Episode 149: Steven Liss of OpenAds on AI start-up, AdCP, UCP, and how The Trade Desk stole his name
Ari Paparo and Eric Franchi sit down with Steven Liss, CEO of OpenAds, to discuss the company’s shift from placing ads inside AI chat to using AI to transform display advertising. They explore why contextual intelligence is gaining new traction, how AI-generated creative performs in practice, and what the new IAB standards AdCP and UCP signal for the future of programmatic buying. The conversation also covers the challenges of embedding interoperability, the complexities of building an AI native DSP, and how agencies are adopting creative generation tools.
Takeaways
-
Ads inside AI chat are constrained by a limited supply, while ads powered by AI offer far more scale
-
New language models improve contextual classification, which leads to more accurate targeting
-
AI creative becomes stronger when grounded in a real webpage context and variety
-
Agencies are driving demand for efficient AI-powered creative, while creative and media teams remain separate
-
Building a DSP from scratch is now possible because coding models accelerate engineering work
-
AdCP helps unify signals between buyers and publishers, which cuts waste in programmatic
-
UCP faces hurdles because embeddings created by different models cannot interpret one another
-
Open source or shared industry models may be needed to support reliable embedding exchange
Chapters
00:00 Introduction and guest setup
02:10 Thanksgiving banter and mailbag announcement
06:40 Stephen Liss joins the show
07:30 The “OpenAds” naming confusion with TTD
10:00 Why ads inside AI did not scale
15:45 Rise of contextual and AI-driven creative
21:00 Building an AI-native DSP
25:25 Examples of surprising AI-generated creative
31:00 ADCP and UCP: standards, embeddings, and challenges
47:20 Why shared models matter + closing thoughts and wrap-up
Learn more about your ad choices. Visit megaphone.fm/adchoices