In Short:
In Bangalore, top tech leaders from Google DeepMind, Microsoft, and Meta witnessed the launch of Sarvam AI, dubbed India’s OpenAI. They introduced voice-based software designed for businesses to interact with customers in ten local Indian languages. This innovation aims to enhance AI accessibility in India, tapping into its vast population. Several startups are creating voice bots to automate tasks and improve customer interactions.
Introduction of Innovative Voice Technology
Earlier this month, executives from Alphabet Inc.’s Google DeepMind, Microsoft Corp., and Meta Platforms Inc. congregated with prominent tech founders in Bangalore to witness a groundbreaking product launch by one of India’s leading AI startups, Sarvam AI. This development holds the potential to revolutionize the utilization of AI technology in the world’s most populous country.
Sarvam AI’s New Offering
Sarvam AI, often referred to as India’s answer to OpenAI, unveiled its innovative software designed for businesses to engage customers using spoken voice communication, rather than limited text interactions. This new technology leverages data from ten native Indian languages and is priced strategically at one rupee per minute to broaden its market reach. During the launch event, Vinod Khosla, a billionaire venture capitalist and investor in Sarvam, remarked, “These voice bots have the potential to reach a billion people.”
Context of AI in India
In the wake of the global artificial intelligence boom sparked by the launch of ChatGPT, India has been striving to harness AI advancements. However, traditional chatbots have faced challenges due to insufficient data across many regional languages. While urban residents can often communicate with chatbots in English, a significant portion of the population lacks these language skills. Consequently, numerous startups believe that voice bots utilizing local language data may reach a broader audience both within India and possibly in international markets.
The Proving Ground for AI Voice Technology
These startups are positioning India as a testing ground for what could emerge as a new frontier in generative AI products, albeit with safety concerns that have been raised in other regions. By integrating AI voice functionalities, technology firms are aiming to create interactive, real-time conversational services that can automatically execute specific tasks. Currently, these developments are manifesting in diverse consumer and business applications across the country.
Examples of Local Innovations
For instance, Gnani AI handles millions of voice interactions daily for some of India’s largest banking, insurance, and automotive companies. CoRover AI provides voice bots in 14 Indian languages for the state-run railway corporation and a regional police department. Meanwhile, Haloocom Technologies uses a five-language voice bot to manage customer service inquiries and assist in screening job applicants.
According to Ankush Sabharwal, co-founder and CEO of CoRover, “The world went from digital first to mobile first to AI first, but voice is the most intuitive way to use technology.” Their voice bot, Ask Disha, was recently launched for India’s train booking entity, IRCTC, enabling users to book train tickets and process payments solely through voice commands. Sabharwal emphasized the necessity for AI agents capable of executing tasks, beyond merely providing information.
Client Engagement and Performance
Gnani AI has developed a voice bot that assists lenders in identifying client financial needs, gathering personal information, and assessing loan eligibility. Additionally, they collaborate with Tata Motors Ltd., one of India’s leading automobile manufacturers, to collect feedback on new car models and promote extended warranties and accessories.
Sarvam AI offers voice bots that can manage mixed-language dialogues and facilitate actions such as appointment scheduling and payment processing. Currently, they have around 50 clients, including the popular devotional app, Sri Mandir, which boasts over 10 million downloads on the Android Play Store. Vivek Raghavan, co-founder of Sarvam, noted, “Try throwing GPT-4 or Claude at Sri Mandir. I can guarantee it won’t work,” highlighting the inadequacy of existing US-based AI models in processing regional dialects and accents.
Market Perspectives and Global Implications
While major US AI firms like OpenAI have created technology capable of generating realistic voice outputs, they have been slow to introduce these innovations to the market. Recently, OpenAI cautioned that users might develop emotional dependencies on its voice products and have indicated measures to prevent impersonation and unauthorized audio replication. Following a delay, the startup has begun to rollout new voice capabilities to a select group of users.
Conclusion and Future Aspirations
Nevertheless, Indian AI startups remain optimistic regarding the future of the technology. Ganesh Gopalan, co-founder and CEO of Gnani, stated, “AI made for specific use cases, languages, and audiences is more accurate, less expensive to run, and has vastly reduced hallucinations,” referring to the phenomenon where AI systems fabricate information. While their primary focus remains on the Indian market, various startups are also looking towards international opportunities, with aspirations to expand into regions like the Middle East and Japan. Notably, Gnani has already deployed voice bots in California, assisting a major Harley-Davidson leasing company in connecting with Spanish-speaking clients.
