Welcome to P3 Media’s AI Commerce Brief, your daily update on the AI and commerce stories shaping how companies build, sell, and grow. It’s Tuesday, October 6. Let’s get into it.
OpenAI is expanding visual advertising inside ChatGPT.
In a product post dated Monday, the company said it will begin testing a new image-based ad format later this month with an initial group of US advertisers.
The ads will first appear during image generation. OpenAI says they will be clearly labeled, remain separate from the generated image, and will not influence ChatGPT’s answers.
For marketers, the bigger signal may be measurement.
OpenAI announced conversion-data integrations with Hightouch, Tealium, and LiveRamp, plus support for attribution partners including AppsFlyer, Triple Whale, Adjust, Rockerbox, Northbeam, and Branch.
It is also developing brand-suitability pilots with DoubleVerify and Integral Ad Science.
The commercial implication is straightforward: OpenAI is assembling more of the infrastructure advertisers expect before they move meaningful budgets into a new channel.
TikTok also used Advertising Week New York to push deeper into AI-assisted commerce and media.
It introduced Buy Direct, an in-app checkout experience for buying from a brand without leaving TikTok, and a Shopping Assistant that the company says can answer questions about products, shipping, sizing, availability, and purchasing in one conversation.
TikTok says it is building those experiences with commerce and payments partners including Salesforce, Shopify, Shoplazza, and Stripe.
For merchants, the main issue is control: faster checkout may reduce friction, but brands will need to watch how customer data, attribution, and economics work in practice.
TikTok also announced its Ad Network for US advertisers, extending campaigns beyond TikTok across what the company describes as nearly 400,000 apps.
It added automated search-ad buying and expanded GMV Max so TikTok Shop sellers can factor affiliate commissions, coupons, and referral fees into return-on-investment optimization.
That brings media buying and marketplace economics closer together.
In enterprise AI, Cohere launched North 2.
The company says the platform adds a redesigned orchestration system for multi-step agents, reusable skills, shared libraries, memory, app creation, and workflow automation.
North 2 is model-agnostic and can run in cloud, private, on-premises, or air-gapped environments.
It also adds admin controls for permissions, usage, token spending, and rate limits.
The product is aimed at regulated organizations that want agentic tools without giving up deployment control.
OpenAI separately laid out its approach to text provenance under the EU AI Act.
API customers globally can now opt in to watermarking for select models, with the feature off by default.
OpenAI says eligible ChatGPT and Codex text outputs in the European Union will receive an invisible watermark over the coming weeks.
Access to its detector will initially be limited to approved researchers and expert organizations.
The company also warned that editing can weaken detection and that a missing watermark does not prove human authorship.
The takeaway for compliance teams is to treat watermarking as one signal, not a definitive authorship test.
In global model news, Aleph Alpha released Kolibri, a German-and-English language model developed and trained in Europe for public-sector and industrial use.
The company says the mixture-of-experts model has 78 billion total parameters, with about 3 billion active per token, and can be deployed on customer-controlled infrastructure.
Kolibri became available on October 3.
Its release keeps sovereign deployment and local-language performance at the center of Europe’s AI competition.
Here’s what to watch.
OpenAI’s US ad test will show whether image-generation sessions become a durable commercial surface. TikTok’s new tools will need real merchant results beyond company descriptions.
And European organizations will now begin testing how usable text-watermarking is once content is edited, translated, or shortened.
That’s your AI Commerce Brief for today. Thanks for listening.
OpenAI is expanding visual advertising inside ChatGPT.
In a product post dated Monday, the company said it will begin testing a new image-based ad format later this month with an initial group of US advertisers.
The ads will first appear during image generation. OpenAI says they will be clearly labeled, remain separate from the generated image, and will not influence ChatGPT’s answers.
For marketers, the bigger signal may be measurement.
OpenAI announced conversion-data integrations with Hightouch, Tealium, and LiveRamp, plus support for attribution partners including AppsFlyer, Triple Whale, Adjust, Rockerbox, Northbeam, and Branch.
It is also developing brand-suitability pilots with DoubleVerify and Integral Ad Science.
The commercial implication is straightforward: OpenAI is assembling more of the infrastructure advertisers expect before they move meaningful budgets into a new channel.
TikTok also used Advertising Week New York to push deeper into AI-assisted commerce and media.
It introduced Buy Direct, an in-app checkout experience for buying from a brand without leaving TikTok, and a Shopping Assistant that the company says can answer questions about products, shipping, sizing, availability, and purchasing in one conversation.
TikTok says it is building those experiences with commerce and payments partners including Salesforce, Shopify, Shoplazza, and Stripe.
For merchants, the main issue is control: faster checkout may reduce friction, but brands will need to watch how customer data, attribution, and economics work in practice.
TikTok also announced its Ad Network for US advertisers, extending campaigns beyond TikTok across what the company describes as nearly 400,000 apps.
It added automated search-ad buying and expanded GMV Max so TikTok Shop sellers can factor affiliate commissions, coupons, and referral fees into return-on-investment optimization.
That brings media buying and marketplace economics closer together.
In enterprise AI, Cohere launched North 2.
The company says the platform adds a redesigned orchestration system for multi-step agents, reusable skills, shared libraries, memory, app creation, and workflow automation.
North 2 is model-agnostic and can run in cloud, private, on-premises, or air-gapped environments.
It also adds admin controls for permissions, usage, token spending, and rate limits.
The product is aimed at regulated organizations that want agentic tools without giving up deployment control.
OpenAI separately laid out its approach to text provenance under the EU AI Act.
API customers globally can now opt in to watermarking for select models, with the feature off by default.
OpenAI says eligible ChatGPT and Codex text outputs in the European Union will receive an invisible watermark over the coming weeks.
Access to its detector will initially be limited to approved researchers and expert organizations.
The company also warned that editing can weaken detection and that a missing watermark does not prove human authorship.
The takeaway for compliance teams is to treat watermarking as one signal, not a definitive authorship test.
In global model news, Aleph Alpha released Kolibri, a German-and-English language model developed and trained in Europe for public-sector and industrial use.
The company says the mixture-of-experts model has 78 billion total parameters, with about 3 billion active per token, and can be deployed on customer-controlled infrastructure.
Kolibri became available on October 3.
Its release keeps sovereign deployment and local-language performance at the center of Europe’s AI competition.
Here’s what to watch.
OpenAI’s US ad test will show whether image-generation sessions become a durable commercial surface. TikTok’s new tools will need real merchant results beyond company descriptions.
And European organizations will now begin testing how usable text-watermarking is once content is edited, translated, or shortened.
That’s your AI Commerce Brief for today. Thanks for listening.
