The Drum Podcast

The episode has a handful of interesting observations - LLMs as regressive models, AI search favouring authentic human content, the death of the MQL - but they are buried under extensive repetition, padding, and the same AI-is-moving-fast rhetoric cycling through all three guests. The insight-per-minute rate is low for a 31-minute episode.

“LLMs really are a regressive model, right? So as advanced as they are, it's a regressive model because it is coming to you with what has happened in the past before.”

“Is the MQL dead? You know, is a, is a download of a report. Is that real lead?”

The central arguments - AI slop is bad, vanity metrics mislead, trust matters more than reach - are entirely recycled takes that have dominated marketing discourse for at least two years. The closing observation about traditional media rising in AI search rankings is one genuinely fresh moment, but it is not developed.

“traditional media is all of a sudden more important again and content from renowned media because also they are ending up higher in the AI search rankings”

“use it smarter to tailor the message to the audience to be a lot more bespoke”

All three guests hold genuine C-suite or director-level positions at recognisable agencies (Havas, Team Lewis, APS Group), so they are credible practitioners. However, they are all agency-side generalists speaking abstractly about client challenges rather than brand-side B2B operators with direct P&L accountability, which limits their relevance for an operator audience.

“we've got 25,000 people around the world, right? And most of our competitor agencies are just talking about the technology”

“we work for a company called Cherry Motors and Chinese EV brand and we've been able to launch amazing campaigns and TV ads at a speed that would have normally been, you know, 10 times slower”

There are a few concrete anchors - the Havas 'Science of Desire' study citing 84% brand indifference, the Chery Motors AI production case, and the AVA internal sandbox platform - but the vast majority of the conversation stays at an abstract level with no metrics, timelines, or dollar figures attached to any claim.

“the last round of research we did at a global scale was called the Science of Desire. And it found that, like, there's about. Of all the brands that we looked at as about 84% of brands in that particular piece of research, which basically, you know, people are pretty indifferent on.”

“we work for a company called Cherry Motors and Chinese EV brand and we've been able to launch amazing campaigns and TV ads at a speed that would have normally been, you know, 10 times slower with real actors”

The host sets up topics competently but rarely follows up with precision or challenge; questions like 'what keeps you awake at night?' and 'are you surprised how quickly it's taken off?' are standard roundtable filler. No claim is ever pushed back on and guests are allowed to end on vague optimism without being asked to substantiate it.

“What kind of fears, what fears do you have in the marketing realm? Obviously we don't need to go into ghosts and goblins and ghouls and things like that.”

“Is there any way of, you know, kind of reining that back?”