Accelerating Financial Inclusion With New Data

Accelerating Financial Inclusion with New Data
A joint report from the Center for Financial Inclusion at Accion and the
Institute of International Finance
MAY 2018
PART FOUR OF THE MAINSTREAMING FINANCIAL INCLUSION: BEST

PRACTICES SERIES
Accelerating Financial Inclusion with New Data 1

Accelerating Financial Inclusion with New Data
authors
Ethan Loufield, Director of Strategy and Operations, CFI
Dennis Ferenzy, Associate Economist, IIF
Tess Johnson, Research Associate, CFI
project directors
Elisabeth Rhyne, Managing Director, CFI
Kristen Silverberg, Managing Director, IIF
Conan French, Senior Advisor for Innovations, IIF
FOREWORD BY
Islam Zekry, Chief Data Officer, CIB Egypt
designer
Mounika Ramaji, Program Assistant Digital Design, IIF
editor
Virginia Moore, Director of Communications, CFI

CONTENTS
Foreword..................................................................................................................................................4
Acknowledgments....................................................................................................................................5
Introduction..............................................................................................................................................6
State of Affairs: How Banks and Fintechs are Using Non-Traditional Data and Tools..............................7
Getting From Here to There with Data................................................................................................8
New Types of Data: Pros and Cons....................................................................................................11
Products and Use Cases.....................................................................................................................13
Snapshot 1: WeBank Uses Tencent Data for Unsecured Personal Loans in China........................14
Snapshot 2: BBVA Bancomer Uses Blend of External and Internal Innovation in Mexico............14
Snapshot 3: Grameen America Takes a High-Touch Approach to Internal Data Collection in the U.S...14
Snapshot 4: SCB Abacus Spins Off from Parent Bank to Accelerate Advanced Analytics in
Thailand.........................................................................................................................................15
Challenges.............................................................................................................................................15
Internal Preparation............................................................................................................................15
Changing the Banking Culture......................................................................................................16
Attracting Technical Talent............................................................................................................17
Modernizing Legacy IT Systems....................................................................................................17
Data Journey Management................................................................................................................18
Identifying the Most Relevant Data and Where to Begin.............................................................18
Slow and Arduous Value Extraction..............................................................................................18
Getting Data Privacy and Consumer Consent Right.....................................................................18
Data Fragmentation and Willingness of External Parties to Share..............................................18
Incomplete Digital Proliferation....................................................................................................19
Regulatory Environment....................................................................................................................19
Opportunities........................................................................................................................................20
Engagement with Regulators and Policymakers................................................................................20
Partnerships........................................................................................................................................21
Emerging Technologies.......................................................................................................................23
The Competitive Landscape Moving Forward......................................................................................25
Appendix A: Study Methodology..........................................................................................................27
Appendix B: Interviews.........................................................................................................................28
References.............................................................................................................................................29

FOREWORD
For the financial institutions that aim to serve the 1.7 billion people globally who lack access to
formal financial services, new data tools offer an unprecedented opportunity. The ability to capture
and analyze vast amounts of data supports real-time, in-depth analytics to improve credit models,
identify new customers, understand the products they need, and create a seamless customer
journey. These new capabilities and improved economics help financial institutions extend quality
financial services to unbanked and underbanked populations, deepen financial inclusion, and
accelerate economic growth.

One country that stands to benefit significantly from advances in data analytics is Egypt, where
two-thirds of the population lacks an account at a financial institution and the informal economy
is estimated to be around USD 90 billion. Currently, borrowers without a financial identity are
automatically locked out of the credit market, and individuals with limited financial data are
categorized as “higher default risk clients” and typically charged higher interest rates or have
greater collateral requirements. Traditional credit scoring methodologies, which exclude most of
the potential customer base, need change.
Commercial International Bank (CIB) Egypt recognizes the prospects of data analytics for financial
inclusion. Converting non-traditional data into credit insights that help banks assess an individual’s
willingness and ability to pay will become the cornerstone for an alternative credit scoring system,
enabling smart lending. CIB’s newly established Analytics and Data Management department is
moving in this direction—our bank is leveraging descriptive and predictive analytics and historical
behavioral trends, while ensuring that all opportunities pursued make financial sense. The ultimate
goal for CIB is to increase profitability, enhance our ability to serve various customer segments,
and quickly tap into the lucrative opportunities that the emerging Egyptian economy offers.
I welcome this report by the Center for Financial Inclusion at Accion and the Institute of International
Finance, and applaud the efforts of both organizations in encouraging the mainstream financial
industry to safely and responsibly leverage data for positive customer outcomes. As digitization
and new data tools continue to enhance our ability to analyze alternative data sources for the
unbanked and identify viable opportunities, our collective efforts will bring about further advances
in financial inclusion.
One can barely begin to imagine the list of potential opportunities that new data holds for the
future. Expanding access to credit, fueling growth in developing economies, increasing the tax
base, optimizing subsidy disbursements, and much, much more are closer to our reach than ever
before.
Islam Zekry, Chief Data Officer, Commercial International Bank (Egypt) S.A.E.

ACKNOWLEDGMENTS
We thank all those who participated in our research—and particularly the executives from 18
financial institutions, fintechs, and other entities who generously shared their insights. We gratefully
acknowledge the following organizations: Artoo, Bancolombia, BBVA Bancomer, Caribou Digital,
Commercial International Bank (Egypt) S.A.E., CIMB Group, DemystData, Destacame, Grameen
America, Kabbage, Konfio, LenddoEFL, Malayan Banking Berhad (Maybank), SCB Abacus,
SMEcorner, Standard Bank, Uganda Communications Commission, and WeBank. Please see the
full list of executives interviewed in Appendix B.
We note with gratitude the insights and guidance of the Research Advisory Group for the
project: Jonathan Hakim, CEO, Cignifi; Peter Cureton, Americas Head of Change Management
for Technology Security Services, Credit Suisse; Anju Patwardhan, Managing Director, Fintech
Investment Fund, CreditEase; Mark Hookey, Founder and CEO, DemystData; Lory Camba
Opem, Program Manager and Lead Specialist, Global Responsible Finance, International Finance
Corporation; Winn Martin, Head of Data Strategy, Kabbage; Mah Kam Lin, Group Data Scientist,
Maybank; Monica Brand Engel, Partner, Quona Capital; and Abdul Musoke, Manager, Economic
Regulation, Uganda Communications Commission. Their support and thoughtful insights have
been essential to this project.
This report is part of a two-year initiative, Mainstreaming Financial Inclusion: Best Practices,
which aims to help advance efforts of financial institutions to reach customers at the base of
the economic pyramid. It is the fourth of six reports in this series. The initiative and this report
were made possible by generous financial support from MetLife Foundation. CFI is also grateful
for generous support from its founding partner, Credit Suisse. However, the views and opinions
expressed in the report are those of the authors and do not reflect the views and opinions of the
interviewees, the advisory group members, or MetLife Foundation. All errors are our own.

INTRODUCTION
As financial services increasingly go digital, finding innovative ways to leverage new kinds of data
and tools is fast becoming a necessity for financial service providers, both to remain competitive
in traditional markets and to tap into vast new markets.
This report, which is based on interviews with banks, fintechs, and other actors,i examines how
new types of data and analytics tools are being used in the financial sector to reach underserved
markets. Our aim is to take stock of how the data landscape is evolving, to describe new types of
data tools, and to assess how firms are innovating around data, and where they have experienced
setbacks. As this report will discuss, there are many internal and external challenges that
providers must address for the promise of new data to be fully realized. From getting the right
culture and technical talent in place to engaging with regulators and partners on data sharing
and management, there are numerous obstacles to negotiate in capitalizing on the explosion of
new data to reach underserved markets with new commercially viable solutions. This is not an
attempt to provide a “how-to” guide for navigating the various stages of the data value chain.
There are too many contingencies to prescribe a “one-size-fits-all” solution. Instead, it is an effort
to shed light on the potential of data to improve financial services for the underserved.
While underserved markets may not be as lucrative as more developed markets today, they
represent a great opportunity for mainstream financial institutions: they are unsaturated and
likely to represent a much larger share of growth in the years ahead.
But along with the influx of new types of data and tools also come new kinds of competitive
threats. The traditional boundaries separating players in adjacent industries are blurring, and
the future market for financial services is up for grabs. Mainstream providers must therefore
accelerate their efforts to confront these challenges and future-proof themselves for what the
data revolution has in store.i
i The scope of this paper prevents us from exploring how data is being leveraged by insurers. For more information on the topic,
please see Inclusive Insurance: Closing the Protection Gap for Emerging Customers and Innovation in Insurance: How Technology
is Changing the Industry.

