Knowledge Graphs for Financial Services 1

Knowledge Graphs for

Financial Services
The path to unlock new insights
from your data
Knowledge Graphs for Financial Services 2

“Logic will get you from A to B.

Imagination will take you everywhere.”
Albert Einstein
Knowledge Graphs for Financial Services 3

Preface……………………………………………………………………………………….….……………. 4

What are Knowledge Graphs?…………………………………………………….…………….…. 5

Use Cases

360° View of Risk & Value ……………………….……….………………..…….……….….6

Compliance Management ………………………………..……………………..……..……7

Data Lineage & Metadata Management………………………………….……..…….8

Fraud Detection & Financial Crime Analytics….………………….….……..………9

Recommender Systems & Conversational AI….…………….…………………….10

Three Pillars of our Knowledge Graph Services….………….……………..…….………12

Conclusion….…….………………………………………………………………….…………..……..…13

Contact….…………………………………………………..….……………………………………………14
Knowledge Graphs for Financial Services 4

Preface
Today, companies are relying more and more on Artificial Intelligence (AI) applications
in their day-to-day decision making. Without context, these AI applications may never
reach their full potential as reliable solutions to address complex business needs.
Knowledge Graphs enable machines to incorporate human expertise for making
meaningful decisions and bring context to AI applications.

The new wave of AI is focused This enables users to quickly access two closely
connected objects and adapt the connection based
on hybrid intelligence that
on the context, for example both a small enterprise
means learning (data) fused and a private individual as customers of the same
with reasoning (knowledge). organization may be given the same risk rating due to
closely matching activities.
Connecting datasets in a meaningful way is
strategic for every business as it enables Major institutions are commonly faced with
decision makers, users, and (above all) thousands of isolated “data silos”, hence facing an
computers gain context on the existing information overload challenge. Knowledge Graphs
knowledge of an organization. For enterprises can serve as a hub of integrated knowledge by
to stay competitive in the current knowledge processing disparate sources and extracting atomic
economy, it is crucial to manage knowledge units of knowledge from diverse datasets. Moreover,
efficiently and be ready for changes that might they provide contextual adaptation, and reasoning
serve either as a threat or opportunity to their capabilities that go beyond the conventional machine
business. That’s where Knowledge Graphs learning approaches.
come into play and why tech giants like
Amazon, Facebook, Microsoft and Google
have invested millions of dollars to create The impact of Knowledge Graphs in financial services
their own Knowledge Graphs. is just in its inception where the role of knowledge
scientists to build bridges between business
Knowledge Graphs allow for processing and requirements, questions and data is becoming more
representation of data and knowledge in a and more important. In this article, we highlight some
format which is very close to the way a human example use cases of Knowledge Graphs to
brain processes and stores information. demonstrate the value of Knowledge Graphs for
data-driven businesses.
Knowledge Graphs for Financial Services 5

What are Knowledge Graphs?

A Knowledge Graph is a means to connect and represent knowledge in an area of interest
using a graph-like structure. It is typically built on top of existing databases to link data
together at web-scale, combining both structured and/or unstructured information. As
opposed to the more commonly used relational data models, a graph model is built as a
collection of concepts or entities and the relationships between them.

Thinking, Fast and Slow

You could see Knowledge Graphs as the brain of a company with two systems of thinking:

System 1 (Thinking fast) System 2 (Thinking slow)

that is intuition-based, that is logic-based, effortful
automatic and error-prone and capable of making
to quickly learn from the complex yet explainable
past observations. decisions.

This system of Knowledge This system of Knowledge

Graphs is powered by Graphs is powered by
statistical AI approaches symbolic AI approaches
such as Machine Learning. such as Knowledge
Representation & Reasoning.
1 2

As an example of system 2, let’s read the following sentence:

“A red truck with a siren and a ladder attached to its side rushed in the city”.
What do you understand from this sentence? As a human you might already think that the
vehicle was most likely a fire truck; or there was probably a fire somewhere in the city; and
many more logical conclusions. However, as a machine it would be very hard to make all those
connections and come up with the above reasoning due to the lack of background and
common-sense knowledge.

A Knowledge Graph brings together Machine Learning and Graph technologies to give AI the context it
needs. Knowledge Graphs represent a complex network of information in a meaningful way (in a similar
way to human intelligence) by integrating data from a wide range of data silos and incorporating
learning and reasoning.
Knowledge Graphs for Financial Services 6
Use Cases

360° View of Risk & Value

Knowledge Graphs allow companies to connect the dots in their business and see the
big picture for evaluating risk and value.

