Four changes landed in the ad industry in the same week. Each one is small on its own. Together they suggest the plumbing of paid marketing is getting rewired faster than most teams are tracking it.
ChatGPT now takes your customer list
Advertisers spotted a new Audiences option inside ChatGPT Ads. It lets you upload raw or hashed email and phone lists and use them to include, exclude, or adjust bids against specific segments. The same first-party targeting you already run on Google and Meta.
The minimum threshold is 25,000 matched users before a custom audience can run. Because match rates on ChatGPT are low, the practical advice is to load well over 100,000 identifiers to clear that bar. OpenAI shipped this quietly. Advertisers found it in the interface before any formal announcement went out.
SAS puts ML models in marketing hands
On July 8, SAS launched SAS 360 Marketing AI, a tool built for marketing teams who own customer data but don’t have a data science queue to work through. It ships with guided workflows and reusable recipe templates for churn detection, next-best offer, cross-sell, customer lifetime value, and audience segmentation. Data prep, feature engineering, model training, and monitoring are all built in.
SAS notes that data preparation alone can consume up to 80% of model-building effort. The tool runs standalone or inside the wider SAS Customer Intelligence 360 stack.

Hundreds of ABM ads in under 10 minutes
AI-native paid media agency Multiply launched 10 Minute ABM, also on July 8. The claim: it builds and activates hundreds of personalised account-based ads for a company’s top accounts in minutes rather than weeks. Each campaign is designed to feed insight into the next, so personalisation scales without adding manual work.
The problem it’s targeting is a familiar one in B2B: teams spending weeks assembling account lists and writing creative before a single ad runs. Whether the auto-built creative is strong enough to put in front of named accounts is the question worth testing.
Google now labels AI-made ads
Google added a “How this ad was made” section to its My Ad Center panel, available globally. Anyone can tap the three-dot menu on ads across Search, YouTube, and Discover to see whether AI was used in creating or editing that ad. Ads built with Google’s own generative tools get disclosed automatically. For ads made with other tools, advertisers must now use a new control to declare AI involvement themselves.
The same week, Meta updated its “About this ad” AI tagging on Facebook and Instagram, and TikTok joined the C2PA content-provenance steering committee. A new New York law also requires conspicuous disclosure of AI-generated content in ads. Across the major platforms, declaring whether AI touched your creative is becoming a standing requirement, not an optional transparency gesture.
Quick hits
- ActiveCampaign launched a Google Ads connector so users can run Performance Max from conversational prompts.
- Akeneo’s summer release adds “Agentic Ziggy,” a layer for managing product-data enrichment agents.
- RainFocus rolled out MCP Profiles, piping live event data directly to marketers’ AI agents.
- ChatGPT now commands 92% of AI referral traffic, based on 6.77 million sessions analysed.
- OpenAI can now generate ChatGPT ad creative for advertisers to review, edit, and approve.
- Zynga launched a new partnership aimed at bridging the gap between advertisers and game developers.
- The Rolling Stones’ album promo is targeting a new generation of Roblox fans and creators.
- Convertr appointed a chief product officer to tighten B2B contact-data governance.
