A 12 June analysis from Digital Applied, drawing on recent CMO survey data, reports that chief marketing officers now allocate 15.3% of their marketing budgets to artificial intelligence. The same analysis finds only about 30% of marketing teams are ready to operate what they have bought.
That gap is the story. Procurement has sprinted. Capability has walked. And EMEA boards approving the next AI allocation are funding a problem they have not yet defined, let alone solved.
The Spend Is Real and the Readiness Is Not
15.3% of a marketing budget is no longer a pilot line. It is a meaningful share of working spend, sitting alongside media, content production, and martech licences that already strain most enterprise plans. When that share goes to AI, the implicit promise to the board is productivity, personalisation, or both.
Yet the readiness number tells a different story. If only three in ten marketing teams can run AI well, the other seven are paying for tools they cannot operate at the standard the business case assumed. The licences renew. The dashboards exist. The outcomes drift.
This is not a failure of ambition. CMOs are responding rationally to pressure from boards, peers, and vendors all telling them the same thing about competitive risk. The failure is sequencing. The spend arrives before the foundation that makes the spend work.
What Readiness Actually Means
A separate Contentstack survey, summarised in the same week's marketing technology roundup, captured the point from the practitioner side. Eighty-eight per cent of enterprise leaders said they wished they had built their content and data foundations before deploying AI tools.
That is a striking admission. It is not a vendor pitching a platform. It is the buyer, after the fact, naming what they should have done first.
Readiness in this context is unglamorous. It means a single, governed view of the customer that AI can actually read. It means content tagged, structured, and rights-cleared so that a generative system does not hallucinate brand voice or republish something it should not. It means permissions, consent, and audit trails that survive a regulator's question. And it means a marketing operations team that knows which model is making which decision and why.
None of that ships in an AI procurement contract. All of it has to exist before the contract delivers value.
Why EMEA Buyers Get This Wrong
The pattern is consistent across enterprises in the region. AI spend is approved at the CMO and CFO level on a productivity argument. The data and content work, which is slower, less visible, and harder to attribute to a single quarter, is deferred to a later programme that competes against new tool launches for the same budget cycle.
Vendors understand this incentive perfectly. The demo shows the model. The contract sells the model. The customer discovers, six months in, that the model is only as useful as the data and content sitting underneath it. By then the renewal is locked and the team is too small to operate what was bought.
Three habits make the gap worse. First, treating AI as a marketing decision rather than a portfolio decision that also involves data, security, and customer operations. Second, buying tools faster than hiring or reskilling the people who run them. Third, measuring AI spend by adoption rather than by outcomes the board recognises as growth.
What to Do Before the Next Allocation
The correction is not to slow down. It is to invest in the right order. Data and content foundations come first. Operating capability, meaning the people, processes, and governance to run AI in production, comes second. Tools come third, against use cases the foundations and the team can actually support.
For EMEA enterprises, this also means stopping the habit of treating AI procurement as a marketing-only conversation. The customer data platform, the content infrastructure, the consent layer, and the contact centre that absorbs the consequences of automated decisions all belong in the same programme. One accountable owner. One readiness standard. One score the board can read.
The 15.3% figure is more likely to rise than fall. The readiness figure will move more slowly, because building data, content, and operating capability is harder than signing a licence.
EMEA boards should demand proof of readiness maturity before approving the next AI allocation. Not a vendor roadmap. Not a pilot result. Evidence that the data, the content, the governance, and the team can carry the spend. Without that proof, the budget line keeps growing and the return keeps slipping. With it, the 15.3% starts to behave like an investment instead of an expense.