The challenge for financial In order to seamlessly integrate

institutions seeking competitive large enterprise data silos, Applications
advantages in the market is not a companies do not only need to
lack of information. It is an over- unify various data models and - Know Your Customer
supply of information from formats, but also need to (KYC)
diverse sources with complex harmonize data based on its - Due Diligence
and vast amounts of data. meaning. - Investment Research
- Insurance Underwriting &
Claim Processing
Big Data technologies have Knowledge Graphs enable
- Commercial Real-Estate
empowered companies to ingest companies to semantically
substantial volumes of data and integrate diverse data and draw
store it in multiple data silos. connections at an unprecedented
Bridging these data silos brings scale. They also allow users to
limitless opportunities for connect external sources of data
companies to assess risk and efficiently, regardless of their
value across multiple domains. underlying data formats and
models. Example

Imagine an international bank that wants to

use their customer data for identification of
high-risk customers to avoid sanctions.
This would require a deep dive into the
customer data, understand the different
attributes and create a risk classification
while ensuring links to the relevant
sources of information such as financial
transactions, customer’s family
relationships, as well as regulations.

Using a Knowledge Graph, the bank can

see all the customers information, risk
dimensions, and regulations that are
relevant for their specific goal, linked
together based on their meaning, in one
place for deep analysis.

Knowledge Graphs empower business

users to see the whole story.
Knowledge Graphs for Financial Services 7
Use Cases

Compliance Management
Interlinking company data with various compliance data sources such as internal
policies, legislations, standards and other private contracts in a Knowledge Graph
enables companies to better manage compliance.

Applications
Finding ways to remain compliant Knowledge Graphs leverage
with ever-changing and growing the power of semantic
universe of regulations, policies and technologies to not only unify - Regulatory Reporting
internal contracts has become one of and interlink various sources - Trade Surveillance
the most visible challenges of today’s of compliance data, but also - Insider Trading Surveillance
companies. to apply complex rules and
patterns for (semi-)
Financial compliance requires automated compliance
companies to maintain sophisticated monitoring.
customer screening and transaction
surveillance systems that pose data In order to optimize
quality and data availability compliance check, Knowledge
challenges. Current compliance Graphs combine contextual
systems are focusing mainly on data domain knowledge with
collection and data consolidation, Natural Language Processing
leaving less time for in-depth (NLP) and Machine Learning. Example
analysis.
Imagine a local company seeking to
branch out to other countries and
markets. In order to trade abroad and to
localize its products and services to other
countries, the company needs to take into
account the legal barriers applicable to the
target market. Dealing with legal and
regulatory compliance data is a
cumbersome task that needs to access
huge volumes of digital compliance
documents and integrate them with
internal company data.

Utilizing a Knowledge Graph allows this

company to efficiently identify relevant
regulations, link its data to those
regulations and to define patterns for
automatic compliance check.

Knowledge Graphs allow companies to bring together different sources of

compliance in a meaningful way and apply automated compliance checks.
Knowledge Graphs for Financial Services 8
Use Cases

Data Lineage & Metadata Management

Knowledge Graphs allow companies to trace their data journey in the right context and
improve the quality of their data flows.

A common issue today’s data- The combination of detailed

driven companies face is tracking metadata and relationships Applications
data throughout its lifecycle. between data lifecycle phases
Whether acquired externally or results in a semantic data layer.
- Risk Data Aggregation &
generated from internal services, A semantic data layer (also known
Reporting
data needs to be tracked as it is as data fabric) enables both data
- Master Data Management
utilized, manipulated, and experts and business - Data Migration for
transformed throughout the stakeholders to take advantage of Consolidating Mortgage
organization. any data asset to which they have Systems or On-boarding
access. New Data Feeds

The challenge is that complex

data neither explains itself nor its Knowledge Graphs provide a
journey. And, data is only semantic layer with full view of
valuable in the context of what it the data lifecycle all the way from
means to its stakeholders. Thus, the business to the most
data must be described with technical components.
relevant metadata and be Example
organized in a way that reflects
Imagine a company that is in a merger
its meaning and fitness to use. process and needs to migrate all the
Customer Relationship Management
(CRM) data from the acquired company to
its current data lake. To ensure the quality
of data, the data managers would like to
know the data changes over time and
know what manipulations were applied to
source data that resulted in its current
form.

