In the previous unit, you explored how Zapier scaled AI fluency by embedding learning into existing work rituals — from formalized Q&A channels to a three-level fluency framework that reshaped hiring and onboarding. But here's the provocation Brandon Sammut delivered next: almost everything covered so far — the fluency programs, the 97% usage rate, the peer support infrastructure — falls under AI adoption, not AI transformation. The distinction matters enormously for where you set your ambition. As Brandon framed it, adoption tends to be "pretty siloed pretty department specific or team specific" and produces results "measured in like marginals like maybe a 10% improvement." Transformation, by contrast, involves "a pretty radical rethink of how stuff gets done" through cross-functional groups "playing for improvements measured in multiples right 2x 3x 5x." Both matter — they're not sequential and not mutually exclusive — but confusing one for the other caps your ceiling before you've even started.

The most compelling proof of this distinction came from Zapier's customer support leader, Lauren, whose approach Brandon highlighted as a model for how to set transformation-level goals. Rather than defaulting to cost reduction or speed alone, Lauren built her goals across three dimensions: efficiency, quality, and employee engagement. Her support team was already tracking handle time and customer satisfaction, but she insisted that how the team experienced the change mattered just as much. As Brandon put it, Lauren understood that She led the team through a 12-month transformation of their ways of working — and the engagement results were striking. The team saw on a scale of 100. That wasn't a side effect of getting efficiency right — it was the direct result of naming engagement as an explicit goal from day one. For anyone inclined to frame AI transformation purely as a cost story, Lauren's framework is a powerful counter-argument: isn't a hope — it's a designable outcome.