Method, , 8 min read
The most common sequence in AI initiatives in marketing is the wrong one: first the tool, then the use case, then the question about data. A readiness check reverses that. It costs a few days, prevents misinvestment and delivers something worth more than any licence along the way, namely a prioritisation that holds internally. This article describes the four dimensions and the questions behind them.
A tool solves a defined problem. If the problem is not defined, it solves none and still creates cost, training effort and expectations. For 2026, Gartner names immature processes, insufficient data foundations and missing governance as the main reasons why only 30 percent of marketing organisations can scale AI. All three are questions that can be answered before a tool is selected.
The second reason is organisational. A readiness check creates a shared picture. Without that picture, marketing, IT and management discuss different problems and wonder why they do not converge.
The first and, in practice, decisive dimension. Salesforce names data problems as the main blocker for effective AI use, with only about half of the surveyed organisations having full access to their own data from service, sales and commerce.
For brands targeting families there is an additional particularity. Behavioural data about children is legally restricted and ethically delicate. The gap here can rarely be closed with more tracking, but with structured market research using consent-based panels.
AI accelerates production. If approval is the bottleneck, it accelerates nothing, it only lengthens the queue. This dimension examines where in the process time is actually lost.
In our experience the greatest leverage lies not in creation but in research, variant generation and coordination. That is good news, because this is exactly where generative AI is reliable today.
Not the question of whether people can write prompts. The question of whether they can judge an AI result on the merits. The Ipsos KI-Monitor 2026 shows that only 15 percent of users in Germany forgo checking the results. That is a good sign and at the same time the actual capability need.
For March 2026, Bitkom names the shortage of skilled staff at 70 percent of German companies as a hurdle for AI. In practice, technical knowledge is missing less often than the ability to recognise a plausible sounding result as factually wrong.
Since 2 August 2026 the transparency obligations under Article 50 of the EU AI Act apply. For brands with a young audience, Article 5 is more relevant, which since February 2025 prohibits exploiting age-related vulnerability. In addition there are the DSA guidelines on the protection of minors and, in Germany, the Jugendmedienschutz-Staatsvertrag.
This dimension is often pushed to the end. That is a mistake, because it determines which use cases are eligible at all. A use case that later fails the legal review has cost money twice.
The four dimensions produce a matrix of impact and feasibility. Use cases with high impact and a good data situation come first. Use cases with high impact and a poor data situation become data projects, not AI projects. That is the most important distinction in the whole procedure.
Realistically, two or three initiatives remain that can show measurable impact within 90 days. For each of them, the target metric, the responsible person and the stop criterion are defined in advance. The stop criterion in particular is almost always missing, and it is the reason why pilots keep running without ever ending.
Depending on the size of the organisation, the check itself takes between two and five working days, usually as a combination of interviews, a review of existing data and a closing workshop.
A readiness check is not a preliminary stage before the actual work, it is the part that decides its success. It costs a few days and prevents the typical chain of tool purchase, pilot phase, fading out and restart.
For brands targeting families there is the added point that two of the four dimensions, data and governance, place particular demands in this segment. Anyone examining them early saves themselves the uncomfortable questions later.