Ads in the Answer Engine: Roger Dunn, CEO, Thrad on Agentic Commerce, ChatGPT Ads & the Snowflake Economy
Roger Dunn, CEO at Thrad, joins Scot to map the new ad layer forming inside AI answer engines: the formats, the economics, and why "keyword jail" is over.
Roger Dunn has watched retail media grow up from the agency side (Mediacom, GroupM, WPP), the tech side (he launched Criteo's retail media business in Australia), and the brand side (a global retail media role at Diageo, home of Johnnie Walker, Guinness, and Baileys). Now he's CEO at Thrad, a Silicon Valley startup building the ad infrastructure for LLMs, "ChatGPT ads, but for everyone else."
Listen/Watch Our Interview:
In this episode, Roger and Scot get into where advertising actually lives inside answer engines, what's working today, and how brands should think about showing up when shopping stops looking like a search box.
Highlights
What Thrad is building: both the supply side (helping retailers and chatbots monetize conversational surfaces) and a self-serve buy side where brands can run ads across existing AI inventory globally, not just in the four countries where ChatGPT ads have launched.
The new ad formats: sponsored messages, sponsored prompts, polling ads, and product carousels, plus the steady stream of formats Google unveiled at Google Marketing Live.
AI-native campaigns: how Thrad turns a URL, a brief, or a prompt into keyword, prompt, and dynamically generated persona targeting, and why creative (human faces beat a bare logo) is where the human edge still lives.
What ChatGPT ads look like right now: CPCs roughly in the $1–$5 range, and why the early “hot mess vs. holy grail” takes are both missing the point.
Roger’s definition of agentic commerce: not a robot doing the final click, but AI shaping every stage of the journey: discovery, comparison, price, review synthesis.
The state of the field: Gemini’s full-court press, Perplexity’s Comet browser, Meta routing UCP checkout into Instagram, the Shopify question Wall Street keeps asking, and where Anthropic/Claude might (or might not) play.
Down Under as a test lab: Woolworths’ Olive, Bunnings’ Buddy, a duopoly retail market, and why Australia keeps getting the pilots.
The advice brands actually need: start small (think $10–20K in a market) but start; get out of “keyword jail”; and the analogy that lands it: keywords are ice cubes, prompts are snowflakes. A few uniform ice cubes everyone fights over, versus infinite unique prompts where a challenger brand can finally own its wedge.
This is a genuinely useful tour of the side of agentic commerce most of us squint at, ads, from someone who has built the playbook three times over. Enjoy the conversation.
🎧 Listen wherever you get your podcasts: Spotify, Apple, etc.
Or, we have a video version on YouTube (make sure to subscribe!) →
Want to Learn More?
Stay on top of what’s happening in agentic commerce, from the tools powering it to the founders building with it by following us across platforms:
Podcast: https://retailgentic.transistor.fm
Instagram: https://www.instagram.com/retailgentic/
We're here to make sense of what’s next in retail + AI and bring you along for the ride. Subscribe and stay tuned!