Running AI Literacy Training for a Marketing Team That Was Skeptical

I was brought in to run AI literacy training for a marketing team that had sat through two previous vendor-led sessions and come out more skeptical than before, not less. The vendor sessions had both opened with a pitch about transformation and productivity gains. The team, reasonably, had heard that pitch enough times from enough tools to tune it out on reflex.

What I changed about the format

I dropped the pitch entirely and opened instead with failure cases — specific, concrete examples of AI tools getting things wrong in ways relevant to their actual work: a hallucinated statistic in a drafted press release, a tone mismatch in generated ad copy, a factual error in a summarized competitor analysis. Showing the failure modes first, before any capability claims, did more to earn attention than any productivity statistic would have.

A skeptical team doesn't need to be convinced AI is useful, they need to be shown you already know where it breaks, before they'll trust anything you say about where it works.

What landed and what didn't

• A hands-on exercise where the team deliberately tried to get a tool to produce a wrong or embarrassing output built more genuine engagement than any feature walkthrough.
• Framing the tool as a draft generator requiring review, never as an autonomous replacement for their judgment, matched how they were actually going to be expected to use it, which mattered to how seriously they took the training.
• A specific, written escalation path for when the tool got something wrong addressed the unspoken accountability worry that no amount of feature demonstration would have touched.

Adoption after the training was noticeably higher than after either vendor session, with no change to the tool itself. The team's original skepticism wasn't a mindset problem to be corrected — it was accumulated experience with people overselling the tool, and the training that worked was the one that didn't repeat that mistake.