Synthesia now lets you make AI avatar presentations using a webcam

Among the new updates Synthesia is launching is the ability to produce AI avatars using webcams or a phone, “full body” avatars with hands and arms, and a screen recording tool that shows an AI avatar guiding you through what you’re watching.

Synthesia, which says it’s used by nearly half of the Fortfune 500, uses AI avatars for all kinds of purposes.

These can range from creating tailored training videos to guide employees around certain processes, or generating promotional material that can be shown in the form of a video rather than an email or other textual communications.

But that hasn’t always been the case. According to co-founder and CEO Victor Riparbelli, in the first three years of the company’s story, Synthesia actually started out trying to sell its technology to Hollywood agencies and big-budget video production companies. The firm used computer vision for an AI dubbing tool that made mouth movements more lifelike for different languages.

“What we figured out was that the quality threshold to do anything with these guys was so big, no matter what we do, we’ll be a very small part of a much bigger process,” Riparbelli told CNBC in an interview at the firm’s London office.

“What was more interesting was the democratization aspect of: There are millions of people in the world who want to make video, but they’re not making video today because they don’t have the budget.”

In an Apple-style keynote, Synthesia’s CEO unveiled the firm’s new products, touting them as a more productivity-focused suite of tools for use by businesses, rather than just a platform that offers AI avatars.

One of the biggest new features the firm showed off was an option to make AI-generated avatars by recording less than five minutes of footage using a webcam or your phone. You can also clone your voice to have the avatars speak in multiple different languages

Typically, to make an AI avatar using Synthesia’s platform, you have to go into a studio in-person. Human actors go into a recording booth, record their voice, and perform lines in front of a green screen on an actual filming set.

This is all training data to provide Synthesia’s AI algorithm with the facial and vocal nuances it needs to come up with humanlike avatars that speak in an expressive way. Earlier this year, Synthesia debuted new expressive avatars that can convey human emotions, including happiness, sadness, and frustration.

But now, Synthesia is introducing new software which will make it easier for users to produce a digital version of themselves from anywhere, using just a webcam and Synthesia’s software.

The company is also launching the ability to create full-body avatars. This is different to Synthesia’s current avatars, which are limited to just portrait view. Now, you can go into a studio with dozens of cameras, sensors and lights all around you to make avatars that can move their hands.

Generating hands is something that’s traditionally hard for AI to do — often because hands are only a small part of the human body and not typically the focus in visual content.

Synthesia also debuted the option to play videos of AI avatars speaking in any language they like, whether it’s English, French, German, or Chinese.

In the future, Synthesia says, it will be able to tailor AI avatars for different countries: For example, a Nigerian avatar running a user through a tutorial rather than an American.

Synthesia also launched a new AI video assistant which can produce summaries of entire articles and documents. This could be a human resources specialist making a quick video explaining company benefits packages, for example.

Another big feature the company is rolling out is a new screen recording tool, which shows an AI avatar guiding you through what you’re watching.

In CNBC’s interview with him, Riparbelli characterized what Synthesia is trying to do as an enterprise-focused product overhaul, which would make it more akin to giants like Microsoft, Salesforce, and Zoom in the enterprise category.

“The world has been blown away by this stuff for the last 12 to 18 to 24 months, which is awesome,” Riparbelli told CNBC.

“But now we have experimented a lot, and we have found out the right use cases for these technologies that have lasting business value. They’re not like just a short-term PR moment.”

“You need to do that business goal of reducing customer support tickets by showing videos instead of text; or sell by making videos instead of just sending out emails,” he added.

“Now people are creating workflows around that. They need better ways to achieve their business goals, not just an interface with AI models. That’s where we’re going as a company.”

Last year, Synthesia raised $90 million from investors including U.S. chipmaker Nvidia and venture capital firm Accel, in a funding round that valued it at $1 billion and giving it “unicorn” status.

The company’s competitors include AI video tools Veed, Colossyan, Elai, and HeyGen. And Chinese-owned social media app TikTok also recently debuted Symphony Assistant, a product that allows creators to make their own AI avatars.

 The company makes money through a number of subscription pricing plans ranging from $22 for a “starter” plan and $67 for a “creator” plan, to custom “enterprise” plans where pricing is based on negotiations with Synthesia’s sales team.