STATE OF AFFAIRS: HOW BANKS AND FINTECHS ARE USING
NON-TRADITIONAL DATA AND TOOLS
FIGURE 1 Over the last decade,
INTERNET-CONNECTED DEVICES
the digital revolution
has intensified with the
proliferation of online
devices (see Figure 1); new
and cheaper solutions to
store, manage, and process
data; a substantial rise in
computing power; a decline
in related costs; and the
ubiquity of connectivity. In
turn, these breakthroughs
have fueled exponential
growth in the amount of data
being generated globally.
According to International
Data Corporation, a market-
intelligence firm, the digital universe is doubling in size roughly every two years. This data
explosion (see Figure 2), combined with a growing capacity to collect, structure, and analyze
data continually and in real time, is fundamentally transforming the financial services landscape
and giving rise to new opportunities for banks, fintechs, and others to reach vast new markets.
FIGURE 2 With over 1.7 billion people still

GROWTH OF GLOBAL DATA
without access to formal financial
services (see Figure 3, pg. 8), one
of the most important implications
of this new data-based economy is
its ability to foster greater financial
inclusion. New forms of alternative
data are playing a key role in
expanding financial services to
lower income customer segments.
Alternative data—including social
media activity, email usage, utility
payments history, mobile phone
records, and psychometric testing—
coupled with advanced data analytics
tools, bring an unprecedented
opportunity to better understand
and serve clients, especially credit-
seeking “thin-file” clients who are
otherwise excluded from the formal
financial sector. A 2015 report by Omidyar Network, a philanthropic investment firm, suggests
that in the world’s six biggest emerging markets alone—China, India, Brazil, Mexico, Indonesia,
and Turkey—big data has the potential to help between 325 million a nd 580 million
people access credit for the first time.1 Moreover, data-driven innovation extends well beyond
credit analytics into applications such as customer segmentation, engagement, and insights;
client onboarding; automation; product design; fraud prevention; identity verification; and the
overall expansion of service offerings to traditionally excluded segments.

FIGURE 3 Getting from Here to
UNBANKED POPULATIONS IN SELECT EMERGING MARKETS
There with Data
In a world increasingly awash in
data, it is challenging to decide
which data to capture and how
to put it to good use. Across
our interviews (see Figure 4),
we observed a stark contrast
between how constrained the
actual use of data is relative to
the vastness of the expanding
universe of data. With both
traditional and non-traditional
data, the practical challenges
of obtaining and converting
raw data into actionable
insights are immense. Pointing
to the lack of open APIs across
banking and other sectors in
Thailand, Natth Bejraburnin of SCB Abacus, a subsidiary of Siam Commercial Bank (SCB), noted
the difficulty of getting third parties to share data and remarked that it could be easier for SCB
to build its own platform for collecting more data internally, rather than relying too heavily on
external partners.iiii
FIGURE 4
PARTICIPATING INSTITUTIONS
However, despite the oceans of data that are out there, in practice, it is difficult for financial
service providers to use more than a few scoops at a time without significant collaboration with
third party data providers, such as utility companies, social media providers, mobile network
operators, and other specialized data vendors. In fact, although progress is being made on this
front, much more needs to be done to build the connections and standards necessary for data
to flow more readily and securely from its point of generation to its final point of utilization. This
is especially true for underserved markets around the world where new sources of data may be
all that is available.
ii Unless otherwise noted, all referenced interviews are listed in Appendix B.

FIGURE 5
DATA JOURNEY

While this report does not provide an in-depth technical assessment of the flow and management
of data, it is useful to take a moment to understand the various steps involved in taking data from
raw form and converting it into valuable insights. With all the new types of data that exist, it has
become that much more important to streamline the process of identifying and connecting to the
growing number of external sources of data. As noted in a recent open letter from the Institute of
International Finance to the U.S. Treasury Department regarding Executive Order 13772 on Core
Principles for Regulating the United States Financial System, “All market participants should be
encouraged to invest in systems, models, and teams that produce and utilize high-quality data,
which in turn enable them to provide customer products and services more economically.”2
Harvesting data is essentially a careful curation process that involves exploring the universe of
data to identify holders of data to connect with, testing that data for usefulness, structuring it for
standardization, and integrating it with internal data and systems for analytics and decisioning.
All of this must also be done in an efficient, secure, and compliant manner (see Figure 5).
Greater collaboration and infrastructure are vital in navigating the increased fragmentation
and complexity inherent in the explosion of data, even for large players, often necessitating
external partnerships. Matt Hennessy of DemystData, a data technology company that helps
financial institutions harness data, underscored the importance of partnerships in helping clients
overcome the general fear that many have about the onslaught of data so they can proceed with
workflow and cost-benefit analyses to identify actionable solutions.
The legal and regulatory environment also contributes to this complexity because of varying
rules pertaining to data privacy, protection, and ownership; know your customer (KYC); anti-
money laundering (AML); personally identifiable information (PII); and data localization (cross-
border data flow) requirements. This increasing complexity not only heightens the need for
specialized expertise in data management and compliance, but also for greater engagement
with, and guidance from, policymakers and regulators. In this regard, the recent implementation
of the General Data Protection Regulation (GDPR) in Europe shows that while new rules can
be cumbersome and involve trade-offs, the silver lining is that they do bring greater regulatory
clarity and direction for providers. As noted in The Economist, given that “many of the firms
preparing for the GDPR’s arrival in Europe enthuse that the law has forced them to put their
data house in order,”3 it would be worth examining whether authorities in emerging and frontier
markets could borrow certain elements of the GDPR to help catalyze collaboration around the
secure and responsible use of new types of data. In many cases, clear regulation, even if not
ideal, is preferable to unclear or continually changing regulation.
One final but fundamental point is the importance of the underlying infrastructure for managing
the flow of data. Just over a decade ago, Amazon Web Services (AWS) came onto the scene
offering cloud computing services, which allow companies to essentially rent access to pooled
computing resources instead of purchasing and maintaining the infrastructure themselves.
Competitors, such as Microsoft, Google, and IBM, later entered the market, helping to drive
down data storage and processing costs and fuel the introduction of new security, compliance,
analytics, application, and other complementary services. In discussing how essential these
cloud services have become, a recent article in Bank Innovation referenced the rapid growth of
Robinhood, a commission-free trading app, noting that just two DevOps professionals at the
startup were able to leverage 18 distinct AWS services to address compliance, fraud detection,
anti-money laundering, and other needs in building out a system at a scale and speed that would
have been impossible without the cloud.4
Clearly, cloud technology has greatly expanded the computing power that is now accessible to
organizations of all sizes. Nearly all of our interviewees, ranging from early stage fintech startups
to large multinational banks, spoke to their use of cloud services. Both Artoo, a provider of
customer relationship management solutions for lenders serving low-income segments in India,
and Destacame, a digital alternative credit scoring platform in Chile and Mexico, use Microsoft

Azure and AWS to store and secure data in the cloud. And Bancolombia, Colombia’s largest
financial institution, has started to use AWS machine-learning tools for its mobile bank product to
scan pictures taken by users at account opening for fraud detection purposes. Bancolombia would
also like to collaborate with competitors in the region to create a centralized image database to
help combat fraud, and they believe cloud services would help to make this possible. As these
examples show, cloud computing is a key component in opening new possibilities for using
data to drive financial services and serves as a foundation upon which a new era of data-driven
innovation is being built.
New Types of Data: Pros and Cons

With the advent of the internet, mobile phones, and smartphones, we have seen a steady influx
of new types of data that either did not exist or were inaccessible just two decades ago. All of
these new forms of data are called “alternative” or “non-traditional” because they are separate
from the credit history data that centralized bureaus have developed an entire industry around.
New categories of data are emerging that span a much wider range of people than ever before
possible. This is creating opportunities to reach markets that have previously been excluded or
poorly served.
While all these new types of data hold great promise, they have one challenge in common:
fragmentation. The surge in data has been spread across a growing number of holders of data,
resulting in much more variation in where data can be accessed and how it is structured and
shared. Just as the traditional credit bureau industry had to solve the problem of aggregating
data from a large number of sources, so too will the alternative data ecosystem. Reflecting on
this point, Spencer Robinson of Kabbage, an online small business lender, said, “I think the way
we look at data is a natural evolution. What we view today as underserved are really populations
[that] did not generate data in a ‘fashion’ that the traditional world today collects it.” This aligns
with the notion that certain segments of the population are underserved in part because they
are financially “invisible.” Solving the broader challenges around sharing and aggregating new
types of data could bring greater visibility into underserved markets and open new growth
opportunities for providers.
The best data to use varies from case to case, and as Figure 6 (pg. 12) shows, various factors
come into play that make different types of data more or less useful, depending on the situation.
Our interviews suggest that no one alternative data type is superior, nor is there one absolute
way to rank them based on their predictive power. Instead, assessing their utility is contextual,
and the appropriate types of data to use depend on the particular product and use case in
question and the corresponding relevance and coverage of available data types.
For example, Spencer Robinson at Kabbage said that for small businesses generating a significant
portion of revenue through online sales, access to eBay or Amazon data would be needed to
make an accurate credit decision. In comparison, for small businesses in construction, Kabbage
would need access to their accounting package to gain more context around the deposits in
their checking accounts. Also, Experian, a leading credit rating agency, has experimented with a
range of alternative data sources in an effort to create credit models for thin-file customers.
Our interviews found mobile phone data to be appealing to many providers. However, getting
mobile network operators (MNOs) to share their data can be difficult. In some instances, larger
financial institutions are able to partner with MNOs or other third parties to access mobile phone
data. Carlos Lopez-Moctezuma with BBVA Bancomer in Mexico said that they have just started
working with intermediaries that aggregate mobile phone data. But that has been the exception
rather than the rule, and many interviewees reported experiencing barriers on this front.

