You are looking for some cool designer shades in metallic blue, so you quickly fire-up your favourite e-commerce app, search for it, add it to your cart and buy it. The difference—you do this without touching your phone, and by speaking to it in Gujarati.
Over the last decade the mobile phone has come to define the digital experience for billions of new consumers. That may be changing. Many voice artificial intelligence (AI) startups are looking to make these kinds of “post-mobile" experiences a reality for their customers globally and in India. With over 1,038 AI startups in India, working on various aspects of AI, speech and context recognition is fast becoming an exciting new frontier for technology. The future of voice AI seems promising, with adoption slated to increase as voice-enabled devices embed themselves in our lives in areas such as smart connected homes, navigation, smart offices and customer support centres. Globally 50% of searches are expected to be voice-based by 2020, with a significant number of them driven through voice-enabled devices like smartphones and smart speakers.
We see five important trends that describe the potential trajectory of use of voice technology in India:
1. Vernacular languages will be increasingly supported as players expand to tier II and tier III cities: Industries including e-commerce, retail and financial services are all looking to tap into the next wave of growth as they look to penetrate tier II cities and beyond. Many of the users in these cities are first-time digital users and will need more evolved user interfaces that are in tune with their experience levels. Vernacular language support will become important to bridge the usability gap and we can expect to see several existing and new voice AI startups support languages beyond English. A key determinant of their success would be measured by how they are able to adapt to the wide variety of dialects and accents that are prevalent across India.
2. Narrow use cases will be the gateway to more widespread voice AI adoption: We believe that voice AI will first become prevalent in more narrow use cases and as users become more familiar with voice AI-based interfaces, voice AI adoption will explode over the next three-four years. We are already seeing narrow use case adoption in customer support in the banking, financial services and insurance space and in product discovery in the e-tailing space.
3. Cracking the data problem will be the key for success: Today Indian voice AI platforms are faced with the unique problem of not having enough data to be able to train their AI platforms for recognizing vernacular languages. This manifests itself in a poor user experience that leads to slower growth. Voice AI startups that manage to solve this data problem will be able to steal a march over their competitors. We believe that non-traditional sources such as leveraging media content platforms might play a role in solving this problem.
4. Big Tech platforms will start focusing on India in the quest for the next billion users: Indian voice AI startups are either building their own proprietary AI platforms to support Indic languages or are relying on voice AI services provided by Big Tech firms. Support for Indic languages has so far been low on these platforms, but as these players recognize the importance of voice in tapping the next billion users globally, they will increasingly start rolling out support for local languages. We are already seeing AI support for Hindi, Tamil, Marathi, Kannada, Bengali, Telugu, Gujarati.
5. Security will become a focus as firms begin to experiment with banking and voice payments: As e-commerce and financial services emerge as key adopters of AI and voice AI in India, security will become important as voice AI firms seek to also implement payment interfaces. In the absence of a clear regulatory framework to allow voice-based authentication for payments, startups will be challenged to come up with new secure means of integrating with existing payment methods that are secure.
As these trends play out in the coming years, we believe that voice AI will emerge as an important tool for designing and deploying user interfaces of the future. As voice AI support for vernacular languages improves, it will serve as a catalyst to help bring digital experiences to the long tail of consumers in India.
Mahesh Makhija is leader (emerging technologies and digital) at EY India.