Separator

Kwantics: Creating Speech AI for the Indian Languages

Separator
Anant Tripathi & Chakrapani Mishra,Co-FoundersLocal languages digital content will witness greater acceptance from Indian language users in near future. In a KPMG report of 2017, the Indian language internet user base will grow at a CAGR of 18% to reach a staggering 536 million by 2021 compared to the English internet user base growing at just 3% to reach 199 million. This in turn has spurred the need to build robust and bolstered speech recognition systems for Indian Vernacular languages.

The need of working on improving real-time translation of local languages using AI and NLP tools has already seen the light of the day. And yet, ample gap lay wide open in speech automation in local languages. Bridging this gap, a Noida, UP-based company Kwantics, is changing the panorama of Speech-based Automation in Indian Vernacular languages with its exquisite flag-ship and patented products.

Enabling Speech AI Solutions in Multiple Languages
Kwantics, has launched Speech AI solutions crafted specifically for Indian use cases. Kwantics provide voice automation or voice bots that are used to automate conversations
with customers. Kwantics' flagship products of Voice Bots or Voice Assistants are enabled in various domains including BFSI, etc. This flagship product produces a highly scalable process, instead of increasing clutters. With each call, numerous leads, information, and data get generated.

With increased vernacular language users for different products, it becomes easier for automated products to track the process, instead of human beings. This increases the cost-effective ratio as well by reducing human agent conversation constraint, and bringing in transparency in the conversations. Because bots are recording the transcription, the bot is also recording several other parameters, which brings out the insights of the conversations automatically.

Kwantics, also provide Other Speech AI solutions like Speech Analytics, Speech Recognition (Speech-to-Text), and Speech Synthesis (Text-to-Speech). These three major tools are specifically crafted to make customer-specific messages out to the customers on phone, Whatsapp, etc. These messages are audio messages that are sent that sound exactly like human and natural-sounding voices. Once a user picks up the phone call, the AI-powered speech tool speaks and re-minds the user about the nature of work or message to be imparted, thereby incredibly increasing customization, and efficiency. Speech Recognition or Speech to Text tool creates more satisfying self-service inter-actions with customers and enables callers to input information using voice commands. With Speech Synthesis solutions, businesses can create brand voices with accents and emotions powered by Speech AI and deep learning.

The Backend Hustle
As challenging and exciting as this area is, AI-based Voice Bots need rigorous training. Kwantics has its trained engines, proprietary speech recognition, and speech synthesis engines which have been trained using 1000s of hours of Indian dialect speech data. With impeccable training resulting in robust flagship and products under patent, this2017 based startup has etched its journey to engage with some of the marquee clients including insurance providers, large credit card companies, etc.

With this journey, Kwantics has seen recognitions coming their way too, one being Nvidia's Inception program, recognized by NASSCOM. Kwantics was also the winner for Speech AI category in the Maxlife Innovation lab Accelerator program as well.