FIGURE 6
PROS AND CONS OF NEW DATA TYPES
Abdul Musoke of the Uganda Communications Commission noted that regulators need convincing
to prioritize taking more meaningful action, while MNOs need more regulatory guidance and
incentives to open up. “Regulation has to help move this forward. MNOs don’t want to lose their
data advantage over fintechs and banks.”
Smartphones, email, and social media platforms have opened new possibilities, but they too
involve a variety of constraints. Although smartphone penetration has been increasing due to
reduced handset costs, data (plan) pricing remains a significant barrier to increased usage and
functionality for low-income consumers. As Thabani Ndwandwe of Standard Bank pointed out,
more and more people have access to smartphones, but in low-income segments they don’t
use data because it is too expensive. In addition to the cost of data plans, the lack of high-
speed mobile network coverage, insufficient battery capacity, low processing power, and many
other technical issues present significant obstacles to the growth of smartphone usage, and the
corresponding potential for increased data capture, in developing markets. As Leon Perlman,
a specialist on digital financial services, cautioned, “While smartphone-using apps and 3G+
networks may become dominant in the future, bringing enhanced services with them, that future
has not yet arrived—especially for customers in the developing world who live in rural areas or
who cannot afford high-end smartphones.”5

As for email, which is widely used, albeit across a number of different platforms, assessing
metadata such as inbox structure, message length, or the timing of when an account was created
can be useful for fraud detection and credit scoring. But as Rodrigo Sanabria at LenddoEFL—an
alternative credit scoring and verification provider active in 20 emerging markets—cautions, “The
time required to gather this data is demanding, so quick decisions can’t be made when analyzing
full bodies of emails.” While email data is not the “bread and butter” for LenddoEFL or any other
interviewees, it was referenced as a widely available source of supplemental information that can
be useful when combined with other types of data.
With the exception of WeBank, China’s first private online-only bank, the interviewees in this
study have not found social media data especially useful. Like mobile phone data, social media
data is held by large players that have their own ambitions in the financial services space. While
social media platforms do have mechanisms through which third parties can obtain customers’
consent for accessing their data, there are limits on the kinds of data available and how it can be
used. Rodrigo Sanabria said that Facebook’s terms of service prohibit the use of data for credit
scoring purposes, for example. And restrictions aside, there have also been questions about
the quality and predictive power of social media data. Mah Kam Lin with Maybank in Malaysia
referred to social media data as “unicorn data,” and added, “I can post a picture of me having a
cup of coffee in Paris even though I’m in Kuala Lumpur … It is aspiration data … I prefer organic
and transactional data. Those are very real.”
We found evidence to suggest that the big tech platforms may be best suited to use social
media data to its full potential in financial services. For example, as we show in the following
section, WeBank is using data from Tencent, one of the world’s largest internet companies, for
credit decisioning. Although banks and fintechs have less traction with social media data, we
expect that social media data will play an important role in the future of financial services.
Bill pay, e-commerce, and psychometric data all have useful applications. The beauty of bill pay
data is its universality and the fact that it reflects actual payment performance, much like traditional
credit history data. How customers pay their phone or utility bills may be fairly indicative of their
creditworthiness. However, an ongoing challenge with bill pay data is the friction associated with
collecting it. For example, Destacame informed us that their customers usually do not have their
bills at hand when accessing the alternative credit scoring company’s platform and that utility
bills are sometimes associated with the household as opposed to the individual. E-commerce
data may be more readily available, but so far, its use is limited primarily to micro, small, and
medium enterprises (MSMEs) that conduct sales online or for online retail finance. Psychometric
data can also aid in scoring those without credit histories or much of a digital footprint, but it
requires more effort by the customer, who must actually take a test.
This discussion gives a brief sense of the landscape of new data and where providers are gaining
traction or hitting friction. Given that many of these new data sources and technologies did not
exist much more than a decade ago, it is encouraging to see how far things have progressed. Next,
we will take a closer look at specific products and use cases fueled by data-driven innovation.
Products and Use Cases

As consumers use financial services, they rarely consider what has to happen behind the scenes
for everything to work. It’s up to providers to figure that out. To provide a fresh glimpse into
how providers are leveraging new data sources, we present snapshots of recent initiatives, as
shared by interviewees. Some of these examples highlight the tradeoffs between data privacy
and access to financial services.

Snapshot 1: WeBank Uses Tencent Data for Unsecured Personal Loans in China
WeBank was launched in 2015 as China’s first private online-only bank. It offers a full
spectrum of consumer financial services, but its unsecured personal loan product is
by far the most active. As of mid-2017, the bank extended total loans of RMB 200
billion (approximately $30 billion)iii in value with a typical loan size of RMB 8,000
(approximately $1,250).6 WeBank has the enormous advantage of having Tencent,
one of the world’s largest internet companies, as its single largest shareholder, which
allows social media data, particularly from Tencent’s WeChat messaging app, to be
a centerpiece of its strategy. WeBank’s Jared Shu noted that WeBank uses blacklist
data from Tencent’s WeChat and QQ platforms as a filter for fraud risk. WeBank is
also experimenting with social graphs derived from social media data to further assess
applicants for fraud. These social graphs, which trace the evolution of profiles and
frequency of communication, help detect whether certain profiles have been fabricated
or are associated with organized crime, for example. WeBank also uses these analyses
to independently assign proprietary social scores to inform loan decisions. WeBank’s
collaboration with Tencent is an ongoing learning process, but with data on over one
billion users, there is great potential, considering that roughly two-thirds of Chinese
people lack traditional credit histories.
Snapshot 2: BBVA Bancomer Uses Blend of External and Internal Innovation in

Mexico
BBVA Bancomer, a subsidiary of BBVA Group and Mexico’s largest bank, has run many
pilots with fintech startups through their open sandbox project. For example, they have
been working with Destacame to test new types of data for alternative credit scoring
and have also worked with Juntos, a U.S.-based fintech, to drive customer engagement
through automated SMS messaging. Today, the bank is pursuing a new internal strategy
using behavioral economics techniques, A/B testing, and transaction data to change
how customers use transactional products, such as checking and savings accounts.
Specifically, BBVA Bancomer is trying to optimize how customers use digital channels.
BBVA’s Carlos Lopez-Moctezuma explained that it was important to get customers
to leave money in their account rather than withdrawing all the cash each time they
receive a deposit. When customers use debit cards and digital banking tools to make
purchases or send money to other people, more data can be captured that can help the
bank better understand their use of their accounts.
Snapshot 3: Grameen America Takes a High-Touch Approach to Internal Data

Collection in the U.S.
Grameen America—a U.S.-based nonprofit microlender which helps low-income
women and immigrants build small businesses through credit, training, and support—
has served about 100,000 customers since its launch in 2008 and aims to reach one
million people over the next 10 years. Following the model of Bangladesh’s Grameen
Bank, the nonprofit works with self-organized groups of borrowers who all have income-
generating activities. Each group must first undergo five consecutive days of training
to learn basic financial skills and demonstrate commitment. Upon completion, each
borrower receives her first loan, which averages about $1,500, and then attends weekly
meetings and makes weekly repayments. Grameen America does not gather credit
reports, scores or other external data to make lending decisions, but instead collects
data directly from customers using mobile devices and stores it on the cloud. They
use this data to assess repayment channels, improve offerings, make follow-on lending
decisions, and inform outreach to the community. Grameen America also works with
Common Cents Lab at Duke University on how to use data to help customers shift
toward use of digital channels.
iii $ denotes U.S. dollar throughout the report.

