Changing The Voice Of Indian Businesses
In the past decade, India has transformed into one of the fastest-growing digital markets in the world. However, the majority of the content online is centered around the urban dweller, who lives in India’s metro cities, has considerable disposable income, and, most importantly, speaks English.
Yet, Google’s ‘Year in Search — 2018’ report points out that more than 65% of searches are coming from Tier 2 towns. Local language searches have grown 10 fold, demonstrating that increasingly, a large portion of India’s digital populace would not want English as their preferred language of communication.
The 2011 Census estimates that nearly 44% of Indians consider Hindi their native language. Bengali, Marathi, Tamil, Telugu, and Gujarati closely follow. And tellingly, of the 25 million English speakers in India, less than 0.3 million speak it as their first language.
These crucial facts essentially demonstrate that vernacular is really a sign of the times and businesses that don’t focus on vernacular would be potentially losing out on revenue.
More services, from banking to healthcare, are becoming digitised, and more people are being folded into the social and economic fabric than ever before. But for these services to reach the length and breadth of their target audience, they need to turn to local languages. Most tellingly, they need to rely on voice.
There’s a gap. And a gap that could be a chasm. Would vernacular and voice be the defining factors bridging this chasm? We, at Vernacular.ai, certainly believe so.
Growing local with voice
According to Rajan Anandan, who was formerly the MD of Google India, voice search queries in India are growing at 270% per year, which is a staggering number.
While the current generation is a mobile-first generation, the future will be a voice-first generation.
An Accenture study showed that India tops a list of 22 countries with 72% adoption of voice assistants.
We are now looking at a practically keyboard-less, more voluble, and language-integrated future.
Connecting voice and vernacular
Currently, many smartphone-based apps assume that Indian users are able to read English and avail of all the services offered through an app. That’s quite an erroneous assumption. When we were testing our product, we met a farmer who told us that he couldn’t read any of the messages on his smartphone, including notifications from banks because they were all in English! He couldn’t even understand if the message was a critical one or just a promotional SMS! He had to travel 10km on his bicycle just to get the messages translated from the bank manager, losing his day’s salary. For people like him, the world is built in English, and he is not part of it. It’s for people like him that businesses need to put in efforts to communicate in his language of choice and help make India an inclusive digital world.
Helping businesses engage with their multilingual customers
Vernacular.Ai is a smart AI and NLP-based platform that aims to bridge the existing gap between the local-language speaking population and businesses who want to reach multilingual consumers. Our products and solutions use voice as a medium to help businesses engage with their multilingual and English-speaking customers at different stages of their life-cycle. We offer two major products for improving customer experience:
- Call center automation
- Voice assistant for apps
How businesses benefit: By adding support for voice-based interactions, businesses can improve customer service and satisfaction by facilitating access to all their services for people who use vernacular languages.
More importantly, businesses can converse in the language of the customer’s choice. Appears simple enough yet there is an additional layer of complexity! India has 22 major languages and 720 recognised dialects! Achieving a high level of accuracy given the complexity of this scenario might seem impossible.
Consider a typical call center. Our research shows that 70% of call center queries are repetitive in nature, which does not require human intervention each time. Again, Vernacular.ai solves this crucial problem by automating repetitive queries for multiple customers across 10 languages using its Vernacular Intelligent Virtual Agent (VIVA).
The solution uses cutting-edge speech recognition and natural language processing technologies allowing for better understanding of the caller’s query and resolving the query through engaging conversations.
Customers have reported higher CSAT (Customer Satisfaction) scores in automated call centers that use VIVA along with human agents compared to call centers with only human agents.
Future ready: By placing itself at the intersection of vernacular and voice, Vernacular.AI is poised to become a vernacular engine for India’s native languages, especially for the next 900 million people who are going to come online.
India’s growth in the use of vernacular in the digital space will mirror more that of China than the US. China has never quite embraced the English language and a billion Chinese use push-to-talk features on smartphone apps rather than type out Chinese on their smartphones. It’s a scenario that might well be replicated in India.
When we had the vision for Vernacular.ai, we realised that there was no vernacular data there to start with. For building a vernacular and voice product, you need data with multiple languages and within that different dialects and accents. You need thousands of hours of annotated data for that. And not just the acoustic data but also the language model, past conversation history, and intent level training data. The accuracy level in each of these factors is important and we are proud to say that we have achieved accuracy levels of 90–95%.
Vernacular.ai resembles a common room where more than a billion people can converse and convey their needs through multiple local languages enabling them to easily access everything that the new age digital world has to offer. And that’s the holistic solution that Vernacular.ai is building.
