The question everyone asks first
Who should I hire to lead AI transformation for my marketing team?
Whoever is accountable for the output of the redesigned work — which is already the person who owns pipeline. Transformation changes how marketing and sales operate day to day, so the decisions are commercial: which workflow to rebuild, what it must beat, and when to stop. That is a CMO or CRO seat with AI underneath.
Notice the question contains an assumption worth testing: that this starts with a hire. Usually it does not. Most departments already have the person who should own it — what they lack is a model for what changes and in what order.
The failure mode is predictable. Companies hand AI to the most technical person available, because AI looks technical. The pilots work and the pipeline does not move, because nobody changed how the department operates.
Key facts
- The short answer
- Whoever is accountable for pipeline. Ownership follows the work, not the technology.
- Why not a technical owner
- The decisions are commercial — which workflow, what baseline, when to stop.
- What changes
- The work itself, who does it, and how it is governed. All three or none.
- Where it starts
- A diagnosis: is the constraint capability, capacity, or focus?
- When you hire
- When the system runs daily across more than one function and the constraint becomes capacity.
Transformation isn’t a hire.
It’s a redesign.
A new seat on the org chart changes nothing by itself.
You can appoint someone on Monday and have the identical department on Friday. What moves is the work: steps removed, handoffs redrawn, output governed.
Diagnose before you change anything.
Four to six weeks of looking, before a single tool decision.
Every stalled department is stalled for one of three reasons: capability, capacity, or focus. The team cannot do the work, cannot get to the work, or is pointed at the wrong work. AI fixes the second and third far more often than the first — and buying tools before you know which one you have is how programs waste a quarter.
i.
Capability
The team does not know how. Training and hiring problem — and the one AI is least likely to solve on its own.
ii.
Capacity
They know how, and there are not enough hours. The clearest case for rebuilding a workflow around AI.
iii.
Focus
Effort is going somewhere that does not pay. No tool fixes this. A decision does.
iv.
Or none of the above
Sometimes the market moved and the team is fine. Read shallow and you will “fix” a team that was never broken.
The deeper you go into the problem, the more obvious the turn becomes.
Leads arriving fast, and nobody calling them.
What a real transformation looked like.
At hh2, a construction software company, inbound leads sat for days before anyone called. The obvious read was a performance problem — reps not working hard enough. The deeper read found two different things: a trust problem (the team had been fed low-quality leads for months and stopped believing them) and a notification problem (a rep on a call has no idea a form was filled).
Before — the department as it ran
- Lead lands in the CRM
- Assigned to an account executive
- Waits for someone to notice
- No signal, no clock, no owner
Leads aged while everyone was busy.Days to first contact
After — the work redesigned
- Routed to a dedicated first responder
- Alert fires where the team already works
- Waiting GONE
- Unclaimed reassigns automatically
Comp tied to the new behavior, so it stuck.Minutes to first contact
Look at what actually changed: routing, notification, and the comp plan. The work was redesigned, the handoffs were redrawn, and the incentive matched the new behavior. AI made the last leg fast — but AI dropped into the old routing would have changed nothing, because the constraint was never speed of typing.
Same thinking, a different motion: rebuilding hh2’s go-to-market from search optimization to answer-engine optimization — being cited by AI assistants rather than ranked by Google.
What happens to the people currently doing the work?
In a working transformation, most of them stay and their jobs get larger. When a step disappears, the person who owned it moves up the stack — from producing the work to directing and checking it. The roles that struggle are the ones defined entirely by executing a step that no longer exists.
This is the part leaders avoid saying out loud, and the silence is what makes teams resist. Say it plainly and early: the work changes, the standard goes up, and the people who learn to direct the system are worth more afterward than before.
It also sets the real hiring test. You are not hiring someone to have the AI skills — you are hiring or promoting someone who can redesign work and hold a standard. See the architecture they end up building.
The first 90 days
The order that works.
- 1Diagnose before you change anything — Four to six weeks of looking. Is the constraint capability, capacity, or focus? Do not buy a tool until you can answer that.
- 2Find where revenue actually leaks — Name the one workflow costing the most pipeline today, not the one easiest to automate. They are rarely the same.
- 3Redesign that workflow end to end — Remove steps, redraw the handoffs, and run it against a measured baseline. Match the incentives to the new behavior or it will not stick.
- 4Govern it — Decide who owns the output, what gets checked, and what stops it. Ungoverned AI gets switched off after the first bad week.
- 5Connect the second workflow — Wire the first system's output into the next. This is where compounding starts, and where most programs stall.
Notice what is not on that list: buying a platform. Tool selection is a consequence of the first two steps, never a substitute for them.
FAQ
Common questions
Each answer stands alone — link straight to the one you need.
Should our CTO own AI in the revenue org?
Your CTO owns models, data, and infrastructure — and should. Ownership of the revenue side belongs with whoever is accountable for pipeline, because the decisions are commercial: which workflow to rebuild, what it must beat, and when to stop.
Do we need to hire anyone to start?
Usually not to start. Most departments already have the person who should own it. What they lack is a model for what changes and in what order — and four to six weeks of honest diagnosis before anything gets bought.
What does an outside operator cost?
A fractional engagement runs a fraction of a loaded CMO salary, and it has an end date. See what a fractional CMO costs.
How is this different from hiring a consultant?
A consultant recommends the redesign. An operator makes it and stays accountable for the number it produces. Ask which one the engagement actually buys you.
Do we need clean data before we start?
Often some — AI acts on data it can trust. That is usually the first workflow rather than a reason to wait. See revenue systems.
Where does this fit in the wider framework?
This is the thinking. The maturity model is the Pilot-to-P&L ladder, and the architecture it produces is the Hamer AI OS.