While they do embrace digital and data-driven tools, Grameen America is
unapologetically high-touch. As Marcus Berkowitz said, “Our most valuable data points
are the ones that we can only collect because we are meeting with someone face to
face.”
Snapshot 4: SCB Abacus Spins Off from Parent Bank to Accelerate Advanced
Analytics in Thailand
SCB Abacus is an advanced data analytics subsidiary of Siam Commercial Bank, the
oldest and one of the largest banks in Thailand. In 2017, SCB Abacus was spun off into
a wholly-owned subsidiary of the bank to give it greater autonomy in pursuing digital
innovation projects for the bank. This flexibility facilitates cross-industry collaboration
on bilateral data partnerships and new business offerings in financial services.
Combining data from non-bank entities with traditional financial data yields deeper
insight into target customers. SCB Abacus has started to use a combination of non-
traditional data for several products. For example, they have developed a customer
rewards feature (named My Deals) through their main banking app that uses artificial
intelligence (AI)iv and machine learning (ML)v to offer personalized promotions and
discounts. The content shown is personalized to each individual user through the use of
an AI-powered recommendation engine. This feature aims to provide the most relevant
offerings available for the bank’s customers. Another project focuses on the area of
credit scoring, by sharpening the bank’s ability to lend to thin-file borrowers through
new underwriting techniques. Alternative sources of data are especially valuable for
such borrowers with limited credit history. This is part of an initiative to better serve
small and medium-sized enterprises (SMEs) in Thailand by providing greater access to
capital to grow their businesses.
These snapshots exemplify how data-driven innovation is creating tangible products and services
for customers. While they represent different philosophies and business models, what they
share is their use of new types of data and tools to reach segments that previously had been
underserved or not served at all. These institutions know the digital revolution is bringing vast
new markets online and that those players that get it right will be rewarded with their business.
CHALLENGES
Despite impressive recent advances, banks face myriad challenges in leveraging data for financial
inclusion, both within the organization itself as well as in the broader ecosystem. We categorize
these challenges into three main buckets: i) internal preparation; ii) data journey management;
and iii) regulatory environment (see Figure 7, pg. 16).
The following paragraphs examine these challenges in detail.
Internal Preparation
The first category of challenges banks must address to get the most out of data-driven
technologies and increase their odds of successfully leveraging data for the purpose of driving
financial inclusion is internal preparation, including changing the banking culture, attracting and
retaining the right talent, and modernizing legacy IT architecture.
iv Artificial intelligence enables software to exhibit human-like intelligence, including learning, planning, reasoning, problem-
solving, and decision-making.
v Machine learning is the science of getting computers to parse through and learn from data in order to perform targeted tasks.

Changing the Banking Culture
A number of interviewees explained that many banks remain risk averse and hesitant to embrace
change. According to Hussam Sultan at CIMB, a commercial bank headquartered in Kuala
Lumpur, the global financial crisis and large regulatory fines imposed on multinational banks in
recent years have helped further entrench this cautious culture.
FIGURE 7
KEY CHALLENGES TO LEVERAGING DATA FOR FINANCIAL INCLUSION
Several interviewees described how important it is to counteract this trend and implement a
culture within their organizations that helps foster innovation and collaboration, promote risk-
taking and experimentation, and capitalize on data-driven technologies. To achieve this, banks
need to introduce an innovative mindset into their culture,7 embrace fresh ways of thinking, and
challenge how operational processes are traditionally performed.
Numerous interviewees emphasized that this cultural shift must start at the top and permeate
the entire chain of command. Without committed support from leadership, including the CEO
and Board, it is extremely difficult to innovate. Establishing a new culture requires leadership to
clearly communicate to all employees the implications for existing roles, incentive structures, and
expectations.8 It also helps when top executives at an institution have an innovation or fintech
background. For example, Bancolombia’s President, Juan Carlos Mora Uribe, previously served
as the banking group’s head of innovation, and BBVA’s Group Executive Chairman, Francisco
González, was a programmer at one point in his career. Both leaders are firm believers in financial
technology and are helping to change the business culture within their organizations.
For several of the financial institutions we spoke with, instilling the will to proactively and
continuously explore and experiment with alternative data and tools is a high priority. Reflecting
Maybank’s efforts on this front, Mah Kam Lin said, “All 44,000 employees are going through an
upskilling program in terms of data awareness, programming, how to use data to make better
decisions, human-centered design, and agility.” She added that the leadership at Maybank is very
supportive of this culture change and wants to make sure the bank future-proofs its workforce, as
organizations that do not will be ill-equipped to support the underserved.

“All 44,000 employees are going through an upskilling program in terms of data
awareness, programming, how to use data to make better decisions, human-centered
design, and agility.”
-- Mah Kam Lin, Maybank, on the financial institution’s efforts to embrace data technology.
Attracting Technical Talent

The scarcity of technical talent challenges incumbents in the transition to a more data-dependent
world. Banks face a tremendous shortage of experienced data scientists and engineers, as
well as AI experts, all of whom are needed to help convert big data “noise” into meaningful
insights. Even relatively basic data skills are lacking in many major institutions, resulting in wasted
resources, project delays, and squandered opportunities. For example, fintechs that partner with
banks often find themselves training their counterparts, slowing down collaboration and product
launches.9
The ability to recruit, develop, and retain workers with proficiencies in fields related to computer
programming and data engineering is increasingly vital to the banking industry’s capacity to
expand services to the underserved and to its overall success in general. According to Gabriel Di
Lelle of Bancolombia, attracting people with the right set of skills is “the number one challenge”
as banks are not the preferred destination for employees with data skills, who typically gravitate
toward cool, innovative, and less formal working environments. Mr. Di Lelle went on to say that
this challenge is not exclusively a banking problem, but one that many industries around the
world are grappling with, as the demand for data specialists far outstrips supply.
Moving forward, banks, especially in certain emerging markets where the talent shortage tends to
be more pronounced, will need to pay special attention to devising effective strategies outlining
how to attract top technical talent.
Modernizing Legacy IT Systems

One of the biggest internal challenges facing banks as they incorporate new data and analytics is
modernizing legacy IT systems. Many of today’s banks rely on infrastructure created well before the
digital era. Over the years, these systems have been updated incrementally to tackle immediate
pain points and have been merged, split, and joined again by other systems as banks went
through various mergers and acquisitions. Furthermore, in many instances, technology teams in
different parts of the organization address pain points independently. While this decentralized,
incremental approach has historically allowed banks to muddle along, it contributes to siloed
systems, redundancies, and increased complexity.10 The rapid digital transformation currently
underway creates an urgent need for banks to completely overhaul IT systems while ensuring
regulatory compliance.11 Not doing so hinders a bank’s ability to collect, structure, analyze, and
leverage data fully, and makes it difficult to collaborate with external partners. Unfortunately, the
end-to-end restructuring of IT systems is a complex, costly, and time-consuming undertaking that
requires long-term vision, dedicated leadership, and cooperation across many departments (not
just IT) and, sometimes, jurisdictions. Moreover, these changes must be done while preserving
the availability and continuity of secure customer services12 and maintaining acceptable quarterly
results for shareholders and investors. Nevertheless, building state-of-the-art IT architecture is
worth pursuing. The resources required up front are likely to pay off in the long-term as companies
become better positioned to capitalize on data, streamline decision-making, mitigate risks,
facilitate automation, and improve overall business efficiency.
An innovative example of a bank modernizing its IT infrastructure for a data-driven world is SCB
Abacus, the first advanced data analytics spin-off of its kind within the Thai financial industry. The
subsidiary entity was created to unleash innovation, experimentation, flexibility, and efficiencies
that were tricky within the traditional organizational and IT structure of its parent company,
according to Natth Bejraburnin of SCB Abacus.

One key challenge with respect to IT infrastructure is security. As data technology permeates
banking, and as the frequency and sophistication of malicious virtual attacks grow, cybersecurity
is increasingly critical. Repeated breaches can damage operations, profits, and reputation,
potentially jeopardizing a bank’s ability to attract and retain customers. Safeguarding sensitive
client information and protecting internal systems is thus essential to the success, and, ultimately,
survival of banks. Astute banks are exploring next-generation cybersecurity technology as they
transition from static security models to dynamic ones that offer new layers of defense, are more
predictive, and are built to not only prevent attacks but also to adapt to various circumstances
and easily detect and resolve breaches within large and complex banking networks.13
Data Journey Management

The second category of challenges facing banks is navigating the broader data ecosystem.
As outlined above, data is abundant and is constantly being produced. However, a number of
obstacles related to collection, structuring, security, and value extraction can arise along the
complex data processing journey, as we will see below.
Identifying the Most Relevant Data and Where to Begin

The sheer volume of available data can be overwhelming. Many banks are unsure of where
to start, what to look for, and how to identify and capture the most appropriate data. Even
tech-savvy banks with dedicated data teams are struggling to extract the full value from their
data, simply because of its vastness. According to Gabriel Di Lelle of Bancolombia, “There is a
belief inside Bancolombia that we are probably not even doing 10 percent of what we could be
doing with our transactional data.” Moreover, during our conversation with CIMB, Hussam Sultan
estimated that, in general, “85 percent of the data made by banks is wasted.” This is one of the
reasons why banks often collaborate with third parties (see partnership section on page 21).
“85 percent of the data made by banks is wasted.”

