Marketers were among the earliest professionals to truly lean into generative AI. We fired up ChatGPT, entered a prompt, and got a response that felt almost magical. We were some of the first to experience those AI “wow” moments and began threading LLMs into our daily routines. By most accounts, marketing sat at the front of the early adoption curve.
But at some point, our progress stalled. A year and a half later, a surprising number of marketing teams are still working almost exactly as they did on day one. They open a chat interface, type a prompt, tweak the response, and move on. The surrounding workflow hasn’t meaningfully evolved. Only one part of the old process has really changed: we swapped the blank page for a rough draft. Nearly everything else remains largely untouched.
How we got stuck
There are understandable reasons for this. Inertia is a big one. It’s simply more comfortable to keep following familiar patterns (see my article about paving cow paths). Early AI outputs also damaged trust. The first time you asked an AI to create something truly important and it hallucinated facts, mentioned a competitor by name, or churned out something painfully generic, you took note. You decided AI needed a short leash. You relegated it to low-risk drafting and kept meaningful judgment squarely with humans. At the time, that was a sensible response.
The issue is that this early lesson hardened into a long-term habit. No one clearly owned AI adoption. In most marketing organizations I’ve talked to…