India’s Voice Opportunity, And The Best One

To be fair, this is hardly the first attempt at creating an Indian language-driven internet experience. The first wave of Indian language internet relied heavily on the vernacularization of English language content and didn’t focus much on context, says Vipin Agarwal, senior director of Fosun Capital.

“There was no interactivity for the user,” he says. A second attempt was made by payment startups like Paytm and Mobikwik, e-commerce firms like Snapdeal, Quikr, Olx, MakeMyTrip who translated their English text interface into regional languages. Dailyhunt, a news aggregator supports news items across 14 regional languages, and social media plays like Sharechat and Kwai use a mix of regional video and text content to attract more regional language users.

Deep Challenges Ahead

For e-commerce apps, however, the access challenge runs far deeper. Even if the text is accessible in a regional language, navigating within apps through a long chain of commands, with stodgy indic characters, makes for an off-putting experience.

This is where stakeholders in the Indian internet ecosystem—developers, internet businesses, device manufacturers—believe that voice as an interface could play a huge role. They imagine Indians speaking in their own regional tongues, some even in their own regional dialects, to browse, learn, discover content and shop online.

“Indian language keypads are very limiting. There is huge potential for voice-based technology to provide for more than just entertainment. Users will also look for utility from these voice-based services,” says Karthik Reddy, managing partner at Blume Venture Partners. Blume was an investor in Vokal, a voice-based, peer-to-peer knowledge sharing app. Agarwal echoes this sentiment. “Voice will surpass text and touch in the next two years,” he says.


The global usage of voice assistants is seeing a slight decline as sales of AI speakers slow down amid consumer confusion over their utility. According to the start-ups and investors The Ken spoke to, though, India might be bucking that trend.

This is because, as the primary interface for internet navigation, the utility would be obvious to a large number of vernacular users. Rangarajan claims that for mobile-first regional language users of the Internet, voice is not just an efficiency hack. “It’s the most natural way to communicate with a machine,” he states.

E-commerce companies, hyperlocal delivery startups, and online ticketing platforms all stand to benefit from the users that voice-based, multilingual interfaces will unlock. However, this will not be cheap or easy.

To be successful, they must create a solid voice interface—with the ability to absorb the diversity of 22 official Indian languages, spoken in over 1,000 dialects. Creating this massive ‘Bharat use case’ for a technology that’s still just getting off the ground hasn’t really worked out for companies in the past.

Flexibility Of The System

Slang Labs, though, believes that it has the answer—customisability.

“If the voice opportunity has to work out for India, it has to look beyond generic search engines”, says Reddy. Slang Labs seems to see things similarly, which is why Rangarajan and his eight-member team aren’t looking to just build another voice assistant a la Siri, Google Assistant, and Alexa. Instead, they’re looking to provide developers with the building blocks necessary to customize a voice interaction layer for third-party apps.

With the application programming interfaces (APIs) that Slang Labs provides, independent app developers and internet companies can build app-level voice controls to help users browse an app through a mix of voice and touch commands.“We have launched a self-serve platform where anybody can sign up to our account and configure their use cases by registering their app,” says Rangarajan.

Most modern mobile apps are optimized for touch-based interaction, where the homepage lists the most important content and controls. However, things get complex when it comes to executing more specific tasks. The experience becomes even more discomfiting for vernacular users as apps aren’t properly customized to regional languages.

This is where Slang Labs enters. “We will come in where a user knows what she wants. But to get those products in, for instance, a Bigbasket app, you have to scroll and navigate several times. With voice, the user can speak to the app and add groceries to the cart,” Rangarajan elaborates.

The voice API tool will also work for users who want to browse for an item. “This takes a lot of filtering, but the user can talk to the app and discover, for instance, a pair of jeans between size 34 and 36, or just blue jeans from a specific brand,” he adds.