Everyone can make AI video now. Very little of it is worth spending money on. Over the past year UGC became every marketer's obsession, and most brands, B2B and B2C alike, were daunted by the cost and operational complexity of finding, hiring, and managing a large roster of creators. Video generation models answered that fever directly, and then handed us a new problem. Producing content stopped being the bottleneck. Deciding what deserves budget became one. This session breaks down how to use AI video to scale an organic and paid social content engine, across both UGC-style ads and brand campaigns. Drawing on live campaigns where AI-generated video was the primary creative supply, the speaker covers what performed, what quietly burned budget, and where human judgment and creativity remain non-negotiable. Expect real examples, real numbers, and honest faux pas from building this at scale. Key learnings: - Which parts of the content workflow AI genuinely supercharges, and which parts still cost more time than they save - How AI creative compares to traditional UGC on the metrics that decide budget: hook rate, cost per acquisition, and return on ad spend - How to tell early which AI assets deserve spend, and where human taste still changes the outcome

Every conference has the sessions where speakers show you the polished version — the campaigns that worked, the tools that delivered, the transformation that went smoothly. This isn't one of those sessions. This panel is built around the conversations that usually only happen in the bar afterwards - where the wins are real, the failures are honest and nobody is trying to sell you anything. Four senior marketers. No vendor sponsorship. No slide decks. Just an unfiltered conversation about where AI is genuinely moving the needle in their marketing functions, where it has consistently disappointed, where the hype hasn't matched the reality and what they'd do differently if they were starting their AI transformation today. What you'll learn: - Where AI is delivering measurable, repeatable value in real marketing functions right now — and the specific use cases worth prioritising - Where it's falling short - the workflows that looked promising, the tools that underdelivered and the assumptions that didn't survive contact with reality - What the panellists know now that they wish they'd known at the start and what that means for how you approach your own AI investment