Natalia Lyarskaya: How ZestMoney is Using AI and Machine Learning to Reach New Customers in India

ZestMoney Chief Data Officer Natalia Lyarskaya explains how
cutting-edge technology is helping consumers in India access credit
where it was previously unavailable

The appetite for credit is growing in India. However, compared to
developed credit markets like the US, India is underserved. There
are only three credit cards per 100 people in India, compared
with 32 per 100 in the US.


This may be starting to change. In this week’s episode of the
Business of Data Podcast, ZestMoney Chief Data Officer Natalia
Lyarskaya explains how ZestMoney is using AI and machine learning
technology to create a transparent and trustworthy credit
solution where traditional banks have been unwilling or unable to
do so.


“There is a kind of chicken and egg problem where someone needs
access to the credit products but has never been in the financial
sector before and other banks, traditional banks, cannot evaluate
their creditworthiness,” Lyarskaya says.


She continues: “We believe that using data and technology, we can
build this affordable, transparent, financial product for the
Indian people that can be used by everyone and can increase also
the trustworthy population in this new credit segment.”
Evaluating customers using data, machine learning and AI

ZestMoney has created a 100% digital user experience that uses an
array of data coupled with machine learning and AI technologies
to evaluate new credit lines in a matter of milliseconds.


“Based on the AB testing that we've done we have collected quite
a good amount of data,” Lyarskaya says. “[We built] some
predictive models that allow us to differentiate between
different groups of users, so we can propose different journeys
and different options for users to apply for our product.”


While the technology behind ZestMoney’s model evaluates new
credit applications and makes the final decision on credit
approval, it also guides the user on a personalized journey
assessing and modifying questions during the application process
based on personal and historical data.


“This [model] is basically behind every decision that we take
along the journey,” she explains. “Like, what kind of questions
we want to ask a user, or do we want to ask this question in one
way or the other?”.


She concludes: “There is a model that stands behind that tells us
what exactly we need to do and who is the user that we see in
front of us. So that is all based, not just on our assumptions,
but on what the data has been telling us.”
Key Takeaways


Machine learning and AI are helping financial firms
reach new credit markets in India. Where traditional
banks have been slow to react, tech upstarts have been able to
capitalize.


Balance privacy and personalization for a better user
experience. Understanding how much data an individual
feels comfortable sharing is an important first step to
creating outstanding user experiences.


AI and machine learning solutions enable better
products but do not create them. Human critical
thinking is needed to make sure a system works. AI and machine
learning make sure that the system is efficient.