By Fred Itua
AT the sidelines of the Grand Challenges
Annual Meeting in Dakar Senegal, been
hosted by the Bill & Melinda Gates Foundation
(BMGF), Daily Sun spoke with Olubayo
Adekanmbi, scientist, innovator and founder
of Data Science Nigeria on his new start-up
which is helping rural women in Nigeria
understand financial literacy
Congratulations on your Grand
Challenges Award, your proposal
is using Artificial Intelligence for
Financial Inclusion especially for
women. What inspired this idea
for you and your team?
As you know, Artificial Intelligence (AI)
makes it faster and easier to identify and
understand patterns and helps simplify
complex everyday process and things around
us. AI is particularly helpful for addressing
financial inclusion issues, which are critical
for building an equitable society. As our world
increasingly relies on digital transactions, the
lack of access to digital financial services stops
many people from enjoying the benefits of
modern living. Our solution is built for the
segment we call the working poor with limited
literacy levels.
What we have built is a solution that allows
people, especially women, who cannot read
or write to track, plan, document, report
and learn about their daily incomes and
expenditures patterns just by talking to their
phones. Then they wait for a voice-based
response in their local language. The solution
can also tell them about the loans or financial
products that are most appropriate for them
based on how much they earn and how they
spend money. The personal voice chatbot is
like recording a voice-note on the phone and
getting the right answers to their financial
enquiries with ease and simplicity.
The solution gives the required convenience
and privacy, and helps users learn more about
their financial position quickly and easily. In
essence, this solution supports transaction
reporting, simplifies accounting, provides
customized financial advice, and guides
daily financial decision-making for informal
businesses, traders, and smallholding farmers.
It bridges the gaps in daily financial decision-
making, financial literacy, and daily transaction
tracking.
You mentioned this is
specifically designed for small
holder women farmers and
women in the rural areas, why this
particular demography?
Enabling financial inclusion is crucial for
women. Recent statistics show that at least
70% of rural women in Nigeria are financially
excluded. A report by the Enhancing Financial
Innovation and Access (EFInA) organisation
states that compared to men, women generally
lack knowledge of the financial options
available to them. The report’s summary
mentions that only 25% of the women in
Nigeria can spend, save, and manage risks in
ways that allow them to be financially healthy,
resilient and seize opportunities. These women
do not have bank accounts, they cannot save
money in banks or get access to loans.
When rural women seek financial options
without right guidance, they are led into scams
and predatory loans because they have minimal
financial literacy. These are serious concerns.
This is why we have leveraged the immense
advantages of Artificial Intelligence to create
a solution that supports women’s economic
empowerment in ways that improve household
earnings, access to financial systems, and
access to guaranteed opportunities to diversify
their income portfolio.
These are what this solution is all
about— empowering women and closing
the gender gap, which is enormous. As you
know so well, women are disproportionately
underrepresented and often work in the
informal economy, which means they have fewer social safeguards – no pension, no
end-of-year bonus or all those guaranteed
income. By designing a solution that follows
the Agenda for Sustainable Development,
particularly Goals 1, 2, 3, 5 and 8, these efforts
will support women and help them become
financially empowered, and help to end
poverty, reduce hunger and inequalities. We
are excited that by tackling financial illiteracy
and inadequate financial planning head-on,
this solution can help break the cycle of
poverty and foster a more inclusive economic
future for all Nigerians, especially women,
enabling everyone to make informed financial
decisions and take control of their economic
destinies.
Now, how big is the problem
of the exclusion of women in the
informal sector in the economy
and what results are you hoping
for through your AI solution?
The exclusion of women from digital
financial services has massive consequences
for our shared well-being as Nigerians. The
truth is that when you enable a woman
financially, you literally help a family and
thereby build a nation. When women are
financially empowered, economies grow.
When more women have chances to
effectively manage their financial realities,
they contribute to the nation’s productivity and
economic diversification. It is estimated that
Nigeria’s gross domestic product could grow
by 23% (US $229 billion) by 2025 if women
participate in the economy to the same extent
as men (CFR, 2022). Unfortunately, this is not
the case. For example, in sub-Saharan Africa,
only 37% of women have a bank account
compared to 48% of men, and the gap has
widened over the past several years. Even with
the pervasive presence of mobile devices and
the drive towards mobile financial access, the
fundamental issue is beyond opening a bank
account with a phone number; it is about
empowering women so they are in control of
their financial well-being, even with the little
they earn. This is why building an AI system
that communicates in the local language is
even more critical.
For us at Data Science Nigeria, we believe
we need to do much more to include women
in the financial system. Inclusion and equity
must be addressed intentionally and frontally.
