Post by Paul Burani

The biggest gains from AI roleplay are not evenly distributed. And that’s exactly why it works! A field study of ~2,000 reps surfaced something important about AI roleplay and sales performance. The impact is highly targeted, not random.  * Low performers saw the biggest uplift, +35%  * High goal-setters improved by ~30% * Other cohorts saw gains too, including highly-experienced reps and those with strong manager support  This groundbreaking research tells us something critical: 𝗔𝗜 𝗿𝗼𝗹𝗲𝗽𝗹𝗮𝘆 𝗮𝗺𝗽𝗹𝗶𝗳𝗶𝗲𝘀 𝗿𝗲𝗮𝗱𝗶𝗻𝗲𝘀𝘀 𝗮𝗻𝗱 𝗲𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝗰𝗼𝗻𝘁𝗲𝘅𝘁. It accelerates people who: * Are motivated to improve * Have enough context to apply what they learn * Are supported by strong coaching That’s exactly how real performance systems behave. Which is why the most effective implementations don’t treat roleplay as isolated training. They: 🎯 Tie scenarios to real customer conversations 🧑‍🏫 Use managers as the bridge between practice and execution 🔁 Create continuous loops of practice, application, and feedback 📈 Track behavioral change inside real sales cycles — At Atlas Primer we believe that AI roleplay isn’t about merely simulating conversations. It’s about improving the ones that actually matter. Link to the study is in the comments. Thank you to Johannes Habel, Michael Ahearne and their colleagues at the University of Houston, C.T. Bauer College of Business for the gift of insight! ❓ How tightly is your training connected to the conversations your reps are having this week?