AI Sales Training Is Producing 'Smarter Idiots,' Says the Analyst Who's Spent Decades Measuring What Actually Works

Dave Kurlan’s warning about AI sales training is blunt: it often produces “smarter idiots” — reps who outsource thinking to a tool rather than using it to sharpen judgment. Writing for Objective Management Group, Kurlan compares eight sales-development formats (AI tools, LMS platforms, live virtual training, in-person workshops, virtual and in-person coaching, blended learning, self-directed content) and finds the format itself barely matters: two reps who sit through the identical workshop can land in “completely different places.” What actually differentiates outcomes, in his framework, is whether a manager diagnoses specific weaknesses, verifies the new skill shows up in real calls, and coaches persistently afterward — not which training vehicle delivered the content. His proposed fix is a five-step sequence (baseline evaluation, process refinement, manager coaching development, salesperson training, AI-assisted call inspection) built around a single named owner accountable for the whole blended strategy, because he argues success rests on human follow-through, not any one methodology or tool.

The mechanism behind “smarter idiots” shows up as a quantified risk in Outreach’s own AI-training research from a month earlier. Outreach’s guide cites Gartner data that reps overwhelmed by too many tools are 45% less likely to hit quota, and names tool fragmentation and treating AI outputs as final verdicts — rather than inputs a manager still has to check — as two of the primary failure modes, alongside poor underlying data quality. The same piece notes only 26% of reps get weekly one-on-one coaching from a manager in the first place, which is the exact gap Kurlan’s argument depends on: AI can’t compensate for coaching that isn’t happening.

For sales leaders buying AI coaching platforms on the promise that the tool itself will fix rep performance, the two pieces together are a specific counter-claim: the tool doesn’t do the fixing. An accountable manager who still checks the AI’s output against a real call does.