The Shadow CMO Exercise

In AI-Native GTM: All the Way or Not at All, I describe the thought experiment that started our rewrite: if I had to fulfill the CMO role from scratch, with zero headcount, zero infrastructure and no legacy commitments, how would I do it AI-native? I called the answer the “Shadow CMO organization.”

This page turns that question into an exercise you can run with an AI assistant. The point isn’t the output. It’s seeing how much of your current operating model exists for reasons that no longer hold.

Before you start

• Write down what your marketing function has to deliver for the business, in outcomes rather than activities.
• List the teams, roles, recurring meetings, reports and tools you run today.
• Note the constraints that are real (regulation, security, budget) and keep them separate from the ones that are habit.

The prompt

Paste the following into your assistant, then add your notes from above.

“I lead a marketing function. Imagine I had to fulfill this role from scratch, with zero headcount, zero infrastructure and no legacy commitments. Start from a blank page. Architect everything around what AI can actually do today, and add humans only where they are genuinely needed and uniquely valuable. Here is what the function must deliver, what we run today and our real constraints: [paste your notes]. Design the operating model: the outcomes it is organized around, the workflows and agents that do the work, and the few human roles that remain. Then list every part of our current setup — reporting layers, cadences, tools, handoffs — that would not survive this design, and say why.”

Reading the answer

• Look first at what doesn’t survive. Reporting layers, cadences, slide-deck shuffling and MarTech middleware are the usual suspects.
• Check whether the design is organized around outcomes or still around crafts like content, digital, product marketing, demand gen, brand and ops.
• Ask what the governed environment would need to be (the “landing zone”) for these workflows and agents to run compliantly, securely and auditably.

Where to take it next

A thought experiment is the spark, not the system. The essay covers the next step: a small team of three — a visionary executive, a senior strategist and an AI architect who can actually build — whose first deliverable is that landing zone. Then ship something real and show it to the people who need convincing.