According to TechCrunch AI, Synthesia's new Roleplay Sessions product represents a notable strategic shift. The company is betting that the real enterprise value lies not in generating training content, but in measuring its impact—a claim that, if true, could help it build a more defensible moat beyond its proprietary avatars. However, the success of this model likely hinges on the quality and fairness of its automated feedback rubrics, an area where AI systems have historically struggled with nuance and bias. In our view, the product's ultimate test will be whether enterprises trust its data enough to make consequential personnel decisions, a leap many may be hesitant to take.