India is emerging as a key producer of AI for virtually every sector, backed by thousands of startups that serve the country’s multilingual and multicultural population and expand to users around the world.
The country is one of the six main world economies leader in generative AI adoption and has seen rapid growth of its startup and investor ecosystem, reaching over 100,000 startups this year, from less than 500 in 2016.
More than 2,000 are part Creation of NVIDIAa free program for startups designed to accelerate innovation and growth through training and technical tools, go-to-market assistance, and opportunities to connect with venture capitalists through the Creation of the VC Alliance.
At NVIDIA AI Summitwhich will take place in Mumbai until October 25, around 50 India-based startups will share AI innovations that will impact areas such as customer service, sports media, health care And robotics. These Inception members will present their solutions on site at the Startup Pavilionduring round tables and in a startup pitch session. Startups can also attend a reverse pitch session where venture capitalists share their vision for the next wave of innovation.
Conversational AI for Indian Railways Customers
Bangalore-based startup CoRover.ai already has over a billion users of its LLM-based conversational AI platform, which includes text, audio and video agents.
“NVIDIA Inception support helps us advance our work to automate conversational AI use cases with large, domain-specific language models,” said Ankush Sabharwal, CEO of CoRover. “NVIDIA AI technology allows us to deliver enterprise-grade virtual assistants that support 1.3 billion users in more than 100 languages. »
CoRover’s AI platform powers chatbots and customer service applications for major private and public sector clients, such as the Indian Railway Catering and Tourism Corporation, the official provider of online tickets, drinking water and food for Indian stations and trains.
Double Ask DISHAbased on the Sanskrit word for direction, IRCTC’s multi-modal chatbot handles over 150,000 user queries daily and has facilitated over 10 billion interactions for over 175 million passengers to date. It helps customers with tasks like booking or canceling train tickets, changing boarding stations, requesting refunds, and checking their reservation status in languages like English , Hindi, Gujarati and Hinglish, a mixture of Hindi and English.
The deployment of AskDISHA resulted in a 70% improvement in IRCTC’s customer satisfaction rate and a 70% reduction in queries through other channels such as social media, phone calls and emails.
CoRover’s modular AI tools were developed using NVIDIA NeMoan end-to-end cloud-native framework and suite of microservices for developing generative AI. They run on NVIDIA GPUs in the cloud, which allows CoRover to automatically increase computing resources during peak usage, for example when train tickets are released.
Watch the CoRover session live at AI Summit or on demand, and learn more about Indian companies build multilingual language models with NeMo.
Powering the future of sports media
VideoVerse, founded in Mumbai with offices in six countries, has built a family of AI models to support content creation in the sports media industry – enabling global clients including the Indian Premier Cricket League , the Vietnam Basketball Association and the Mountain West Conference for America’s college football to generate game highlights up to 15 times faster and increase viewership.
“Short-length video clips that can be easily shared on social media can also help lesser-known sports attract public attention and grow their fan base,” said Vinayak Shrivastav, CEO of VideoVerse. “AI-assisted content creation allows emerging sports like longball and kabbadi to gain exposure with a limited marketing budget.”
VideoVerse’s enterprise solution, called Magnifi, uses AI technologies such as vision analytics, natural language processing and optical character recognition to streamline editing workflows by detecting players, by identifying key moments and tracking the movement of the ball from multiple camera angles. Magnifi also automatically adjusts video sizes for horizontal and vertical formats on laptops, tablets and phones, ensuring the main action remains centered in the frame.
VideoVerse uses NVIDIA CUDA libraries to accelerate AI models for image and video understanding, automatic speech recognition, and natural language understanding. The company runs its custom AI models on NVIDIA Tensor Core GPU for inference.
Watch the VideoVerse session live at the AI Summit or on demand.
Rewriting the narrative of business efficiency
Mumbai-based startup Smooth AI offers generative AI chatbots, voice call robots and a range of application programming interfaces to improve business efficiency. Its AI tools can access an organization’s knowledge base to provide teams with information, reports and insights, or to help answer questions accurately.
Fluid AI chatbots can be applied to customer service to increase agent productivity and reduce response times, generating accurate results in real time. Organizations can also choose to deploy them to sales and customer-facing teams, using them for tasks such as creating slide presentations in less than 15 seconds.
Smooth AI Taps NVIDIA NIMs microservices, NVIDIA NeMo platform and the NVIDIA TensorRT inference engine to provide a comprehensive and scalable platform to develop personalized generative AI for its customers.
The company is also studying the use of NVIDIA Riva microservices to develop a voice experience for its chatbots that will help significantly reduce latency and deliver more fidelity experiences. Its AI models run on NVIDIA GPUs in the cloud.
“Our work with NVIDIA has been invaluable: the low latency and high fidelity we deliver for AI voice calls comes from the innovation that NVIDIA technology allows us to achieve,” said Abhinav Aggarwal, founder of Fluid AI.
Watch the Fluid AI session live at the AI Summit or on demand.
Delivering data work to bridge the digital divide
Karya, based in Bangalore, is a smartphone-based digital workplace platform that enables members of low-income and marginalized communities across India to earn additional income by completing language tasks that support the development of business models. Multilingual AI.
Nearly 100,000 Karya workers record voice samples, transcribe audio files, or check AI-generated sentences for accuracy in their native languages, earning nearly 20 times India’s minimum wage for their work. Karya also pays royalties to all contributors whenever its datasets are sold to AI developers.
“By compensating these communities fairly for their digital work, we are able to improve their quality of life while supporting the creation of multilingual AI tools that they can use in the future,” said Manu Chopra, CEO of Karya.
Karya’s work helps companies accelerate the design and data collection process, enabling the creation of deployable AI solutions that cater to non-English speakers in India. The company will use NVIDIA NeMo and NVIDIA NIM to build its AI platform, which offers custom AI model training and pre-trained models tailored to customers’ business needs.
Companies and research centers can purchase the datasets collected by Karya to train diverse and multilingual AI models. For example, Karya is working with the Bill and Melinda Gates Foundation to create the largest open-source, gender-intentional AI dataset in Indian languages to date. Karya employs over 30,000 low-income women across six language groups in India to help create the dataset, which will support the creation of various AI applications in agriculture, healthcare and the bank.
Watch Karya’s session live at the AI Summit or on demand.
To learn more about the AI Summit, Watch NVIDIA Founder and CEO Jensen Huang’s Fireside Chat with Mukesh Ambani, Chairman and Managing Director of Reliance Industries.
