Diagnose skills versus hours before spending on AI or new hires

Key points

Gartner's 2026 CMO Spend Survey of 401 marketing leaders shows labor's share of marketing budgets rising from 21.9% in 2025 to 24.5% in 2026.
In the same survey, 38% named a lack of internal AI expertise and talent as the top barrier to AI-driven efficiency.
An Optimizely-commissioned Savanta survey of 2,003 marketing leaders found 76% spend at least three hours weekly correcting AI output.
MarTech suggests three checks: review what gets corrected, look for single-person bottlenecks, and pause one recurring project for a month.

Why it matters for marketers: Buying AI or adding staff without diagnosing skills versus hours can raise output while loading review work onto your best people. A quick diagnosis comes first.

A MarTech analysis argues that many marketing teams pour AI licenses or new hires into problems they have not diagnosed. A skills gap means the team cannot do the work the tools now demand. A headcount gap means the team knows the work but lacks the hours. The fixes differ, and AI behaves differently in each case.

The data cited points to the strain. In Gartner's 2026 CMO Spend Survey of 401 marketing leaders, labor's share of the budget rose from 21.9% to 24.5%, and 38% said missing internal AI expertise was the top barrier to efficiency. A Savanta survey of 2,003 marketing leaders, commissioned by Optimizely, found 76% spend at least three hours a week editing or fact-checking AI output. Another 54% said leadership underestimates that effort.

For marketers, the risk is that AI on top of a skills gap yields more output at the same quality, plus a new review job. Reviewing also takes skill: someone who does not understand segmentation cannot catch a bad segment. In the same study, 46% worried heavy AI use will erode junior marketers' skill development. On a skilled team short on hours, the article says, AI can take volume off capable people and the saved time stays saved.

The piece offers three checks. First, ask reviewers what they correct: judgment errors such as the wrong audience point to skills, while typos and missed steps point to hours. Second, look for work queued behind one person who knows how to build a journey or audience, which signals a concentrated skills gap and a retention risk. Third, pull one recurring project for a month. If other work improves, hours were the issue. If not, skills were. The test costs a month of that project, so choose one you can pause.