--Hussam Sultan, CIMB
Slow and Arduous Value Extraction

The amount and complexity of data, especially semi-structured and unstructured data, often
require multiple players to turn it into something of value. Collecting, structuring, managing, and
analyzing the data is a lengthy process that can involve several collaborators. Depending on the
technical abilities of each partner, this process may be slow and arduous.
Getting Data Privacy and Consumer Consent Right

In a world of serial data breaches, hacking, and misuse, a practical challenge is protecting
consumers’ data privacy and maintaining their trust. Consequently, banks are oftentimes
reluctant to turn to outside partners that could help them harness consumer information as they
worry about the third party’s commitment to safeguarding data.14 Foregoing collaboration with
specialized data companies, however, can limit the bank’s ability to extract data’s full value. In
addition, customers can also be reluctant to provide consent. According to Thabani Ndwandwe
of Standard Bank, one of the biggest challenges when it comes to alternative data is obtaining
customer consent, as many customers do not feel comfortable giving banks access to additional
personal data such as social media activity. Customers, however, are usually more willing to
provide consent if it is in exchange for a valuable service that they need, and do not have access
to, such as credit. In the current climate, being able to demonstrate particularly strong and
cohesive processes can be a competitive asset for banks. Enhanced transparency towards the
customer vis-à-vis other competitors could also enhance customer trust and their willingness to
give consent.
Data Fragmentation and Willingness of External Parties to Share

According to DemystData, “While, historically, the bulk of actionable data was controlled by a
small number of players, the proliferation of data means there are now a myriad of data owners,

each of which contribute only fragments of a customer’s full data profile.”15 Moreover, some data
owners are hesitant to share information. Despite numerous partnerships between banks and
telecommunication companies (“telcos”), for example, some telcos hesitate to share their mobile
data with banks. According to one bank executive we spoke with, trust between telcos and banks
is difficult to achieve in certain markets, particularly when they are competitors offering similar
financial services. Other interviewees suggested that telcos are not sharing their data as much as
they could. This might be due to a number of reasons, including regulatory concerns as well as
protecting and capitalizing on their own data. Therefore, this data fragmentation makes it harder
to obtain and derive value from the consumer’s full data profile.
Incomplete Digital Proliferation

While the digital revolution has reached a large number of the underserved and unserved
population worldwide—making it easier for financial service providers to harness the data
they generate to better serve them—inadequate data still persists for certain sub-sections of
this community.16 For example, some older customers are not comfortable using computers
or mobile phones. Marcus Berkowitz of Grameen America told us that many of their older
members appreciate the fact that they can participate in the loan programs without having to
use new technology. Carlos Lopez-Moctezuma at BBVA pointed out that a major challenge to
obtaining data on certain customer segments is incentivizing customers away from cash. Even
with the ability to use mobile wallets—which provide a gold mine of data—some lower income
consumers can be reluctant to give up their cash and convert it into digital money, as they’ve
grown accustomed to the tangibility of paper notes and coins. For many, cash is indeed still king.
Other factors contributing to a lack of data on particular segments include lower income and
education levels—which can limit an individual’s usage of data-capturing technology—as well as
certain cultural and social norms that can lead to distrust of large organizations, such as banks,
and restrictive roles for women.17 Therefore, thanks to economic, demographic, psychological,
social, and cultural factors, the digital footprints of these types of consumers tend to be narrower
or non-existent, reducing banks’ ability to harness data to serve them.
Regulatory Environment
The third category of challenges facing banks is the current regulatory landscape within which they
operate. National and international regulations on data sharing and privacy have a large impact
on how banks can use consumer information. Complicating matters for banks operating in various
jurisdictions are overlapping and sometimes contradictory regulatory requirements, which make
compliance a complex and expensive undertaking. Given the patchwork of regulations related to
new data sources and tools, it is not surprising that banks and fintechs alike seek guidance from
policymakers on these issues.
In many instances, relevant regulation simply has not yet been developed or is relatively light,
but this too creates its own barriers. It can increase uncertainty and cause financial service
providers to hesitate to pursue data-dependent product offerings. Although lack of regulatory
clarity leaves some providers searching for answers, others adopt best practices from markets
with more mature regulatory landscapes as a preemptive move. In a separate project, the IIF is
working to inform and engage regulators globally on the transformative potential of AI and ML
technologies in credit and risk modeling. As part of this project, the IIF surveyed 60 member
financial institutions and will share key findings in a forthcoming paper.
Fintechs sometimes follow regulations placed on their more regulated partners to inform their
compliance standards. For example, Artoo is not directly subject to lending regulations, but finds
that it must comply with the data management standards that its partner lending institutions
must meet. It obtains consent of the borrower through SMS at the point of data capture. Beyond
its direct lending operations to small businesses and consumers in the United States, Kabbage
has developed a white-label Software as a Service (SaaS) platform to enable partnerships with
lending institutions overseas, namely, ING, Santander, and Scotiabank. While some partners

connect to the lending platform directly, Kabbage has also established overseas data centers
in other cases to handle certain country-specific data storage requirements. Considering this, it
views PII as a liability, preferring to work with anonymized data unless absolutely necessary to
the credit approval process.
Data localization regulations are particularly challenging to financial institutions with regional
and global reach, especially those that are looking to leverage the benefits of cloud computing
services. New data technologies work best to generate insights when data can be stored and
accessed in centralized, consistent, and structured ways; however, restrictions on the ability to
share, store, and access data across borders complicate this. According to a 2017 article by the
American Bar Association, “more than two dozen countries have enacted or considered policies
that require retention of data within their borders.”19 Data localization rules could lead large
banks to avoid serving certain jurisdictions entirely or require them to build and maintain multiple
data centers across the globe.
Another overarching challenge is that policymakers are cautious in setting new standards and
face competing regulatory priorities. In Uganda, where telecommunications regulators are
focused on ensuring competition in the telco market (and tend to limit data sharing), they are
not necessarily incentivized to prioritize financial inclusion or work towards getting telcos to open
their data to banks and fintechs. Until greater consensus on how to best proceed is reached
internally among regulators, actors looking to leverage new data sources—banks, fintechs, and
others—will either need to proactively seek out opportunities for data collaboration or continue
to make do with what they have.
In jurisdictions where regulation is more rigid, financial service providers face obstacles in
deploying innovations. Colombia’s data protection regulations, based in part on European
models, have restricted the ability of financial service providers to partner with telcos and use
specific types of data (e.g., geolocation data, information on customers’ physical movements)
that could improve credit scoring models and product offerings. Bancolombia has had even less
success in convincing traditionally-minded regulators in other Latin American countries of the
benefits of working with new sources of data.
In South Africa, according to Thabani Ndwandwe at Standard Bank, credit regulations designed
to protect consumers have made it much more challenging and costly for regulated providers
to lend to underserved, low-income clients. Specifically, rules around documentation to verify a
client’s income prevent lending to consumers earning 1,000 rand a month or less (approximately
$83), given that many low-income clients work outside the formal economy and do not have
bank statements or formal pay slips.
OPPORTUNITIES
Engagement with Regulators and Policymakers
Despite obstacles outlined above, a significant number of interviewees shared positive
experiences working with regulators and are looking for regulatory guidance on these issues.
Many institutions have benefitted from engaging directly with regulators and other policymaking
bodies. In India, Sameer Segal of Artoo highlighted that their legal counsel has drafted model
legislation on compliance issues that are important to their partner lending institutions, such
as whether lenders working with Artoo comply with data privacy regulations when using the
company’s machine learning services and whether they need explicit consent from customers to
use these services.vi
vi To avoid running into issues regarding the handling of personally identifiable information (PII), Artoo’s machine learning service
only uses anonymized data.

