This session is a field report from inside a role that only exists because the shape of marketing work changed. I spent over a decade in marketing, most recently as Director of Paid Media at NP Digital, managing large teams of media buyers. I now manage a fleet of AI agents instead. This is an honest account of what that actually looks like, what it means for the rest of the industry, and what to do about it if you're earlier in your career than I am. We'll start with what's genuinely changing. Google and Meta are steadily removing advertiser control and pushing automation into places that used to require a person. Campaign building, QA, pacing checks, reporting — the execution layer is becoming close to free. That isn't a forecast, it's visible in their own product announcements, and we'll look at them directly. Then the useful question: if execution is commoditised, what are marketers actually for? My answer is judgment, taste, strategy and relationships — but I want to make a sharper version of that argument. Machines aren't accountable. The humans behind them still are. When a campaign misfires, when a claim is wrong, when a budget runs somewhere it shouldn't, no agent answers for it. You do. That accountability is not a consolation prize; it's the part of the job that can't be automated, and it's worth understanding deliberately rather than by accident. I'll then walk through my actual week. Not a tools demo — three specific agents and what each one taught me. A morning briefing that reads my calls, calendar and inbox and tells me what matters. A commitment tracker that listens to my week and writes my task list overnight, so I wake up to work I didn't plan. Media agents that watch budget pacing and creative fatigue and raise a flag when something moves. I'll be precise about what is monitored automatically versus what still requires a human to decide, because the difference between those two things is the entire argument. I'll also cover what broke. Some of this didn't work, some of it created new problems, and the governance questions are real: what these systems can reach, what they should never touch, and how mistakes surface before a client sees them. Finally, the part most relevant if you're early career, between roles, or watching your agency's delivery model shift: the gap between using AI and building with it. Level one is asking a chatbot questions. The real move is building end-to-end workflows and connecting systems together. That gap is steeper than most talks admit, and I'll be honest about how long it took me and where it was hardest. You'll leave with a clear view of which parts of your role are exposed, which parts just became more valuable, and one specific thing you can build this month. This is not a talk about humans being replaced. It's a talk about what humans are for when the typing stops being the job.