AI Avatars in 5-6 Years? (Eric Yuan, Zoom CEO)

For us - it's a matter of weeks

We can beat them to it

Our meeting bot has the unique ability to speak in meetings. By combining it with Speech-to-Speech APIs and Image to Video APIs we can hit a ping as low as 500 milliseconds, while human vocal latency typically ranges from 200 to 600 milliseconds.

We're on track to deliver realistic speaking meeting bots within the next month, on Zoom, Google Meet and Microsoft Teams.

From passive listeners to Active Participants

Hundreds of use-cases

Have you tried speaking with GPT-4 with voice-mode activated and a simple prompt? If so, you'll understand the number of use-cases possible: teacher, salesperson, HR recruiter, project management, knowledge-base, interpreter, ...

It's not a gimmick - AI agents with access to personal and company knowledge are already on the market.

Companies such as Twin.so and Lindy AI are offering digital clones of human office workers that can complete tasks. These tasks can, of course, include conversational ones.
Today for this session, ideally, I do not need to join. I can send a digital version of myself [...] in order to listen to the call but also to interact with a participant in a meaningful way [...]
Zoom CEO

Digital twins in your meetings

Eric Yuan is steering Zoom with a focus on AI-powered "digital twins", recruiting AI engineers and restructuring the company (i.e., firing some departments). The goal is to develop avatars that can attend meetings for users within 5-6 years.

That's a new gold rush

Let's sell picks and shovels

Meeting Bots as a Service: Attend, Interact

Attend: We already offer and distribute meeting bots which passively record and transcribe.

Interact: Our API will allow anyone to programmatically send virtual clones into meetings that can speak.

Key Advantages

Multi-Platform Support: One single API for Meeting Bots on Google Meet, Zoom, and Microsoft Teams.

High Barrier to Entry: Building and maintaining meeting bots is resource-intensive. Building requires months of work by experienced developers, and maintaining the system requires a dedicated team due to constant platform updates.