Similarly, in South Africa, Standard Bank has done advocacy work, engaging with parliamentarians
and regulators to convince them of the virtues of using new data sources and tools for the benefit
of reaching low-income segments.
In Malaysia, Maybank interfaces regularly with the central bank (Bank Negara Malaysia) and has
played a significant role in supporting Bank Negara’s regulatory sandbox and other efforts to
support fintechs and innovators in this space.
Financial service providers and regulators can work together to troubleshoot unanticipated
obstacles as they arise. For Bancolombia’s mobile-only banking product, Nequi, it was essential
to work with Colombia’s regulators to remain in compliance with KYC regulations. Launched
in 2017, Nequi accounts feature a simplified KYC process, with balance limits that justify the
simpler identification requirements. However, as Nequi customers began to hold larger balances,
Bancolombia had to develop a solution to graduate these clients to different account types
in a way that would be easy for the customers. In-person interviews are extremely costly, so
Bancolombia came to an agreement with regulators which allowed for chat-based interviews to
be conducted over Facebook Messenger.
Regulators recognize their responsibility to take the lead and initiate broader ecosystem
discussions pertaining to new and alternative data sources and tools, especially given the diverse
array of players looking to get into this space. In Egypt, where financial inclusion is a top priority
of the government’s strategy for 2030, the central bank has kicked off initial conversations on
cloud computing—a crucial piece of technology that CIB Egypt, the largest private bank in the
country, expects will help minimize the costs of going down-market—and is drafting data sharing
regulation similar to the GDPR.
Abdul Musoke of the Uganda Communications Commission is in favor of a regulatory sandbox

whereby regulators can oversee the development and testing of products within the confines of
a controlled environment. SCB Abacus is currently going through such a process with the Bank
of Thailand. In order to bring digital credit products that leverage machine learning techniques
to market, SCB Abacus and their compliance team submitted a proposal to the Bank of Thailand
detailing the types of data and models they would use, and the key performance indicators
they would measure—all in advance of submitting a separate application to join the regulatory
sandbox. Even if the specific products do not graduate from the sandbox into the market, such
experimentation would serve well to establish greater trust and collaboration among banks,
fintechs, and regulators.
As data-driven technologies fuel further innovation and transformation, it will become increasingly
important for banks to engage with officials and encourage standard-setting bodies to develop
regulatory guidance that protects consumers while also addressing cross-border inconsistencies
and regulatory fragmentation. This will help ensure that banks can effectively leverage data
technology across jurisdictions to improve their ability to expand quality financial services to
emerging customer segments.
Partnerships
Partnerships between banks and fintechs offer enormous opportunities to leverage data for the
purpose of driving inclusion. Fintechs, which are increasingly viewed by banks as valuable partners
rather than direct competitors, face fewer legacy hurdles and benefit from greater specialization,
risk tolerance, and agility. This complements banks’ deep pockets, brand recognition, large
customer base, and existing data.20 Because innovation within many banks is fairly difficult due
to their size, legacy constraints, and extensive approval processes, collaborating with fintechs is a
great model as it typically enables greater flexibility for trial and error, speeds up time to market,
and reduces the risks associated with innovation. These mutually beneficial partnerships allow
fintechs to scale their technology and access capital, while banks gain assistance in harnessing

customer data for purposes such as improving product offerings, increasing efficiency, lowering
costs, and expanding their customer pool.
Our publication on bank-fintech partnerships noted, “As it turns out, these are all goals with
special relevance to low-income customers who look for products that are more convenient, less
expensive, and higher quality, and that makes financial institution-fintech partnerships a crucial
strategy for meeting the financial needs of the underserved around the world.”21
We will briefly spotlight a few banks we spoke to that are partnering for the purpose of harnessing
data. CIB Egypt—the first bank in the Middle East with a dedicated data science team, according
to Islam Zekry—has begun working with Fawry, a local electronic payments provider, to offer a
bill payment solution on its CIB Smart Wallet. The expectation moving forward is that the smart
wallet will provide data on customers’ payment habits which can be used to profile and target
customers for lending and deposit products. CIB Egypt is also collaborating with Careem, the
region’s leading ride-sharing company. The new relationship provides Careem-Egypt with an
integrated set of digital financial solutions to efficiently manage driver incentives and payment
disbursements via the CIB Smart Wallet. Moreover, the mobile wallet allows drivers to receive
payments from Careem, send money, pay bills, and make deposits or withdraw cash through the
bank’s large network of ATMs and agents. What makes this relationship relevant from a data use
perspective is that CIB merges transaction data from the wallet with driving performance and
rider ratings from Careem to develop a persona for each driver and assign a credit score, which
enables CIB to lend to high-performing drivers.
Other examples of banks involved in data collaboration include BBVA and Bancolombia.
BBVA is exploring partnerships with several intermediaries that leverage telco data for use by
financial institutions. These intermediaries purchase the data from the telcos, structure it, and
analyze it to provide key insights to banks. This type of data could provide valuable insights
into an individual’s willingness and ability to repay loans. For example, in some data models,
initiating calls (as opposed to receiving them) and lengthy conversations tend to correlate with
creditworthiness. On the other hand, above-average call activity during regular business hours
and a small network of people in one’s calling circle correlate with low credit scores.22 Finally,
Bancolombia is exploring a pilot with a fintech focusing on social media-based credit scoring.
This type of data can also potentially help assess whether a person is likely to pay back a loan.
Possible indicators of creditworthiness include longstanding social media accounts and a large
network of contacts.23 It is worth emphasizing that these exploratory partnerships are in the early
stages of collaboration, and the effectiveness and viability of these alternative modeling systems
are still being assessed by the banks.
Some of the fintechs we interviewed work primarily by forming partnerships that support financial
institutions. Matt Hennessy at DemystData—a SaaS provider—stated, “A huge opportunity
for us is leveraging partnerships with consulting and analytics firms who specialize in helping
financial institutions implement solutions faster.” The fintech has contracts with approximately
30 financial institutions worldwide and is exploring relationships with 10 others. DemystData
provides a technology platform to standardize and clean partner banks’ data, source additional
data from third parties, and test data in an efficient and secure manner through a single integration
point. The firm’s platform allows banks to verify customers in real-time and assess credit risk
for potential new customers by enabling access to new data sources for customer verification
and underwriting. Banks can review information from various data sources, conduct tests and
validation, and then select customers based on their internal risk appetite. The banks still own all
the customer data and access DemystData’s cloud-based platform through an API connection.
These kinds of partnerships help financial institutions around the world harness data to better
serve more consumers.24

In India, Artoo, a company offering a cloud-based platform that digitizes the entire credit
process, works with eight institutions, including Ujjivan, a small finance bank, to help expand
credit throughout the country. Artoo’s platform allows Ujjivan field agents to input customer
data and onboard customers digitally, which helps improve the tracking of sales and portfolio
performance. Moreover, because Ujjivan agents can capture more data points on potential
customers, higher quality information is generated for the bank to use in credit decisions. And
while data is captured on the fintech’s platform, it is owned by the bank, just like in the previous
example. According to Sameer Segal of Artoo, it typically takes three months to go live with a
partner and the benefits for both parties are tremendous. Forecasts suggest that over the next
couple of years Artoo’s partners will disburse approximately $2 billion in loans leveraging the
fintech’s data platform. Finally, Destacame’s Sebastián Ugarte informed us that his company
works with approximately 20 financial institutions in Chile and Mexico, including BBVA Bancomer.
Partnerships, however, do come with their own set of challenges. A number of organizations
we spoke with highlighted regulatory challenges surrounding data sharing arrangements. For
example, in its relationship with Ujjivan, Artoo had to switch from Amazon Web Services, with
servers housed in the U.S., to a local Microsoft-based cloud service provider, as customer data
had to remain within the country’s borders. Interviewees based in Latin America informed us that
regulation often limits the flow of data between banks and alternative risk modelers, for instance.
This is a point of much contention in many markets and can create some tension in certain
partnerships. Banks seek to maintain control over customer financial data, while fintechs hope to
gain access and use data for a variety of applications (e.g., credit assessment, lead generation,
etc.). Conversely, fintech partners aim to keep their risk models proprietary while banks face
regulatory requirements to disclose their models.25
Other obstacles can arise around negotiations, integration, approvals, coordination, and
getting institutional buy-in—even between institutions that are well-organized to innovate
and partner. Nevertheless, we believe the challenges that emerge during collaboration tend
to be outweighed by the complementary strengths each party brings to the table. Moreover,
collaborative experiences over the past couple of years have taught banks and fintechs valuable
lessons that are now being heeded.
Moving forward, partnerships will continue to facilitate expansion of services to new customer
segments and drive bank innovation. Through partnerships, banks will learn more quickly what
is possible and shift their strategies accordingly. Greater collaboration will enable them to test
technology in low-risk ways, understand how it works for their customer base, and speed the
time to roll out a new product. As the number of successful examples of partnerships rises and
their impacts become increasingly apparent, we expect to see more, and stronger, collaboration
between banks and fintechs.vii
Emerging Technologies
Two emerging technologies with the potential to create enormous opportunities around data and
how it can be leveraged to deepen financial inclusion in the longer-term are artificial intelligence
and blockchain.
According to Sameer Segal of Artoo, AI is opening up a new dimension and will play an important
role as human-machine collaboration expands what is possible. Bancolombia, BBVA, CIB Egypt,
SCB Abacus, and Standard Bank are just some examples of the companies we interviewed
that are already using artificial intelligence in some way. AI is helping organizations collect,
connect, structure, and analyze enormous amounts of data more efficiently. Moving forward,
the technology is expected to play an increasing role in modernizing processes and digital
infrastructure of financial institutions.
vii For more information and detailed case studies on how partnerships between financial institutions and fintechs drive financial
inclusion, please see the first report of the Mainstreaming Financial Inclusion: Best Practices series by clicking here.

