Synthesia, a digital avatar startup originally based in the U.K., has created an interactive virtual avatar for a TechCrunch journalist. This marks the first time Synthesia has developed such an avatar for a journalist or anyone outside of its head of corporate affairs.
The journalist's interactive avatar is trained exclusively on their story concerning why venture-backed startups reportedly commit more fraud than non-VC-backed startups. It is designed to answer questions solely on this specific topic, redirecting queries that fall outside its training.
The creation process involved capturing numerous photos and a two-minute voice recording of the journalist. The interactive avatar's technology stack includes Synthesia's video and voice models, along with voice-to-text, agentic language, and text-to-voice models.
Synthesia offers various products, including a video-creation platform with classic avatars, an agentic platform called Sessions for interactive roleplay, and an API platform for combining its models with other tech services.