under Construction

A day before the rollout, The Wall Street Journal detailed how quickly this business has scaled. The report framed the ad push as central to funding the free ChatGPT tier and offsetting compute costs ahead of a planned IPO.

OpenAI has said the ads business is tracking toward an annualized $1 billion revenue rate. CFO Sarah Friar repeated the number in September and compressed the strategy into a single analogy at a Goldman Sachs event.

‘if Google and Meta had a baby’

That hybrid framing matters for marketers. ChatGPT merges high-intent search with personal context, meaning ads can sit closer to consideration than to broad audience targeting.

The Revenue Rationale: Free Tier, Compute Costs, and IPO Positioning

The ads operation is not a side experiment. OpenAI has told investors to expect a $100 billion revenue trajectory within four years, a threshold that took Meta 17 years to reach.

The company has also said ChatGPT passed 900 million weekly active users, with about 20 percent of prompts carrying direct commercial intent. That reach gives the ad platform a distribution foundation few competitors can match.

According to Digiday, the European footprint expanded August 24 across 31 markets. Consent rules under GDPR meant personalized targeting had to clear explicit permission, making Europe a slower but structurally important test.

Why Buyer Caution Is as Important as Early Performance

Media buyers say the platform is still effective but not as mature as Google’s and Meta’s ad stacks. The rapid build-out has also produced administrative friction and technical hiccups.

Agency teams still describe parts of the product as ‘unproven.’ Victor Batista, senior director of paid media at Jellyfish, has pointed out that the platform live today is no longer the same one released in the pilot phase.

The Marketing Reality: High-Intent Context Without a Native Format Yet

For now, ad units remain relatively simple placements shown under ChatGPT answers for users on free and Go plans. Sarah Friar has noted that a format truly native to AI has not yet shipped.

That split should shape budget decisions. The platform offers high-intent relevance and a growing signal layer, but the creative format is still catching up to the distribution engine.

The AI-Native Ad Format Hasn’t Arrived—but the Infrastructure Has

For marketers, the September 16 update is less about a finished ad format and more about the enabling rails: agentic post-click paths, prompt-based operations, and CRM-native activation. For marketing teams that need to turn AI-native campaign data into search-visible growth loops, Andres SEO Expert’s programmatic SEO and AI automation work is built for exactly this shift — contact us.