“The analytical power you can get through machine learning, deep learning, neural
networks, and other types of AI technologies will really push the boundaries of credit
scoring and serving customers in a much more personalized way and without human
intervention.”
--Gabriel Di Lelle, Bancolombia
Thanks to its ability to independently gather, process, and analyze enormous amounts of data
quickly and effectively, AI could help banks make more informed decisions, automate processes,
and reduce operational costs so that it becomes more profitable to provide affordable services
for low-income customer segments. Furthermore, because AI software is easily scalable and can
enhance automation, banks utilizing the technology could offer services at a level of sophistication,
customization, and scale never before possible, while also reducing operational costs as the need
for manual tasks gradually declines. According to Gabriel Di Lelle at Bancolombia, “The analytical
power you can get through machine learning, deep learning, neural networks, and other types of
AI technologies will really push the boundaries of credit scoring and serving customers in a much
more personalized way and without human intervention.”
AI software could, for example, automatically collect real-time credit data on an individual
seeking a small business loan and then make a near-instantaneous decision on their application
and set a personalized interest rate based on their risk profile. Samir Bhatia of SMEcorner, an
online lender in India, envisions future underwriting to be performed via interviews conducted
on video conferencing platforms where AI technology will both ask questions and analyze
answers, facial expressions, and emotional cues to help determine credit decisions. This could
help bring greater consistency to decision-making within the industry and make processes more
cost-effective, efficient, and timely, with less human involvement. These scenarios, however, will
undoubtedly be influenced by the evolving national and international regulations on data and
technology.
Similarly, blockchain, because of its enormous potential to form the underlying architecture
for myriad applications, could also help deepen financial inclusion. For example, distributed
ledger technology could increase efficiency by automating identity verification—the lack of
reliable identity papers among lower income segments in emerging markets being a significant
obstacle to their inclusion. This is especially true for the most vulnerable populations, including
refugees. According to ID2020, a public-private partnership dedicated to solving challenges of
identity through technology, 1.1 billion people globally lack an officially recognized identity,
leaving them susceptible to economic exclusion. Creating a digital identity through blockchain
could help bring them into the formal economy. Blockchain could also help banks streamline
KYC processes and ameliorate the customer onboarding experience for less vulnerable but still
underserved populations.
By acting as an information-sharing mechanism and providing a universal source of truth that is

tamper-proof and distributed, blockchain could lead to increased efficiency and improvements,
including establishing a level of accountability and transparency that hitherto was impossible;
reducing human error, fraud, data duplication, processing delays, and transaction costs; and
providing easier and improved data access to both internal and external parties. Moreover, with
the potential to create reliable decentralized databases, give people control over their data, and
create borderless, immutable financial identities, blockchain could have a considerable effect
on emerging customers and financial inclusion. That being said, blockchain is unlikely to have a
significant impact on emerging customers in the near or medium term.
While these emerging technologies demonstrate valuable applications today and show significant
future promise, some interviewees warned not to overhype the potential impact. According to Marcus
Berkowitz of Grameen America, emerging technologies have “the potential to be transformative in
certain ways and for certain people … but they are not a silver bullet for financial inclusion.”

THE COMPETITIVE LANDSCAPE MOVING FORWARD
Data can be a powerful force for financial inclusion, and its promise is alluring. The business
opportunity at the inclusion frontier is indeed enormous, as there are billions of underserved
customers globally, and internet-connected devices such as smartphones continue to proliferate
(see Figure 8). For example, according to CIB Egypt, the business opportunity presented by the
untapped market within Egypt is comparable to the traditional banked customer segment—
around 1.8 trillion Egyptian pounds (approximately $90 billion). And as we have pointed out,
although banks face a number of challenges to harnessing data’s full potential for expanding
financial services to new customer segments, many are embracing innovation, integrating
data-driven technologies into their “organizational DNA,”26 and pursuing mutually beneficial
partnerships with fintechs.
As they seek to become competent data deployers, and consider leveraging data to reach
emerging customer segments, incumbent banks must also reckon with growing competition.
Many of our interviewees stated that banks that fail to capitalize on the explosion of data and
continue to rely on legacy systems will find themselves losing market share in existing markets
and failing to gain a foothold in new ones.
Ultimately, what is at stake may be the continued centrality of banks in the financial lives of
customers.
With new players like Alibaba and Tencent growing rapidly and others like Amazon and Facebook
dipping their toes into financial services, incumbents feel pressure to innovate to protect their
market share. Many bankers view these “BigTechs”—firms that provide intuitive online platforms,
excel in utilizing customer data, and are well capitalized—as their primary future competitors.
Such platform companies already have powerful multi-dimensional data on their users, providing
them with a more complete picture. If firms like these expand their role in financial services, as
anticipated, traditional banks will need to innovate significantly in order to compete. If not, they
could fade into the background, becoming transaction processors with limited direct customer
relationships. Jared Shu at WeBank suggests that this is already happening in China, where the
FIGURE 8
GLOBAL SMARTPHONE SUBSCRIPTIONS AND PENETRATION RATE

customer-facing entity is not usually a traditional incumbent.27 He foresees increasing separation
between financial product manufacturing and distribution in China. Banks will produce the
financial products, but tech companies with large networks will distribute them to the customers.
The bank’s role will become increasingly invisible to customers. This potential disruption to the
industry will not be an exclusively Chinese phenomenon.
Banks also face competition from other banks that embrace innovation. Astute banking
executives increasingly view data-driven technologies and innovation in general as key to
remaining competitive and long-term success. This has led many banks to challenge the status
quo and accelerate the shift toward digitization, by changing the banking culture within their
organizations, partnering with third parties, and modernizing and integrating digital systems
and infrastructure. As a result, it is becoming ever more important for all banks to embrace data-
based technologies to improve efficiency and their understanding of customer needs. Speaking
to this point, Carlos Lopez-Moctezuma of BBVA, told us, “We are trying to generate as much
analysis as we can from the market and our current customers so that we can be relevant in
the different product offerings that we have.” Moving forward, banks that lead in the space
will be in a much better position to overtake the competition and increase market share while
simultaneously cutting operational costs.
We end this report as we began, with a call to action. Banks recognize the need to use new
data and tools to tap the vast underserved market segments at the base of the pyramid. But
much of the real work lies ahead. A greater sense of urgency is needed, especially from the
leadership of mainstream financial service providers, to embrace data-driven innovation as a
key to unlocking underserved markets in a commercially viable way. Adopting data-focused
strategies—including developing a culture of innovation, attracting technical talent, modernizing
IT systems, collaborating with fintechs and other parties, and engaging with regulators—can
help banks reach these markets more quickly and efficiently, and in doing so, support financial
inclusion, poverty alleviation, and economic growth. Moreover, such strategies can help banks
future-proof their business models and increase their chances of success.

APPENDIX A: STUDY METHODOLOGY
This study examines how financial service providers are leveraging new (and existing) data
sources and tools to address challenges associated with serving customers at the base of the
economic pyramid, primarily in emerging markets. CFI and the IIF carried out in-depth interviews
over two months with 21 people from 18 financial institutions, fintechs, and other entities. A list
of interviewees can be found in Appendix B and a sample interview guide is available on request.
In addition, an advisory group of stakeholders and industry experts (see “Acknowledgments”
section) provided the researchers with guidance.

