Local Language Technologies — Tipping Point for Next Billion Indians?
The Ankur Capital team caught up with Girish Varadarajan (Head Customer Analytics, Aditya Birla Capital), Dr. Manuj Garg (Co-Founder, myUpchar), Anirudh Maitra (Product Manager, Google Assistant for Next Billion Users) and Kushagra Sinha, (Founder, Jiny, an Ankur Capital Portfolio Company) to discuss the tipping point for digital inclusion in India.
While Anirudh was able to provide a wider lens on the subject by leveraging his experience at Google and Girish and Manuj were able to provide deep sectoral insights from a healthcare and financial services perspective, Kushagra gave us first-hand insights from his experience with businesses looking to adopt tech solutions to target the next billion Indians.
We kicked off the discussion with each of the panelists giving their take on how they understand and classify the target segment. Anirudh spoke of how India has the complexity of twenty different countries and that the first aspect is understanding the market and the market opportunity. For this, he uses a socio-economic classification resulting in creation of three different cohorts:
India 1- 170–200 million users; monthly household income of more than USD 500 per month; 70–80% internet penetration; digital journey involves mainly use of English language
India 2- 200M users. monthly household income is between USD 90–500; 40% internet penetration; little or no use of English language; Use low to medium end smartphone
India 3- 100M users; monthly household income is USD 50–90 monthly; a large percentage of people in this segment use Jio feature phones which provide high speed connectivity and access to apps; 25% internet penetration
From the perspective of building the product, he added that optimizing user experience in the first day and first week is critical. It is important to analyse the device model and language settings. Once people start using the product, the learnings from these use cases can be leveraged to roll out customized content.
Kushagra offered a different take. At Jiny, while helping businesses make their product Bharat-friendly, the segmentation of users is primarily based on where the users lie on the digital adoption curve. The classification is based on capability and not on income-levels. For instance, while the people in India 1 may have a high purchasing capacity, they might not rank as highly when it comes to digital literacy. Several urban users, though equipped with feature-filled smartphones, use only Whatsapp. While even in Tier II cities where people have less income, there are several users who are relatively more digital savvy! Another relevant parameter is the medium of education and ends up playing a significant role in segmenting users.
Manuj pitched in by telling us that in his experience at myUpchar, the classification was based on language. A barrier and challenge they wanted to tackle at myUpchar was to reach non-English users? The second factor that help identify the relevant target was to segment the users based on engagement and measuring the amount of time users spent on the platforms and whether they transacted using the platform.
Girish mentioned that while socio-economic classification is a relevant parameter, they look at product, geography and engagement. At Aditya Birla Capital, while formulating strategies to ensure that financial services reach the next billion Indians, an important learning has been that customers have to be spoken to and onboarded in a manner that builds trust and makes them understand the financial products and potential upside.
Our key takeaway was that when it comes to classification of users to identify the target market, a one-size-fits-all strategy cannot be adopted and as our panelists discussed, the segments can be based on socio-economic classification, language, geography or even tracking product usage and engagement.
We subsequently discussed some of the key challenges that our panelists thought were the biggest hurdles for greater tech adoption within the next billion. Manuj highlighted how the existing natural language processing tools are not built to tackle non-standard spellings or words that are often by-products of the amalgamation of two or more languages. On a similar note, Anirudh mentioned that presently the web and app ecosystem is not customized and built keeping in mind the next billion users. He added that decoupling device language and app language is important since people want to use their phone in English but the mobile apps in local languages.
Kushagra felt that technology is not the only problem since the technology has the capability to tackles several of these issues but just that the developers have not provisioned for it. It is important that the technology, which is merely an enabling tool, is built with the objective of solving real needs of the people.
“There is no point of building a better and more real AI tool if there is no understanding of the pain points of the end consumer”
We need to adopt a more holistic approach and solve for the larger issue of improving digital literacy!
Girish echoed Kushagra’s view by emphasizing on the importance of getting a deeper connect with the users — this translates to better trust and in-turn, business!
In terms of how our panelists are tackling this issue, Manuj spoke of how at myUpchar, which has seen a recent spike in the number of users on its platform, in order to build trust and connect with their users, they still provide call support to users who come on the platform but are unable to transact due to limited understanding of the tech and UI.
Our panelists also offered some interesting takes on what they felt the tipping point is / will be for greater digital inclusion. Girish re-iterated how trust is paramount, especially in the financial services sector where transactions aren’t small-ticket purchases but in fact, people parting with their life-savings or opting for a life-insurance policy, etc. The question that needs to be answered is how do we use AI to bridge this gap and ensure that the “AI has a brain and heart”.
Kushagra reminded us of the uphill task we have on our hands and that multiple aspects need to be tackled in parallel — a medium that can be exploited further is voice. Audio is a great enabler and can solve the problem of comprehension effectively. A humanized assisted experience is required to reduce the cognitive load.
Closing out the discussion, Anirudh aptly summarized what must happen for us to remove the tech barrier and reach the next billion users. Three fundamental aspects must come together — 1) Economic growth, in China, the tech revolution was preceded by a spurt in socio-economic growth, whereas in India, we have seen a reverse cycle where digital adoption has bypassed economic prosperity; 2) Technology, while we may have reached the inflection point for English language such that AI-based voice tools are better than humans in terms of comprehension, we are yet to achieve the same for other languages — having said that, we aren’t far from that as well; 3) Business model innovation, all users are aspirational and we must look at building products and adopting business models that will leapfrog segments.
All these three things must come together and a good starting point could be the ‘Sachetisation’ of services and products, allowing next billion users to test tools and the technology can then be retrofitted.
In case you were not able to attend the session, you can catch it on YouTube or listen to the podcast here.
