📈 Meta taps AI chats, OpenAI’s Sora 2, & YouTube’s 200× AI-search edge

Meta says from 16 Dec it’ll fold what you type or say to Meta AI into ad targeting and feed recommendations. There’s no opt-out, though it applies only if you actually use Meta AI. Rollout excludes the UK, EU and South Korea for now due to local laws. Meta claims sensitive topics (health, politics, etc.) won’t be used for ads. Marketers should expect new intent signals to flow into ad delivery and Reels placement.

Why it matters: First-party “chat intent” becomes a targetable signal at massive scale. So plan tests, but review consent flows and regional exclusions. 📰 Reuters

OpenAI is prepping a standalone app built on Sora 2, with a vertical feed, swipe UI, and a For You page, except the videos which are model-made. Identity-verification features let people manage likeness usage, and the company is beefing up content-owner controls and monetisation paths. The move lands as rivals push AI video (Meta’s Vibes, Google’s Veo on YouTube).

Why it matters: If AI video feeds gain traction, brands will need new creative ops for prompt-native ads, rights management, and UGC-style iterations at speed. 📰 Wired

TikTok rolled out an AI-assisted travel ad product that pairs brands’ catalogues with its travel-intent model. The format can auto-generate creatives, set up campaigns, and optimise delivery to highlight hotels, flights and destinations. Early positioning: more relevance and engagement with less manual build.

Why it matters: Travel’s back…and performance teams get lower-lift, AI-assisted creative/targeting loops on the platform Gen Z actually uses to dream and book. 📰 Media Post

Fresh analysis shows YouTube is overwhelmingly the video source of record cited by AI search systems (Google AI Overviews/Mode, Perplexity, ChatGPT). For marketers, it means AI surfacing is increasingly tied to YouTube presence, structure, and metadata.

Why it matters: Treat YouTube like an SEO property: consistent publishing, chapters, transcripts, and authoritative channels will compound AI visibility. 📰 Search Engine Land

Banijay, the producer behind Big Brother and MasterChef, is using AI to index speech, scenes and sentiment across its archive to pick high-performing clips for social platforms. Trials cut clip selection time by -80% and costs from $800 → $10 per clip, while expanding digital ad monetisation.

Why it matters: Legacy content libraries + AI retrieval = scalable, profitable social programming. Brands with archives should copy this playbook. 📰 Business Insider

Christy on the the topic of Meta confirming that chats with Meta AI will influence content and ad targeting from 16 Dec (outside UK/EU/South Korea).

Shrinking budgets and tighter turnarounds are pushing more marketers to use AI-generated visuals in campaigns. Many are doing it to cut costs and accelerate testing, while avoiding the PR splash that sometimes comes with overt “made with AI” creative. Creatives report cautious adoption: internal and low-risk use cases first, consumer-facing later.

Cost vs. craft
AI art can slash production time and spend, especially for variant testing and concepting. But buyers still worry about uncanny-valley moments, brand safety, and dilution of distinctive visual identity. Rights checks (models, textures, style look-alikes) add legal overhead.

Where it’s landing
The safer beachheads: background swaps, product set-ups, iterative resizes, and filling gaps in omnichannel toolkits. Teams keep human direction and QA up front to maintain taste and cohesion, with AI doing the heavy lifting underneath.

Signals to watch
Expect clearer asset provenance (watermarks, C2PA), and brand standards that include prompt libraries and negative-prompt guardrails. As models improve, the trade-off between “cheap and fast” versus “ownable and premium” will narrow, but not disappear.

Team shape-up
Winning orgs pair creative ops + data: prompt specialists, rights/legal partners, and measurement loops that connect image variations to lift in CTR, CVR and ROAS—not just “looks cool.”

Our takeaway: Use AI art as a multiplier, not a replacement. Start where speed matters and risk is low, prove uplift, then scale with provenance and brand-consistency guardrails. 📖 Read the full article here

Fast walkthrough of the week’s Google AI Mode changes (more visual for shopping queries), reporting tweaks, and agentic behaviours to watch. This recap from Barry Schwartz is useful for PPC/SEO teams syncing messaging and measurement - Watch it here.

Two days focused on customer intelligence and AI-powered decisioning for marketers, including CI360 roadmap insights, live demos, and peer case studies. If you care about journey orchestration, optimisation, and how automation is reshaping engagement, this is squarely in the AI-marketing wheelhouse.

Three new roles we think are perfect if you’re ready to upgrade your title and your tech stack:

Was this email forwarded to you? Sign up here.