Freedman

An AI agent can now make 900 ads a week and localise them. Who in the market is checking them? Runway, the AI video company, launched Runway Ads this week. Their announcement is here: https://lnkd.in/grDc5ug3 The agent makes variations of an ad, swaps the on-screen text, screenshots, and voiceover for each market, publishes them, and reads the results back. Runway says its own team went from 77 ads a week to about 900 using it. To be fair, the controls are better than most ad hoc local work. You can set brand voice rules and terms that must never be translated. Human approval is on by default. But two things bother me. Swapping text and voiceover is translation. It is not adaptation. Nobody has asked whether the idea works in that market. And approval can be switched off, campaign by campaign. At 900 ads a week, approval will be switched off. No team reviews that volume properly in 12 languages. I have no evidence the output is bad. The numbers are Runway's own, about its own marketing. There are no customer results yet. My worry is simpler. International marketing debt builds when work goes out unchecked. This makes unchecked work very cheap. If you are piloting tools like this, who signs off in each market?

AI won't fix a broken global marketing operation. It will probably expose it faster. Gartners latest AI Hype Cycle argues that the central enterprise question is moving from what AI can do, to whether organisations can account for what AI systems do, what they cost and how they are governed. It specifically highlights the rapid adoption of AI agents and agentic commerce. For global marketing teams, AI creates another layer between the central brand team and local execution. The issue is no longer simply efficiency. It is control, accountability and consistency at scale. This is where I'd strongly agree with Gartner, but I'd add something that is usually missing from the AI discussion: marketing debt. Most organisations already have fragmented brand guidelines, approval processes, DAMs, agency relationships, localisation workflows and market exceptions. Putting agents on top of that infrastructure does not remove the debt. It can make it run faster. The strategic question for global marketers is therefore not where can we use agents? It is what parts of our global marketing system are sufficiently well defined for an agent to operate safely? https://lnkd.in/gBv8yNC8

International marketing debt doesn't come from making mistakes, it happens through sensible decisions. A brief adapted from the home market because there's no time to rebuild it. A cultural review skipped because the schedule's tight and it reads fine on paper. Defensible calls, made by capable people under pressure. That's what makes it hard to stop. A mistake gets caught and fixed but this gets approved on purpose, by someone doing their job well, and then it compounds. Read below for my second article in my series on International Marketing Debt. Looking forward to discussing this at Sifted Summit. #SiftedSummit2026

Today is international translation day, perfect for our panel discussion topic starting at 1pm on the main stage - how to set key milestones when entering new markets. Looking forward to the discussion, see you there!

Day 1 at Sifted Summit - we've arrived! I'm already feeling the energy - it was amazing to walk into the event and see the freedman name everywhere. First off, i'm focused on the panel discussion I'm joining on the main stage at 1pm: How to set key milestones when expanding into new markets. Hope to see you there.

Publicis London's World Class report is being used to argue that global campaigns don't work. That's wrong and the report's own footnotes prove it. The gap behind the claim isn't statistically significant. The study doesn't account for budget, category or brand size. And it counts a campaign that ran in two countries the same as one that ran in forty. Read the data properly and it says the opposite. Bold, distinctive global work pays. Safe global work doesn't.