APPENDIX B: INTERVIEWS
Institution Interviewee Title
Artoo Sameer Segal Founder and CEO
Bancolombia * Gabriel Di Lelle Vice President of Innovation and Digital Transformation
BBVA Bancomer / BBVA * Carlos Lopez-Moctezuma Head of New Digital Businesses and Financial Inclusion
Caribou Digital Maha Khan Senior Manager, Research and Knowledge Management
CIMB Group * Hussam Sultan Head, Transaction Banking and ESG Integration
Commercial International Bank (Egypt) Islam Zekry Chief Data Officer
S.A.E. *
Commercial International Bank (Egypt) Amin Khairy Risk and Support Functions Analytics Manager
S.A.E. *
DemystData Matt Hennessy Director, Strategic Accounts

Destacame Sebastián Ugarte Co-Founder
Grameen America Marcus Berkowitz Senior Director, Technology and Innovation
Kabbage Spencer Robinson Head of Strategy
Kabbage Winn Martin Head of Data Strategy
Konfio David Arana Co-Founder and CEO
Konfio Francisco Padilla Sánchez Co-Founder and CTO

LenddoEFL Rodrigo Sanabria Director, Partner Success, Latin America
Malayan Banking Berhad (Maybank) * Mah Kam Lin Group Data Scientist
SCB Abacus / Siam Commercial Bank * Natth Bejraburnin Senior Data Scientist
SMEcorner Samir Bhatia Founder and CEO
Standard Bank * Thabani Ndwandwe Head, Credit for Personal and Business Banking, South Africa
Uganda Communications Commission Abdul Musoke Manager, Economic Regulation

WeBank * Jared Shu Head of Strategy
* Denotes IIF member.

REFERENCES
1
Costa, Arjuna, Anamitra Deb, and Michael Kubzansky, “Big data, small credit,” Omidyar Network, 2015.
2
Adams, Timothy, “Core Principles 4th Report,"Institute of International Finance, April 30, 2018.
3
“Copy That,” The Economist, April 5, 2018.
4
Mullins, Scott, “Leveling the Playing Field: Cloud as the Enabler of Innovation in Financial Services,” Bank Innovation, February 26,
2018.
5
Perlman, Leon, “Technology Inequality: Opportunities and Challenges for Mobile Financial Services,” The Center for Financial Inclusion
at Accion, April 2017.
6
“Bank of the Future: The ABCs of Digital Disruption in Finance”, Citi GPS: Global Perspectives & Solutions, March 2018.
7
Partnership for Finance in a Digital Africa (FiDA), “Snapshot 9: Best Practices in Big Data Analytics,” Farnham, Surrey, United Kingdom:
Caribou Digital Publishing, 2017.
8
Yaworksy, Kathleen, Dwijo Goswami, and Prateek Shrivastava, “Unlocking the Promise of (Big) Data to Promote Financial Inclusion,
Accion, March 2017.
9
Partnership for Finance in a Digital Africa (FiDA), “Focus Note: Can Big Data Shape Financial Services in East Africa?” Farnham, Surrey,
United Kingdom: Caribou Digital Publishing, 2018.
10
Desmet, Driek, Markus Loffler, and Allen Weinberg, “Modernizing IT for a digital era,” McKinsey, September 2016.
Partnership for Finance in a Digital Africa (FiDA), “Snapshot 9: Best Practices in Big Data Analytics,” Farnham, Surrey, United Kingdom:
11
Caribou Digital Publishing, 2017.

12
Ibid.
13
Innovation by Mid-Sized Banks, Institute of International Finance, November 2016.
Cheston, Susy, Tomas Conde, Arpitha Bykere, and Elisabeth Rhyne, “The Business of Financial Inclusion: Insights from Banks in
14
Emerging Markets,” Center for Financial Inclusion at Accion and Institute of International Finance, July 2016.
15
“Cleaning Up the Universe of Data,” DemystData, July 2017.
Cheston, Susy, Tomas Conde, Arpitha Bykere, and Elisabeth Rhyne, “The Business of Financial Inclusion: Insights from Banks in
16
Emerging Markets,” Center for Financial Inclusion at Accion and Institute of International Finance, July 2016.
17
Partnership for Finance in a Digital Africa (FiDA), “Snapshot 2: Which Attitudes, Behaviors, Experiences, and Beliefs Influence Digital
Finance Adoption?” Farnham, Surrey, United Kingdom: Caribou Digital Publishing, 2017.
18
Accion, March 2017.

19
Cohen, Bret, Britanie Hall, and Charlie Wood, “Data Localization Laws And Their Impact on Privacy, Data Security And the Global
Economy,” Antitrust, Vol. 32, No. 1, Fall 2017, American Bar Association.
20
Innovation by Mid-Sized Banks, Institute of International Finance, November 2016.
21
Kelly, Sonja, Dennis Ferenzy, and Allyse McGrath, “How Financial Institutions and Fintechs are Partnering for Inclusion: Lessons from
the Frontlines,” Center for Financial Inclusion at Accion and Institute of International Finance, July 2017.
22
McEvoy, Michael J., “Enabling Financial Inclusion Through ‘Alternative Data,’” MasterCard Advisors, 2015.
23
Ibid.
24
25
Ibid.
26
Accion, March 2017.

27

The Center for Financial Inclusion at Accion is an action-oriented think tank that engages
and challenges the industry to better serve, protect and empower clients. We develop insights,
advocate on behalf of clients and collaborate with stakeholders to achieve a comprehensive
vision for financial inclusion. We are dedicated to enabling 3 billion people who are left out of –
or poorly served by – the financial sector to improve their lives.
For more information visit www.centerforfinancialinclusion.org
The Institute of International Finance is the global association of the financial industry, with
close to 500 members from 70 countries. Its mission is to support the financial industry in
the prudent management of risks; to develop sound industry practices; and to advocate for
regulatory, financial and economic policies that are in the broad interests of its members and
foster global financial stability and sustainable economic growth. IIF members include commercial
and investment banks, asset managers, insurance companies, sovereign wealth funds, hedge
funds, central banks and development banks.
For more information visit www.iif.com.
This research was made possible by generous support from the MetLife Foundation.

Accelerating Financial Inclusion With New Data

Uploaded by

Copyright:

Available Formats

Accelerating Financial Inclusion With New Data

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Accelerating Financial Inclusion With New Data

Uploaded by

Copyright:

Available Formats

Accelerating Financial Inclusion with New Data

PART FOUR OF THE MAINSTREAMING FINANCIAL INCLUSION: BEST

Accelerating Financial Inclusion with New Data 1

Accelerating Financial Inclusion with New Data 2

Accelerating Financial Inclusion with New Data 3

Accelerating Financial Inclusion with New Data 4

Accelerating Financial Inclusion with New Data 5

Accelerating Financial Inclusion with New Data 6

FIGURE 2 With over 1.7 billion people still

Accelerating Financial Inclusion with New Data 7

ii Unless otherwise noted, all referenced interviews are listed in Appendix B.

Accelerating Financial Inclusion with New Data 8

Accelerating Financial Inclusion with New Data 9

Accelerating Financial Inclusion with New Data 10

New Types of Data: Pros and Cons

Accelerating Financial Inclusion with New Data 11

Accelerating Financial Inclusion with New Data 12

Products and Use Cases

Accelerating Financial Inclusion with New Data 13

Snapshot 2: BBVA Bancomer Uses Blend of External and Internal Innovation in

Snapshot 3: Grameen America Takes a High-Touch Approach to Internal Data

iii $ denotes U.S. dollar throughout the report.

Accelerating Financial Inclusion with New Data 14

The following paragraphs examine these challenges in detail.

Accelerating Financial Inclusion with New Data 15

Accelerating Financial Inclusion with New Data 16

Attracting Technical Talent

Modernizing Legacy IT Systems

Accelerating Financial Inclusion with New Data 17

Data Journey Management

Identifying the Most Relevant Data and Where to Begin

“85 percent of the data made by banks is wasted.”

Slow and Arduous Value Extraction

Getting Data Privacy and Consumer Consent Right

Data Fragmentation and Willingness of External Parties to Share

Accelerating Financial Inclusion with New Data 18

Incomplete Digital Proliferation

Accelerating Financial Inclusion with New Data 19

Accelerating Financial Inclusion with New Data 20

Abdul Musoke of the Uganda Communications Commission is in favor of a regulatory sandbox

Accelerating Financial Inclusion with New Data 21

Accelerating Financial Inclusion with New Data 22

Accelerating Financial Inclusion with New Data 23

By acting as an information-sharing mechanism and providing a universal source of truth that is

Accelerating Financial Inclusion with New Data 24

Accelerating Financial Inclusion with New Data 25

Accelerating Financial Inclusion with New Data 26

Accelerating Financial Inclusion with New Data 27

DemystData Matt Hennessy Director, Strategic Accounts

Kabbage Winn Martin Head of Data Strategy

Konfio David Arana Co-Founder and CEO

Konfio Francisco Padilla Sánchez Co-Founder and CTO

Uganda Communications Commission Abdul Musoke Manager, Economic Regulation

* Denotes IIF member.

Accelerating Financial Inclusion with New Data 28

Caribou Digital Publishing, 2017.

Accion, March 2017.

Accion, March 2017.

Accelerating Financial Inclusion with New Data 29