As businesses across segments seek new ways to transform customer interactions and improve retention, advancements in AI like natural language processing, speech recognition, and computer vision technologies are contributing to new market opportunities and innovation. DeepBrain AI, a leading conversational AI startup from South Korea, specializes in video synthesis with a focus on creating realistic AI avatars that enable businesses to automate and enhance customer interactions and increase operational efficiency. The company has secured a strong presence in the market across industries, including media, finance, and retail.
I sat down with CEO and founder Eric Jang to discuss his vision behind DeepBrain AI, the evolution of their AI technology, what makes them stand out from the competition, and what being a member of the Pegasus Program has meant for his company.
Providing a superior product in a growing field
DeepBrain’s revolutionary media synthesis technology enables the creation of AI avatars capable of responding to natural language questions. By capturing video of a human model in a studio and training a machine learning system, DeepBrain generates real-time avatars with synchronized lip, mouth, and head movements. Broadcasting companies, for example, use this technology to create AI news anchors with a synthesized voice and face.
Eric told me that the spark behind DeepBrain’s creation was initially a chatbot system developed for a major bank in South Korea, but quickly realized that wouldn’t suffice to capture the market.
“Because there are a lot of companies in competition for text, we wanted to improve the conversational aspect of our solution,” Eric said. “We needed to focus on our main core technology, so we developed a speech synthesis tool to add to the text, and then finally a video component. This development roadmap led our conversational AI to focus on our AI avatar and video synthesis technology.”
Taking advantage of a sizeable market lead
Powered by Azure OpenAI Service and ChatGPT, DeepBrain offers highly engaging and interactive customer support experiences. With the use of AI avatars, the startup’s clients can provide personalized and efficient customer service, resulting in increased satisfaction and loyalty. And since they have been finetuning their technology since 2018, Eric says they have a significant head start compared to other competitors.
“We have more sophisticated lip-sync technology, for example, that is more enhanced than other companies,” Eric explains. “We are currently providing 4K high resolution video for our conversational AI with reduced latency—within 1.5 seconds, sometimes as little as 1.2 seconds. We were the first player in the AI avatar market, and our tech continues to stay on the cutting edge.”
Using AI technology in a responsible way
Concerned about the discourse regarding the responsible use of AI, Eric and his team have developed a tool to combat what he sees as potential for malicious use of video synthesis solutions.
“With deepfakes and face swaps, many people are worrying about the abuse of AI,” Eric says. “Even though AI avatars are a different technology, we have also developed a deepfake detection tool that we provide to our clients to defend against potential abuse.”
Leveraging Azure to level up their AI offerings
The foundation of DeepBrain’s tech stack relies heavily on Microsoft solutions. In addition to Azure OpenAI Service, Eric says the raw processing power required to run their AI avatars would not be achievable on their own.
“We require a lot of GPUs to provide our machine-learning service,” Eric explains. “We already have on-prem GPUs, but we rely on Azure cloud for our research and to deliver service to our customers.”
“ChatGPT can not only create an auto-generated script, but it can make the whole video. It can create a title, subtitle, images, and script to automatically generate everything the user needs.”
One crucial aspect of DeepBrain’s service is employing a reliable and accurate speech-to-text (STT) engine to seamlessly convert a user’s voice into the written word. Eric says that after rigorous testing of competitors, Azure’s STT engine outperformed them all.
“It was the clear choice for us,” he told me. “Its exceptional performance ensures that we capture every nuance of the user’s voice accurately.”
In addition to leveraging Microsoft’s STT engine, the company has integrated Azure OpenAI Service’s cutting-edge solutions into their application, specifically for their text-to-video solution and AI interview service.
“OpenAI’s advanced technology allows us to provide an unparalleled experience, enabling us to create compelling videos from text and deliver an innovative AI-driven interview service,” Eric says. “Microsoft has long been a pioneer in shaping our digital landscape, revolutionizing personal computing with their widely adopted PC operating systems.”
Automatically generating content for clients
DeepBrain’s integration with ChatGPT has leveled up the company’s AI Studios platform, which allows clients to create and customize AI avatars. A deep learning algorithm generates realistic and detailed avatars that look like real people, but traditionally required the client to write and upload a script for the avatar to recite. With ChatGPT and Azure OpenAI Service integration, however, scripts can be created automatically with only a few prompts inputted by the client.
“ChatGPT can not only create an auto-generated script, but it can make the whole video,” Eric says. “It can create a title, subtitle, images, and script to automatically generate everything the user needs.
“One more thing ChatGPT is helping us with,” Eric continues, “is an AI interview service. It’s like a real video job interview, but with an AI avatar as the interviewer. Questions can be automatically created with ChatGPT, and then clients can analyze the answers afterward.”
Pegasus Program brings cross-selling opportunities
Eric is quick to credit some of DeepBrain’s recent success with attracting new clients to the company’s participation in the Microsoft for Startups Pegasus Program. Intended to provide growth-stage startups with access to Microsoft technology, expertise, and resources to help them accelerate their growth, Eric says the most important benefit of the Pegasus Program is the ability for startups to connect with new potential customers via Microsoft’s network of clients.
“As a startup, we need clients and we need sales,” Eric says. “We need to represent our business. The Pegasus Program provides us with a chance to get introduced to Microsoft’s Azure Cloud Service clients and cross-sell our solution. So that has been very helpful for our company; it’s the most beneficial part of the Pegasus Program.”
Free credits help deployment into Azure cloud
Pegasus startups also have the opportunity to reach new customers by applying for credits for Azure, GitHub, and LinkedIn. DeepBrain received $250,000 in these credits, a number Eric calls “huge.”
“It’s a huge number,” he gushes. “We’ve used these credits for CPU and GPU consumption, which helped deploy our solution into the Azure cloud service. It’s a real benefit to joining the Pegasus Program.”
As DeepBrain AI has successfully differentiated its services across various industries, the applications of their AI avatars are vast and continue to grow. With a focus on innovation and global expansion, DeepBrain AI is poised to continue its position as a major player in the AI avatar space, transforming the way businesses and consumers interact.
For more tips on leveraging AI for your startup and for access to Azure’s AI services, sign up today for Microsoft for Startups Founders Hub. Members receive Azure cloud credits that can be used toward Azure OpenAI Service or OpenAI to help build their product.