The truth remains that when a rural woman
wants to open a bank account, seek a loan,
or plan a long-term investment, she does not
have the basic financially literacy to know
what the best deals are. When she visits an
ATM or opens a financial app, she should be able to use her local language and get all
the facts she needs in a manner that relates
to her reality. When she receives a 10-page
document to thumbprint for a loan, she needs
to know and understand the details of the
financial commitment she is making in her
local language. This is why we are using the
power of Artificial Intelligence to eliminate
these complexities in the processes that have
excluded many women from the financial
sector.
What data is behind this
invention and what is the
likelihood of acceptance by the
women especially small holder
farmers?
The informal women traders, small farmers
and self-employed artisans in sub-Saharan
Africa, particularly Nigeria, represent the
largest segment of people living below the
poverty level. Their continued poverty
has been directly associated with financial
illiteracy, poor tracking of daily financial
transactions, the absence of self-moderated
financial reporting or a simple accounting
system, and poor financial decision-making.
Statistics show that one out of two Nigerian
adults do not use any formal, regulated
financial services, and up to 133 million
people are below the poverty line. About 31%
of the nation’s population cannot read and
write, and only 3% of Nigerians have access
to loan services. These are worrisome data
that clearly show we need to seek simpler and
more nuanced channels for making financial
services available in the users’ local languages.
To encourage acceptance of our solution,
we have designed a simple interface. It is
voice-driven as much as possible, and users do
not need a smartphone. We are also working
on a back-end integration so users can phone
a number for an automated response. This
means the solution will work even where there
is no internet service.
Also, we are working with women’s groups
to help them understand the solution and
how they can use it. We are also learning from
them about additional features we could add
to support their everyday financial well-being,
focusing on ease, convenience, and privacy
without compromising their personal data. We
have built an anonymous and de-indemnified
interface so that every recommendation and
the data input are not traceable to anyone.
Have you done any test runs,
if you have, what was the
experience like ?
The test data is very promising. It shows
that about 93% of the testers confirmed the
accuracy of the financial explanations provided
by the AI-powered chatbot, which was higher
than that of an average person. It was at par
with a financial expert’s understanding. In the
same vein, 80% of the testers confirmed the
high level of convenience and privacy from
the AI-generated financial advice compared
to working with a human. Also, 90% of the
testers confirmed an increased interest in
digital financial access because of the AI-
enabled financial advice and accounting
support.
Two key differentiators of our platform
are its abilities to read and explain financial
documents. A user can upload their statement
of accounts, or a loan agreement and our app
solution will explain the details to them. In
addition, the solution is based on fine-tuned
large language models. We used ChatGPT,
which has been fine-tuned with a lot of
nuanced and Nigerian-specific financial
information and consumer slang. For example,
if a user tells the app that it seeks a loan of
‘100 Figo’, the app knows that the user is not
speaking about the great Portuguese football
player but ₦500 in 100 places. This is street
language, and understanding it is a central to
our solution.
How soon should we expect the
deployment of this innovation and
how many languages are you
planning to develop?
At the moment, we have piloted our app
in the Yoruba language, and we hope to scale
it by building more Yoruba language audio
datasets that our large language model will
learn from. Our approach is to be as inclusive
as possible. We have recently opened a
language recording lab, which allows us to
build the large language voice data required
to train the AI to speak and understand
more Nigerian languages. However, that
takes time and it is a huge investment. We
are building the “Siri” of Africa that can
speak, understand, and engage in at least
the 10 most popular African languages. One
additional commitment we are making is
to build rich audio datasets that recognize
dialect differences in local languages. That
will enable the solution to serve as many users
as possible in villages and local communities,
since helping them is our principal goal. For
example, when our system is fully completed,
it will be able to support Yoruba spoken in
Abeokuta (Egba) or Ijebu, Ekiti, or even the
variations in Togo or the Benin Republic.
Tell us about your organization
and what you have done in
innovation in the past
Data Science Nigeria (DSN) is sub-
Saharan Africa’s premier artificial intelligence
technology enterprise. It is dedicated to
fostering Africa’s AI talent ecosystem and
pioneering cutting-edge solutions in health,
retail and finance. DSN’s proven track record
includes successfully developing major AI
solutions in the financial services sector,
such as SpotOn, an integrated addressing,
marketplace and mobile wallet platform for
small businesses. In 2022, the prestigious
International Research Centre on Artificial
Intelligence (IRCAI) recognised DSN’s AI-
powered SMS Fraud Detector and Instant
Fraud Prevention Call Alert as one of the
Top 100 Global AI Solutions. In 2021, the
organization also listed two of our products
as global Top 100 AI Products. This wase in
addition to earning several academic awards.
Our start-up, DSNai, focuses on health,
education, financial inclusion and agriculture
for emerging market customers. DSNai is
also a global winner for using AI to generate
automated and personalized learning content
based on learners’ nuances, style and life
development stages.
We support many global organizations
including the Bill & Melinda Gates
Foundation, Mastercard Foundation, World
Bank, Microsoft, Google, Meta (Facebook),
Tony Blair Institute and TikTok, among others.