Marketing was the first function to adopt generative tools and is the one where the output is public, so the cost of a wrong figure is a correction everyone can see. This track is about volume with a check on it: produce more, then prove it is right before it ships.
Generative tools produce confident numbers. “Companies using X see a 40% increase in Y” reads well, fits the slide, and frequently traces back to nothing: no study, no survey, no source that says that. Published, it is a claim your competitors can check and your own sales team will repeat. The drill in this track is mechanical and it works: every number in a draft gets an origin written next to it before the draft moves on, and anything without one is cut rather than softened.
Summarising a thousand open-text responses is the use case where these tools genuinely beat a human on speed, and it is also where a plausible theme can appear that nobody wrote. The method we teach is to make the model quote: every theme has to come back with the verbatim responses it was built from, so a theme with three weak quotes behind it gets dropped instead of becoming a roadmap item. It is slower than asking for a summary, and it is the difference between analysis and decoration.
Teams quickly learn to paste the tone-of-voice guide into the prompt and are then surprised that everything still sounds the same as everyone else's. The useful exercise is the reverse one: take a piece of output, mark the three sentences that could have been written by any company in your sector, and rewrite only those. That is a skill that survives the next model release, which is more than can be said for a prompt that worked last quarter.
Diagnostic, acceptable use policy, interactive tracks by function, adoption follow-up. The four steps work together, and the free 30-minute call scopes all of them.
See the four stepsAI training for marketing teams
The opposite is the point. Most of the track is about what to remove: the unsourced number, the sentence that fits any company, the structure the model defaults to. Teams that only learn to generate produce more average content faster; the editing judgement is the part that is worth training.
It depends on the tool, the contract behind it and whether the responses identify anyone. That is one of the rules the acceptable use policy settles for your company, naming your approved tools rather than leaving each marketer to decide alone.
We teach on the stack you have already decided on, whichever assistant that is. We do not resell software and we take no commission on any vendor, so the choice of tool stays a decision you make on your own grounds.
We go through how your teams use AI today, what is missing to frame it, and what the first quarter would look like. You leave with enough to decide, whether you work with us or not.