CDP Institute

Poe, a generative AI platform offering a choice of third-party LLMs, reports that image generation pioneer DALL-E has seen its usage share drop almost 80% as the number of models in the space has grown from three to around 25. Text usage across Anthropic (Claude) and OpenAI (ChatGPT) models has been nearly equal reflecting intense competition.

More than 40% of consumers trust AI search results more than paid search, while a striking 85% trust it more than search ads. Consumer research platform Attest reports that 68% of actual gen AI users trust information from chatbots and other AI tools (an impressive jump from 43% of consumers in general).

Consumers might have less trust in AI search if they read this Newsguard report, which shows Russian sources grooming ChatGPT, Microsoft Copilot and other models by providing them with genuine-looking web content packed with Russian propaganda. A propaganda network called Pravda created 3.6 million such articles in 2024 alone.

Is LinkedIn cracking down on companies that scrape member data in violation of its Terms of Service?  The company hasn’t said anything but two lead gen providers, Apollo.io and Seamless.ai, quietly vanished from the LinkedIn platform on the same day.  Both companies offered Chrome extensions that sales people could use to collect data from LinkedIn.

Composable CDP Hightouch continues to flesh out its identity resolution product.  The latest enhancement, called Golden Record, lets marketers set rules for picking the preferred identifier for each field in the customer profile.  It’s an old concept but not necessarily present in every identity management product.

Dynamic creative platform Jivox has released a new tool, Creative Compliance Checker, that uses AI agents to check ads’ compliance with retail media networks requirements.  Users upload compliance guidelines and ad specifications to the compliance checker in ordinary language. The checker then generates a report, assigns a score to the ad, and identifies rules that have not been followed.

Data unification CDP Reltio can now deliver unified data delivered in less than 50 milliseconds. The Lightspeed Data Delivery Network supports personalized customer experiences at any point of engagement, whether that be a mobile app, a kiosk, or even in a car. Reltio also announced enhanced capabilities in Reltio Data Cloud including data masking for security and bulk-matching to accelerate data unification.

Social customer engagement vendor Emplifi has launched Emplifi Fuel, an out-of-the-box solution that combines voice-of-customer data, brand content, industry knowledge, and generative AI to build automated workflows based on user-specified business objectives. The system measures progress against objectives, suggests tactics, and automates user-selected tactics using Emplifi AI capabilities.

Basis Technologies is integrating with IRIS.TV to support contextual targeting for CTV programmatic advertising. Basis is an advertising automation platform, IRIS.TV a content data marketplace built for streaming. The potential to target ads based on the nature of video content has long been available for linear TV, but not readily for CTV.

Publicis Groupe has acquired data collaboration and connectivity platform Lotame. Lotame will be integrated with Publicis Epsilon, expanding its portfolio of (non-cookied-based) consumer profiles to some four billion. This continues the trend of the large agency holding companies investing heavily in tech.

Microsoft is launching some ad categories designed specifically for the Copilot AI assistant. Microsoft Advertising Showroom ads will create an immersive environment for users to interact with products. Dynamic filters will enable users to refine search results based on their recorded preferences rather than by typing additional search prompts.

Without fanfare, Anthropic, proprietor of the Claude AI model, has withdrawn from voluntary commitments made under the previous U.S. administration. These include security testing models before release, watermarking AI-generated content and research on bias and discrimination (according to the Midas Project). The commitments were quietly removed from Anthropic’s Transparency Hub.