product-launch · July 22, 2026
By Joe Jurczyk, Editor
Synthesia moves from video generation into live AI role-play training
Synthesia launched Roleplay Sessions, in which employees practise conversations with a responding AI avatar that pushes back and scores them against a rubric. It is the first product under a broader Sessions platform and is enterprise-only at launch.
On 22 July 2026 Synthesia introduced Roleplay Sessions, an interactive training product in which employees rehearse workplace conversations — sales pitches, performance reviews, customer complaints — with an AI avatar that responds, offers pushback and scores performance against a rubric. Synthesia described Roleplay as the first release under a wider "Sessions" platform it intends to extend to job interviews and candidate screening.
CEO Victor Riparbelli framed the product around practice rather than information delivery, saying "we learn the best by actually practicing something rather than just reading it". A company spokesperson said early commercial rollouts include one of the top three European companies by market cap, one of the top five Fortune 100 and one of the largest recruitment companies; none were named. The product is enterprise-only at launch, with plans to reach small businesses, prosumers and education within months. No pricing was disclosed.
Why it matters for digital humans: This is the clearest signal yet that the avatar-video category is moving from rendering to interaction. Synthetic presenter video is a commodity with heavy price pressure; a real-time agent that listens, interrupts and evaluates is a different product with different technical requirements — latency budgets, turn-taking, and an assessment layer that has to survive contact with L&D buyers. The customer claims are unnamed and company-supplied, and the scoring rubric is the part that will decide whether this is training or theatre: automated assessment of soft-skills conversation is not a solved problem, and the article reports no validation evidence.
Separately, on 15 July 2026 Synthesia released an Interactive Avatar API for embedding real-time, lip-synced avatars, alongside an Assistant feature, the Express-3 avatar model, and Dubbing 2.0 with improved lip sync and translation (source: Synthesia blog).