OpenAI is testing Sponsored Agents inside ChatGPT ads

OpenAI began testing Sponsored Agents with a group of US advertisers on September 16. Click an ad in ChatGPT and the next step can be a separate conversation with an agent sponsored by that business. The thread is labeled. OpenAI says it sits apart from ChatGPT’s independent answers and from the conversation the user started. The user can ask follow-ups, then leave for the advertiser’s site. TNW

The beta can carry conditions such as creative automation, automatic bidding, and a minimum daily ad spend over 28 days. OpenAI The Ads Manager plugin lets advertisers create campaigns, edit them, and read performance with natural-language prompts. It reads a landing page and campaign objective, then suggests copy and images. The advertiser still reviews the work. OpenAI Reuters Existing headlines and descriptions can be rewritten for the conversation and translated into the user’s language. TNW

HubSpot is the first CRM partner, and Shopify is the first ecommerce partner. Starting September 16, HubSpot accounts can connect a ChatGPT Ads account, putting ad creation, performance tracking, and lead follow-up in one screen. US Shopify merchants can run ads from products already connected to Shopify Catalog; the app reaches other ChatGPT Ads markets on September 23. TNW

The channel has been moving fast. OpenAI shifted to cost-per-click pricing in April. It passed a $1 billion annualized revenue run rate on August 31, then started an Amazon advertiser pilot for ChatGPT inventory on September 11. TNW

For an indie product, the first answer becomes part of the ad

For a small product team, this is more than another paid channel to watch. The front door moves. If the click opens a conversation, the hero line and the first answer share the job of converting the visitor. A beautiful ad that promises fast shipping loses force if the agent cannot answer where the product ships and how long it takes.

Shopify lowers the integration work because the catalog is already there. HubSpot lowers the switching cost because ads and lead follow-up sit in the same workflow. The test is easier to start. Messy product data shows up sooner.

Copy also moves into a different rhythm. When the system drafts headlines and images from a landing page, the bottleneck shifts toward review and measurement. The test is plain: put generated variants beside your own copy from the same page. Then compare click-through, conversion, and where the conversation stalls.

Start with one product-answering agent before buying much traffic. Load product data, dimensions, compatibility, shipping, care instructions, and the questions buyers ask every week. Then check which promises in the ad the agent can answer. Run a conversational landing page against the current page with the same ad. See where people leave.

There are still soft edges. A labeled sponsored thread may still be read as ChatGPT’s answer. Advertisers need more data on where chats end and whether the leads pay back the spend. Cost per click can rise if buyers crowd into the new surface, and sellers outside the US still need to check their market timing after the September 23 Shopify app rollout.

The rest of today’s news

  • Anthropic merged Claude chat and Cowork into one interface: documents and slide creation came with it, rolling out to Pro and Max first. TechCrunch
  • Zed opened the Delta public beta: a multiplayer environment for coding with agents and reviewing their work; its own repository runs without pull requests. Zed
  • Google Home opened MCP server early access: agents such as Claude and ChatGPT can control smart home devices and read event history. TechCrunch
  • Mozilla’s Firefox Smart Window beta runs on Mistral models: France and North America first, with the UK and Germany later this year. Mistral
  • Arcee AI closed a Series B above a $1bn valuation: it says the open-weight Trinity family cost about $20m to build. TNW
  • Cohere and Aleph Alpha signed a definitive combination agreement: the combined company keeps the Cohere name, two headquarters and more than 1,000 staff. TNW
  • Google DeepMind launched the DeepMind Institute: Shane Legg, Demis Hassabis and James Manyika direct it, and Legg called an AGI declaration premature. TNW
  • The EU Commission president will invite the frontier labs in: Ursula von der Leyen said she wants to support industry efforts to pace the frontier. Politico
  • The EU competition chief would entertain an antitrust exemption for safety coordination: Teresa Ribera said no formal request has arrived. Politico
  • California made AI ad disclosure law: Governor Gavin Newsom signed SB 1050, which requires explicit labeling when AI-generated performers appear in video or audio ads. Governor of California
  • Bipartisan AI safety talks picked up in the Senate: Ted Cruz wants a Commerce Committee markup this month, and Amy Klobuchar is negotiating a bill on model approvals and imminent-harm alerts. Semafor
  • Vishal Sikka’s Hang Ten Systems added $53M to its seed: five weeks after the first $32M, for $85M total. TechCrunch