Computer vision applications have received a major boost from advances in smartphones and a combination of AI that helps bridge the gap between what everyday devices can't see. With these tools, a Munich startup called Beyond Presence believes it holds the key to what's next. It is a hyper-realistic avatar that looks and sounds exactly like a human and can be used in real-time conversational situations.
The startup announced today that it has raised its first external funding, a pre-seed round of $3.1 million. German firm HV Capital led the round, with participation from individuals from 10x Founders, Alba VC, Meta, DeepMind, and Zalando.
Beyond Presence hasn't actually released its technology to the world yet, but the company's CEO and co-founder Awais Shafique told TechCrunch that it will do so in the near future.
He said some of the new capital will be used to continue developing the company's foundational model, and some will fund the release of a public beta by the end of this month.
So far, around 300 companies are on the waiting list for the product, with potential early adopters in areas such as customer service and support, recruiting, sales and e-learning. These are all areas where companies are looking to expand their interactions with users without having to hire or train more people.
So how did a startup raise a multi-million round before even launching a product and get to this point as a resource-intensive AI company without any outside funding? The answer lies in the founder's background.
Shafique previously co-founded Presize, a computer vision startup that can accurately measure a user's measurements by recording a clip of the user turning once in front of a smartphone camera. These measurements could potentially be used to purchase apparel online. The Munich-based startup has gained huge attention in its home country after its founders took part in the German TV version of Dragon's Den/Shark Tank in 2020, raising a record 650,000 euros. Collected.
Then, in April 2022, it was revealed that Meta had acquired Presize. Although the price and exact date of the deal were not disclosed, there are some strong signs pointing to a major exit.
Meta spent $774 million on acquisitions from January to March 2022 and $1.15 billion in the first six months of 2022, according to SEC filings. Presize was the only deal announced at the time. Meanwhile, Wikipedia's list of meta acquisitions has been compiled to a much more conservative estimate of $100 million.
Sources told TechCrunch that in this case, the last nine digits of the transaction amount are more accurate than the 10 digits.
Regardless of where Presize falls on that spectrum, Shafiq estimates that he and his co-founders are doing very well considering the startup has only raised “a few million.” I can say that I have made an exit.
“I make enough money that I don't have to work,” he said in an interview.
(They're all still working anyway. In addition to Shafique's latest startup, Presize's other co-founders, Leon Szeli and Tomislav Tomov, are still at Meta, working on generative AI projects. They keep in touch with Shafique. Investing in Beyond Presence ).
At its heart, Presize aims to solve some of the toughest problems in computer vision, and Beyond Presence aims to do the same.
Admittedly, my initial thoughts on Beyond Presence's pitch were skepticism.
Digital avatars aren't entirely uncharted territory. Chatbots have not only been around for decades, but have only recently gained attention and perhaps even some hype due to the latest innovations in AI and processing, not to mention the general chatbot craze. It is.
There's also no clear idea of what will work as a long-term business in the Avatar space. Some companies, like Hyper and Ready Player Me (backed by Amazon and A16Z, respectively), are based on gaming culture and the theory that avatars shouldn't be realistic versions of the people using them. People who want to hide their identity to protect their privacy may choose these approaches rather than those that recreate their exact identity.
Some, like Synthesia, allow you to create a likeness of a person. But they are not the basis for the main product, which requires a visit to a studio and uses an unobtrusive device like a phone camera to capture the user's facial expressions and enhance an avatar that bears no resemblance to the user.
Others, like Veed and Deepbrain, are trying to build tools that can be manipulated using everyday devices.
Beyond Presence aims for avatars that look like you and can be used in real-time, unscripted conversational situations, as well as avatars that you can create on the fly using your phone. Something like the video version of text-based chatbots from OpenAI, Anthropic, Meta, and others.
Of course, this startup is not the only one moving into this field. Zoom appears to be working on ultra-realistic custom avatars, and CommonGround is also working on something in a similar direction.
Currently, Beyond Presence is focused on building something unique in the video component of its products.
Eleven Labs powers the audio portion of the experience, and OpenAI's GPT is a generative AI model that powers the content. Business customers have the flexibility to change the model they use.
“You could leave everything to me, but… [or use] Any LLM or any voice agent. Then I draw a face on it,” he said. He noted that this feature addresses the fact that enterprise customers may already be working with a particular LLM provider or may not want to share their proprietary data externally.
There are some interesting details about Beyond Presence's early funding decisions. It turns out the company was approved to join Y Combinator earlier this year, but turned down the prestigious incubator.
YC certainly provides access to a great network, but Shafique and his co-founder Felix Altenberger (CTO, former deep learning specialist at Presize) are not giving up equity and autonomy to access it. I chose that.
While YC's partners may have had their own opinions on how to grow the company through sales, Shafiq and Altenberger said they had a “horizontal” approach they wanted to pursue.
“We wanted to push the core technology and underlying model layer further,” he said.