Vocinity and Synthesia Debut Custom Virtual Being Brand Assistants
Video experience developer Vocinity has partnered with virtual being startup Synthesia to produce customized interactive AI described as ‘conversational brand assistants’ starting with a virtual human named Gabbi. This virtual assistant uses AI to mimic the look and sound of a human being that can interact with customers, carry out administrative tasks, and train new employees in their roles, among other functions.
Vocinity’s platform offers a multimodal conversational AI service with minimal coding necessary to design the avatar. The new brand assistants simulate a video conversation with someone using advanced models to layer appropriate visuals and audio over the AI’s responses. The conversations are realistic to the point the AI will even pause when interrupted by the user. The avatar can be activated from different points, as seen in the image on top. The brand assistant might be deployed on websites, digital signs, or a QR code could pull it up on a mobile device. The multilingual assistant will then try to help the user or transfer them to a human if not able to complete the requests.
“Real-time video engagements are more complex than voice and chatbot technologies,” Vocinity CTO Nathan Stratton explained. “So to ensure great user experiences, we made NLU enhancements and developed proprietary error correction, noise reduction, adaptive bandwidth throttling, and other algorithms.”
Video-based conversational AI has exploded as a concept over the last few years. The MetaHuman Creator released a year ago by Unreal set off an accelerating rollout among voice and visual AI developers. Startups like Hour One, Supertone, Resemble AI, Veritone, and DeepBrain are all grabbing cash and clients as AI-powered versions. Most recently, metaverse human avatar generator Ready Player Me raised $13 million for improving its three-dimensional digital character platform. Virtual beings are popping up everywhere. YouTube star Taryn Southern’s virtual clone, the story-telling virtual William Shatner, Nestle Toll House’s virtual “cookie coach” Ruth, and Sber’s Russian television host are just the beginning.
Synthesia, which raised $50 million in December, analyzes video to generate realistic simulacra able to do and say new things from a script that are unrelated to the analyzed video. The user can customize the final video or audio product and fine-tune it to match their vision without repeated retakes or the more usual editing software suite. Vocinity had used Synthesia’s existing avatars before, but Gabbi is the first avatar explicitly designed for a brand. Gabbi serves as both a helpful tool and a demonstration of how Vocinity can leverage Synthesia’s tech on its platform.
“We tested several custom AI avatars before recommending Synthesia to our clients,” Vocinity CEO Michael Riemer explained. “We have been working with them now for more than a year and are very pleased with the results,”