Should Agencies Teach Prompting Or Thinking?
Should agencies teach prompting or thinking? Training budgets are already moving towards AI: workshops on tools, prompt libraries, internal guides on which model to use for what. The case for prompting is obvious, because people who use the tools well are faster. The case for thinking is less visible, and it is the one more likely to be quietly dropped.
Both matter. They do not matter equally over time.
Why agencies teach prompting first
Because it pays back immediately and can be measured. A team that learns to brief a tool well produces more in a week, and the improvement is visible in output.
The IPA Agency Census 2025 found 88.3% of UK agencies saying AI had considerably affected how they work. Training follows that kind of change, and prompting is the most teachable part of it.
The tool changes every six months. Knowing what to ask it for lasts a career.
What thinking covers that prompting does not
Knowing what the problem is. Telling a strong idea from a plausible one. Understanding why a piece of work will land with one audience and fail with another. Arguing for an idea when the client is nervous.
Those used to be learned slowly, by doing the early drafts and having them torn apart. When the tool produces the early drafts, that apprenticeship thins out unless someone rebuilds it deliberately.
What a balanced programme looks like
Teaching the tools as tools, briefly and practically, and spending the larger share of time on critique: reviewing work together, explaining decisions, asking juniors to defend a choice rather than generate another option.
It is the slower investment. Agencies that skip it may find in a few years that they have many people who can operate a tool and few who can say whether its output is any good.
More on Ad Tribe Work Life.