Before and now The CMO role: Before, the job was choosing technology vendors, choosing agencies, measuring campaign and pipeline results, optimizing for performance, and delegating the technical layer without asking too many questions about how it worked underneath. Now, the job includes understanding what obligations each AI system you bring in actually carries. Demanding transparency from vendors before you sign, not after. Documenting why you use each AI system for the decisions that matter. Knowing when a customer needs to be told they’re talking to an AI. Sharing this responsibility with your CTO and your AI lead, rather than handing…
A Tuesday with AI in your marketing department 8:30. A generative tool drafts the first version of a campaign brief. 9:15. Meta automatically decides where to push more budget. 10:05. A chatbot answers a prospect’s question. 11:20. Your CRM decides which lead your sales team should call first. 12:40. An AI-generated avatar presents a sales proposal to a client. In under four hours, you’ve used five different AI systems. You’ve delegated five decisions. And you’re still accountable for every one of them. What those five AI decisions have in common None of them was made by you directly. Each was…
Why guaranteed results didn’t exist A guaranteed outcome requires controlling the variables that determine it. In advertising, those variables were always distributed: platforms controlled their own inventory, agencies controlled their own process, and the space between them belonged to nobody. So the risk stayed with the client. Not because the industry was dishonest — because the architecture to back a real guarantee didn’t exist. Projections, targets, benchmarks, best-efforts commitments: these weren’t evasions. They were the honest upper bound of what any system could promise when it couldn’t control what produced the result. What controlling the variables actually requires Speed: decisions…
The difference between data and pattern recognition All campaigns generate data. Essentially, what distinguishes a system that has run 12,000 campaigns from one running its first isn’t access to more data from a single campaign. In other words, it’s the ability to recognize what the data means because it has been seen before, in comparable contexts, with known outcomes. For instance, a signal that looks ambiguous in isolation becomes interpretable when you’ve seen it in 400 comparable campaigns and know how it resolved in each one. A market condition that looks unprecedented to a system with no history is recognizable…
What decision latency actually is Decision latency is the interval between the moment a signal appears in your campaign data and the moment your system acts on it. In a human-managed process, that interval includes: someone noticing the signal, flagging it, the team reviewing it at the next available checkpoint, agreeing on a response, and implementing the change. At its fastest — a responsive team, a clear signal, no approval friction — that process takes hours. In normal operating conditions, it takes days. In complex organizations with multiple approval layers, it can take a week. The market doesn’t pause while…
At 3am, the report and your CPM disagreed Your CPM spiked 28% on two exchanges between midnight and 6am. The platform reported average CPM for the day. Average looked fine. It wasn’t fine. A significant portion of your overnight budget ran at a price that a real-time system would have redirected. The average absorbed the problem and made it invisible. This isn’t a reporting failure. It’s a structural issue. Daily averages are designed to summarise, not to signal. When the signal matters most, in the hours when no one is watching, the summary is all that remains. You paid for…
How agencies design their KPIs No one sat down and decided to design a KPIs framework that would obscure performance. That’s not how these things happen. Agency KPIs emerged the way most professional standards do: organically, over time, shaped by what was possible to measure, what was possible to influence, and what was possible to demonstrate to a client in a quarterly review. The result is a framework that centers variables where human expertise shows and attributes clearly. Reach. Frequency. CPM efficiency. Campaign setup quality. Reporting turnaround. Strategic recommendations delivered. These are real inputs. They reflect genuine work. The problem…