Sontana is a conversational avatar created by NECTEC-NSTDA. It is designed to be used by businesses as a virtual assistant to provide a unique and personalized experience to their customers.
Dr. Chaianun Damrongrat, member of NECTEC Speech and Text Understanding Research Team, said that over the years, NECTEC has developed a range of artificial intelligence (AI) platforms for both speech and text recognition, including a speech-to-text platform, a text-to-speech platform, a text analysis platform and a chatbot platform. These natural language processing tools work in both Thai and English.
“To create a solution to labor shortage in customer service, the research team integrated three platforms that are already in service, namely a speech-to-text software Partii, a chatbot software Abdul and a text-to-speech software VAJA and combined them with a computer-generated 3D character to create Sontana - a conversational avatar for responding to frequently asked questions (FAQs) provided by clients. Sontana can respond to questions in just 5 seconds with lifelike facial expressions and gestures.”
Dr. Ausdang Thangthai, member of NECTEC Speech and Text Understanding Research Team, further explained that the blendshapes technology was employed to create facial expressions for the computer-generated character to match the speech, making it more human-like. Sontana’s voice, appearance and facial expressions can be customized to meet the requirements of clients.