The Gap Is the Story
AI has moved into marketing faster than the operations underneath it have matured. That gap is where the returns leak out, and it's what this library is about.
AI is taking over the production layer of marketing: drafts, variants, tagging, reporting, QA, the routine optimization that used to fill most of the org chart. The operation underneath it is moving much slower.
When production stops being the bottleneck, what's scarce changes. Somebody still has to decide what should exist, what good looks like, and when the machine is wrong, and none of that speeds up because the making did. That work also sits on a foundation that was easy to ignore when humans did the producing: the data and the architecture that connects it, the permission to use it, measurement that can tell activity from effect, governance that can say no, and an operating model that puts people where the judgment calls are.
How far a team has pushed AI into the work tells you one thing. Whether the foundation can carry it tells you another. I've spent twenty-five years on both sides of that problem, client side and agency side, across more brands than I can list, and the two have never moved at the same speed. The distance between them is where the money goes. AI scales whatever operation it lands on, including a weak one. The industry research keeps finding the same split: a small group whose early investment in the foundation is paying off, and a much larger one whose pilots never make it to production.
Even when the gap is plain to see, closing it usually comes down to nerve. Most organizations can afford the capability now. Far fewer will move against their own economics while those economics still work.
The library splits in two. Signal is instruments: frameworks, reference architectures, and tools you can put your own organization against, starting with a maturity framework for marketing capability and a working console that shows how customer signals clear the permission gate, or don't. Fieldnotes is arguments: where the new marketing jobs come from as the production layer thins, why most organizations are data rich and activation poor, and why the hard part is usually nerve.
None of it is finished, on purpose. The field is moving too fast for a fixed position to stay honest, so I publish what I think and revise it as the evidence moves. More will keep showing up here. If you run marketing, data, or AI and want to test your own read against mine, start with the framework, or measure the gap in your own operation. If you'd rather argue with something, start with Fieldnotes.