Pressure Is Not a Plan
By Jake the WizardEighty percent of marketing leaders feel pressure to adopt AI within a year. Fifteen percent have a plan with a named owner and real metrics. That gap is not a soft readiness problem. It is a business about to get hurt.
Thirteen years in marketing and two exits taught me the hardest system to debug is people. Not the tool. The people who never decided what the tool was for. That failure has a new costume this year, and the numbers on it just landed.
Why doesn’t feeling behind count as a plan?
Because a mood is not a decision. Pressure moved everyone. Nobody aimed the movement anywhere.
Supermetrics surveyed 435 marketing leaders, director level and above, across the US, UK, Germany, Australia, and Singapore, and published the results as its AI Readiness Gap report on July 14, 2026. Eighty percent feel moderate or significant pressure to adopt AI within twelve months. Fifteen percent report a clear strategy with defined owners and success metrics.
Read those two numbers next to each other. Almost everyone feels the heat. One in seven turned the heat into a decision. The rest are carrying a feeling and calling it a roadmap.
Feeling behind is a mood you can hold for years without changing anything. A plan is a decision you can be held to. Only one of them shows up in the work.
What does “nobody owns it” actually cost?
It costs you the fix. Eighty-five percent of the organizations in that survey had no formal AI strategy or clear ownership. What nobody owns, nobody improves.
This reads as a paperwork problem and it is not. MIT’s NANDA group ran 150 leadership interviews, surveyed 350 employees, and analyzed 300 public AI deployments for The GenAI Divide, and found 95% of enterprise generative AI pilots produced no measurable return. The models worked fine. What broke was ownership, the definition of success, and workflows nobody redesigned before bolting a tool onto them.
Nothing announces that kind of failure. The work drifts, everyone assumes a colleague has it, and four quarters disappear while the company tells itself it is exploring.
Does using ChatGPT count as adoption?
No. That is a person with a chatbot, not a company with a strategy.
The same survey found 52% of SMB teams using AI for content creation and 37% for reporting. Real usage, counted fairly. Then it asked the harder question: only 6% described AI as fully embedded in their workflows.
So the usage everyone points to as proof of transformation is task level. Captions. Summaries. A first draft of the report. Useful work, and none of it is a decision about what the tool is for. The report counted all of it and the hole was still there.
A habit is not a system. You can run the habit for two years and end up exactly where you started, with better captions and no answer to what changed.
Why AI with no plan is faster chaos
Point a fast tool at a broken process and you get more broken output, sooner. That is the whole mechanism.
AI does not diagnose. It executes. Hand it a muddled offer and it will produce forty confident variants of a muddled offer by Thursday. Wasted spend, off brand work, a mess with a clean font on it. The tool did exactly what you asked. You just never decided what to ask for.
This is the sequence problem wearing new clothes. A pen topples a skyscraper in twenty nine steps because the order is doing the work, not the volume. AI with no plan is tactics with no sequence. Same failure, faster clock. The Eight Dominoes run the same way: the first decision sets what every decision after it is allowed to be, and running the later ones harder never repairs a missing first one.
A plan alone does not finish the job either. The machine still cannot supply the idea sitting at the center of it, which is why you have to be the originator and not the operator. Decide the job, then bring the idea. Skip either one and the tool fills the gap with the average.
The plan fits in one sentence
One owner. One sentence about what AI is for here. One number that tells you if it worked.
That is the whole minimum. Not a document, not a hire, not a transformation office. The 85% did not skip a budget line, they skipped an afternoon.
| The 15% have | The 85% have |
|---|---|
| A named person accountable for it | A general sense that someone is on it |
| A sentence defining the job | A list of tools people liked |
| A number that can come back negative | Usage stats that only go up |
| Something to correct next quarter | Something to restart next quarter |
Look at the last row. That is the compounding difference. A wrong plan you can steer, because you wrote down what you expected and reality can argue with it. No plan gives you nothing to correct, so you begin again every time the tools shift. The 15% get to edit. The 85% get to start over, forever.
On the number: pick one that can embarrass you. I have used the same core metric for years, which is revenue, because if the needle does not move, was there really value? Usage dashboards never embarrass anyone. That is why they are popular.
The falsifiable version
The claim, stated so it can be wrong. Take two similar businesses eighteen months out. One named an owner and wrote a sentence. One felt pressure and bought tools. The first will not have better AI, it will have a shorter list of abandoned experiments and at least one process that measurably improved. If the improvisers come out ahead on anything but tool count, I am wrong and the plan is overhead.
One caveat on the source, because it matters. Supermetrics sells data integration, so a report concluding that data plumbing is the blocker serves Supermetrics. Discount that particular finding. The pressure and ownership gap it measured stands on its own, and MIT’s failure data arrives at the same place from a different direction.
The pressure is real and it is not going away. The only question is whether it made you decide something or just made you nervous. The bill on the second one is already running. Most people have not opened it yet.
No spell here. Name the owner. Write the sentence. Pick the number that can tell you no.
Straight Answers
How many companies actually have an AI strategy?
About 15%. Supermetrics surveyed 435 marketing leaders across five countries for its July 2026 AI Readiness Gap report and found only 15% had a clear strategy with defined owners and success metrics. The other 85% either had no formal strategy, partial clarity on use cases, or a mention of AI in planning with nothing prioritized.
Does using ChatGPT for content count as AI adoption?
No. That is usage, not adoption. In the same survey 52% of SMB teams used AI for content creation, but only 6% described AI as fully embedded in their workflows. One person with a chatbot is a habit. A company with an owner, a defined job, and a number attached is a system.
Is it smarter to wait on AI until the tools settle down?
Only if waiting is a decision with an owner and a trigger. A plan is not a bet on a specific tool, it is a decision about the job you want done. Tools change every quarter. The job does not. Name the outcome, name the owner, swap the tools underneath.
Who should own AI on a marketing team?
One named person, and it can be a person who already has a job. Ownership means someone is accountable for what AI is for here and what happens when it does not work. What nobody owns, nobody improves. Ownership is the thing 85% of surveyed teams are missing, not budget.
What does a minimum AI plan actually look like?
One owner, one sentence about what AI is for in your business, and one number that tells you if it worked. That is an afternoon, not a budget line or a consulting engagement. The companies with no plan skipped the afternoon, not the money.