Nearly half of test subjects mistook Tavus’ AI video avatar for a real person on a one-minute call
What happened
Tavus launched Griffin, an AI video avatar designed for real-time, two-way video calls. Griffin processes facial expressions, tone, and gestures during live interaction. In a Tavus study, 48 percent of participants thought Griffin was a real person after just one minute on the call. By comparison, earlier AI video systems only convinced up to 2 percent.
Why it matters
Griffin raises the bar for AI in customer engagement, digital sales, and virtual support. The ability to mimic real human interaction with facial cues and voice nuances makes AI-driven video calls more credible and persuasive. This pushes companies relying on scripted or one-way video to rethink their outreach strategies. It also intensifies pressure on AI detection systems and platforms to identify synthetic personas, since nearly half of users are now tricked. For businesses, this means AI avatars could scale personalized, live-like conversations without human reps, cutting costs while maintaining engagement.
What to watch next
Watch how Griffin integrates with existing CRM and sales workflows and whether it gains traction beyond demos and tests. Competitors will need to upgrade AI avatars with similar or better real-time interaction to stay relevant. Regulators and platforms may also review transparency rules and limits on AI impersonation to combat fraud risks as this tech becomes more convincing. Finally, adoption by enterprise customer support and marketing teams will reveal if the tech shifts budgets from live agents to AI-powered avatars in high-volume user outreach.
AI Quick Briefs Editorial Desk