Your AI pilot is a license, not a strategy
Most companies we meet have three AI subscriptions, two pilots and no number. That is not a strategy. It is a cost line waiting to be noticed.
Ask a leadership team for its AI strategy and you usually get a list of tools. A writing assistant license. A pilot with a vendor someone met at a conference. One team doing something interesting that nobody outside it can explain. Everyone agrees it matters. Almost nobody can say what it has returned.
That is where most of the market sits. Gartner's 2026 CMO Spend Survey of 401 marketing leaders found CMOs now put an average of 15.3% of marketing budgets into AI. Only 30% report mature AI readiness, and 70% say their internal processes are not mature enough to implement and scale it. The money is moving faster than the operating model.
A stack of subscriptions is not a strategy. A pilot with no metric is not a pilot. It is a license.
How it happens
We have watched this sequence from the vendor side and the buyer side. A demo lands well in a leadership meeting. Someone senior wants to be seen moving, reasonably. Procurement signs. A small team adopts it with enthusiasm, but nobody measured what the process cost before the tool arrived. Six months on, the tool is still billing, the team likes it, and finance can't say whether it is worth the money. There was never a baseline.
Repeat across three or four functions, each tool with its own champion and renewal date, and you get real spend, real enthusiasm and no number anyone would defend to a board.
The question that fixes it
Which workflow, at what volume, costing what today, will cost what after the tool is in?
If nobody can answer, the tool isn't deployed. It is installed.
Start with the process and the number. Pick the tool second. Doing it the other way round is how a company ends up paying for three overlapping products that all claim the credit.
The one-page pilot card
Before anything is bought or renewed, we ask for this, filled in.
| Field | What good looks like |
|---|---|
| Workflow | One named process in one function, such as invoice coding or campaign QA |
| Volume | Units per month, pulled from a system, not estimated |
| Baseline | Cost, cycle time and error rate measured before the tool |
| Target | The number that makes it worth scaling |
| Owner | One person, accountable for the result |
| Decision date | Sixty to ninety days out: scale, fix or stop |
| Adoption | Share of eligible users using it every week |
If a line can't be filled in, the pilot isn't ready.
What working looks like
Narrower than most companies expect. One high-volume, judgment-heavy workflow: enough repetition that a modest gain is worth real money, enough judgment that ordinary software hasn't already automated it. A tool chosen for that job, not for its logo or its funding round. A short runway to production with adoption included, not a technical demo nobody opens.
One honestly labeled experiment with a stop date beats five quiet subscriptions. The signal to watch is not how many tools you run. It is how many have an owner and a number.
Then the next workflow.
It won't make an exciting slide. It works. And the first thing a properly measured pilot tends to find is licenses already being paid for that nobody uses, which often funds the next one.
Sources: Gartner, 2026 CMO Spend Survey (May 2026).
What is the difference between an AI pilot and an AI strategy?
A pilot without a baseline and a target is just a subscription. A real pilot names one workflow, measures what it costs today, sets the number it should reach and has an owner and a decision date.
How much are marketing teams spending on AI?
Gartner's 2026 CMO Spend Survey found CMOs allocate an average of 15.3% of marketing budgets to AI, while 70% say their processes are not mature enough to scale it.
How should a company structure its first AI pilot?
Pick one high-volume, judgment-heavy workflow, measure cost, time and error rate first, name one owner, and give it sixty to ninety days to reach production with real adoption before deciding to scale or stop.
We would rather look at your numbers than talk about ours. If you have AI subscriptions and no number, an opportunity assessment gives you both in three weeks.