As companies throughout segments search new methods to rework buyer interactions and enhance retention, developments in AI like pure language processing, speech recognition, and laptop imaginative and prescient applied sciences are contributing to new market alternatives and innovation. DeepBrain AI, a number one conversational AI startup from South Korea, makes a speciality of video synthesis with a give attention to creating real looking AI avatars that allow companies to automate and improve buyer interactions and improve operational effectivity. The corporate has secured a robust presence available in the market throughout industries, together with media, finance, and retail.
I sat down with CEO and founder Eric Jang to debate his imaginative and prescient behind DeepBrain AI, the evolution of their AI know-how, what makes them stand out from the competitors, and what being a member of the Pegasus Program has meant for his firm.
Offering a superior product in a rising discipline
DeepBrain’s revolutionary media synthesis know-how allows the creation of AI avatars able to responding to pure language questions. By capturing video of a human mannequin in a studio and coaching a machine studying system, DeepBrain generates real-time avatars with synchronized lip, mouth, and head actions. Broadcasting firms, for instance, use this know-how to create AI information anchors with a synthesized voice and face.
Eric advised me that the spark behind DeepBrain’s creation was initially a chatbot system developed for a serious financial institution in South Korea, however shortly realized that wouldn’t suffice to seize the market.
“As a result of there are loads of firms in competitors for textual content, we needed to enhance the conversational facet of our resolution,” Eric mentioned. “We would have liked to give attention to our primary core know-how, so we developed a speech synthesis software so as to add to the textual content, after which lastly a video element. This improvement roadmap led our conversational AI to give attention to our AI avatar and video synthesis know-how.”
Benefiting from a sizeable market lead
Powered by Azure OpenAI Service and ChatGPT, DeepBrain gives extremely participating and interactive buyer help experiences. With using AI avatars, the startup’s shoppers can present personalised and environment friendly customer support, leading to elevated satisfaction and loyalty. And since they’ve been finetuning their know-how since 2018, Eric says they’ve a big head begin in comparison with different rivals.
“Now we have extra subtle lip-sync know-how, for instance, that’s extra enhanced than different firms,” Eric explains. “We’re at the moment offering 4K excessive decision video for our conversational AI with lowered latency—inside 1.5 seconds, generally as little as 1.2 seconds. We had been the primary participant within the AI avatar market, and our tech continues to remain on the innovative.”
Utilizing AI know-how in a accountable method
Involved concerning the discourse concerning the accountable use of AI, Eric and his crew have developed a software to fight what he sees as potential for malicious use of video synthesis options.
“With deepfakes and face swaps, many individuals are worrying concerning the abuse of AI,” Eric says. “Though AI avatars are a distinct know-how, we’ve got additionally developed a deepfake detection software that we offer to our shoppers to defend in opposition to potential abuse.”
Leveraging Azure to stage up their AI choices
The inspiration of DeepBrain’s tech stack depends closely on Microsoft options. Along with Azure OpenAI Service, Eric says the uncooked processing energy required to run their AI avatars wouldn’t be achievable on their very own.
“We require loads of GPUs to offer our machine-learning service,” Eric explains. “We have already got on-prem GPUs, however we depend on Azure cloud for our analysis and to ship service to our prospects.”
“ChatGPT can’t solely create an auto-generated script, however it could make the entire video. It may well create a title, subtitle, pictures, and script to robotically generate all the things the consumer wants.”
One essential facet of DeepBrain’s service is using a dependable and correct speech-to-text (STT) engine to seamlessly convert a consumer’s voice into the written phrase. Eric says that after rigorous testing of rivals, Azure’s STT engine outperformed all of them.
“It was the clear alternative for us,” he advised me. “Its distinctive efficiency ensures that we seize each nuance of the consumer’s voice precisely.”
Along with leveraging Microsoft’s STT engine, the corporate has built-in Azure OpenAI Service’s cutting-edge options into their software, particularly for his or her text-to-video resolution and AI interview service.
“OpenAI’s superior know-how permits us to offer an unparalleled expertise, enabling us to create compelling movies from textual content and ship an revolutionary AI-driven interview service,” Eric says. “Microsoft has lengthy been a pioneer in shaping our digital panorama, revolutionizing private computing with their extensively adopted PC working methods.”
Routinely producing content material for shoppers
DeepBrain’s integration with ChatGPT has leveled up the corporate’s AI Studios platform, which permits shoppers to create and customise AI avatars. A deep studying algorithm generates real looking and detailed avatars that appear to be actual individuals, however historically required the shopper to put in writing and add a script for the avatar to recite. With ChatGPT and Azure OpenAI Service integration, nonetheless, scripts could be created robotically with just a few prompts inputted by the shopper.
“ChatGPT can’t solely create an auto-generated script, however it could make the entire video,” Eric says. “It may well create a title, subtitle, pictures, and script to robotically generate all the things the consumer wants.
“Another factor ChatGPT helps us with,” Eric continues, “is an AI interview service. It’s like an actual video job interview, however with an AI avatar because the interviewer. Questions could be robotically created with ChatGPT, after which shoppers can analyze the solutions afterward.”
Pegasus Program brings cross-selling alternatives
Eric is fast to credit score a few of DeepBrain’s current success with attracting new shoppers to the corporate’s participation within the Microsoft for Startups Pegasus Program. Supposed to offer growth-stage startups with entry to Microsoft know-how, experience, and sources to assist them speed up their progress, Eric says crucial good thing about the Pegasus Program is the flexibility for startups to attach with new potential prospects through Microsoft’s community of shoppers.
“As a startup, we’d like shoppers and we’d like gross sales,” Eric says. “We have to symbolize our enterprise. The Pegasus Program gives us with an opportunity to get launched to Microsoft’s Azure Cloud Service shoppers and cross-sell our resolution. In order that has been very useful for our firm; it’s probably the most useful a part of the Pegasus Program.”
Free credit assist deployment into Azure cloud
Pegasus startups even have the chance to succeed in new prospects by making use of for credit for Azure, GitHub, and LinkedIn. DeepBrain obtained $250,000 in these credit, a quantity Eric calls “large.”
“It’s an enormous quantity,” he gushes. “We’ve used these credit for CPU and GPU consumption, which helped deploy our resolution into the Azure cloud service. It’s an actual profit to becoming a member of the Pegasus Program.”
As DeepBrain AI has efficiently differentiated its companies throughout numerous industries, the functions of their AI avatars are huge and proceed to develop. With a give attention to innovation and world enlargement, DeepBrain AI is poised to proceed its place as a serious participant within the AI avatar area, remodeling the best way companies and shoppers work together.
For extra recommendations on leveraging AI on your startup and for entry to Azure’s AI companies, sign up today for Microsoft for Startups Founders Hub. Members obtain Azure cloud credit that can be utilized towards Azure OpenAI Service or OpenAI to assist construct their product.
