The AI Skills Gap and the Bitter Truth About Skills

Marketing leaders are investing in artificial intelligence with increasing confidence, but many organizations still lack the right skills, technological infrastructure, and consumer trust for these investments to start delivering real returns.

Most CMOs Can’t Distinguish Generative AI from Predictive Models

The problem starts with the term “AI” itself, which is often used too broadly as a label for everything from generative language models to predictive analytics systems and machine learning. This, in turn, leads to oversimplifications such as “AI did this,” even though, in practice, the work involves specific tools, data, and the people who implement them.

“AI” is becoming an umbrella term for everything from text generation to medical data modeling. This may be convenient for the media, but it leads to oversimplification, obscuring the difference between an automation tool and a predictive system that supports scientific discovery.

AI Skills Among CMOs: The Bitter Truth

According to research by Findem and CMO Huddles, 26% of newly hired CMOs have some AI-related skills, compared with just 9% of those who have held the role for more than four years. Companies are clearly seeking leaders who understand new technologies, yet most current marketing chiefs still have ground to make up.

In many organizations, “AI skills” unfortunately begin and end with the ability to write a prompt and generate an image or text. That is useful, but it is not yet genuine AI proficiency. It is more like operating an interface than understanding the system: the data, models, limitations, risk of errors, privacy, quality of results, and when generative AI is needed versus predictive AI.

The technology itself is also not ready for widespread implementation. In a global study by Lippincott and Bloomberg Media, only 11% of marketing leaders rated their organization as “excellent” at implementing and innovating in marketing technology. Just 7% rated it as “poor,” suggesting that most companies fall somewhere in the middle: they have the tools, but not yet the full ability to put them to work.

Ambitions for AI Investment

Investment ambitions remain high, however. According to Gartner’s 2026 CMO Spend Survey, as many as 75% of senior marketing leaders plan to increase spending on AI assistants next year.

But here, too, categories are being confused: a generative assistant, a recommendation system, a predictive model, and an analytics tool are not the same thing. “Campaign personalization” usually does not come from a prompt; it depends on working with data, integrations, and predictive models. These are precisely the kinds of solutions most often behind the effective tailoring of communications to audiences.

Consumer trust is another factor. Kantar’s Media Reactions 2026 finds that only 23% of consumers trust recommendations from AI assistants. This matters because even the best automation will not automatically translate into brand credibility or business results. Recommendations from AI assistants are based on publicly available data; predictive models are not behind them. AI can support data analysis and improve processes, but simply invoking the term “AI” guarantees neither sound decisions nor a competitive advantage.

Building AI Skills Step by Step

That is why an effective strategy is not about implementing AI everywhere at once. Skills are better built step by step: by training teams, clearly distinguishing generative AI from predictive and analytical AI, starting with low-risk applications, and measuring business outcomes rather than just time saved. In customer-facing areas, a human still needs to have the final say–not because AI is useless, but because brand trust cannot be generated automatically.

AI is indeed changing marketing, but only when companies understand what kind of AI they are using. CMOs who do not treat everything as one convenient shorthand will be better prepared to turn technology into real results.