Using a Knowledge Graph solution, the

company can get an overview of the data
lifecycle together with the sources used
and all their dependencies to allow data
managers make better decisions for data
integration and quality improvement.

Knowledge Graphs allow to keep track of your data

journey meaningful to its different stakeholders.
Knowledge Graphs for Financial Services 9
Use Cases

Fraud Detection & Financial Crime Analytics

Augmented analytics provided by Knowledge Graphs & Machine Learning enables
companies to identify fraudulent patterns and investigate specific criminal links.

Today’s financial institutions are Trying to scale these large

contending with measures for Anti- queries to return real-time data Applications
Money Laundering (AML), Suspicious is impossible in most cases as
Activity Reports (SAR), counterfeiting, the number of connections and - Spotting fraud, which applies
as well as synthetic, first-party, and patterns grow. to fraudulent transactions
and applications in banking,
card-not-present fraud.
benefits fraud in
Knowledge Graphs empowered
government, tax fraud,
Frequently, fraud happens relatively by machine learning and applications and claims
quickly and with a convoluted reasoning capabilities allow fraud in insurance.
network of transactions that mix and companies to better identify - Anti Money Laundering
blend the identity and funds through fraudulent patterns by
multiple channels that are hard to traversing many hops on very
detect. large amounts of
interconnected data in
The traditional relational database real-time.
systems for fraud detection require Example
a lot of complex series of joins that
Imagine an insurance company that wants
are often complex to build.
to detect fraud schemes. Insurance fraud
is usually from complex networks that are
difficult to detect for insurance and
financial institutions. Fraudsters usually
forge fake identities, file several claims and
cash the insurance checks. Creating fake
identities requires forgery of personal
information (social security numbers,
addresses, credit cards, etc), to submit to
insurance companies to become a
customer. Fraudsters often recycle this
data to create several fake identities.

Using a Knowledge Graph for entity

resolution and a visual analytics interface,
the company can easily detect clusters of
fake claims.

Knowledge Graphs allow monitoring, categorizing, and

predicting potential points of threats.
Knowledge Graphs for Financial Services 10
Use Cases

Recommender Systems and Conversational AI

Knowledge Graphs enable fact-based recommendations that accumulate contextual
knowledge with each conversation.

The goal of a recommender system A Knowledge Graph is built as a

is to deal with an overload of large semantic network of Applications
information by filtering vital entities and their attributes.
information fragments out of a Therefore, it allows finding the - Conversational Commerce
large amount of information. best matching entities based on - Customer Service Chatbots
semantic similarities between - Product Recommendations
- Enterprise Search & Content
In order to provide the entities. Knowledge Graphs
Classification
recommendations that match also allow enriching the context
user’s intent, preferences and of data by incorporating
interest, a recommender system domain-specific knowledge
needs to know about the context vocabularies, taxonomies or
of conversation and must be ontologies.
capable of not only looking at
structural similarity between items The personalized experience
but also their semantic similarity. provided by Knowledge Graphs
Example
For example the word “balance” enables conversational banking
might convey different meanings tools to more effectively interact Imagine a big retail company that wants to
when used in a financial with customers on their build a conversational commerce platform
conversation as opposed to a banking needs. where its customers could interact with a
bot similar to the way they interact with a
medical context.
salesperson in their shop. The company
has more than 10,000 product categories
and more than 100,000 attributes such as
brand, colour, and any other attribute
about an object that customers can buy.
They failed when they tried a rules-based
engine to support such a huge catalog,
with more than million items in inventory.

Using a Knowledge Graph solution, the

company can store relationships between
various sources of information such as
products, customer interests, and
purchase history and quickly query it to
make recommendations that are
personalized and relevant to customers.

Knowledge Graphs provide context-driven recommendations.

Knowledge Graphs for Financial Services 11

A Knowledge Graph is worth

a thousand words.
Knowledge Graphs for Financial Services 12

Three Pillars of our Knowledge Graph Services

The effective design and implementation of Knowledge Graphs requires 3 components:

a) bridging diverse data silos regardless of data formats, serializations,
conceptualizations, and technology ecosystems, b) investigating interconnected data to
find out insightful patterns and c) deriving context-relevant knowledge from the large
amounts of integrated data.

This component deals with extracting This component deals with different kinds of
knowledge from data. graph analysis to identify patter ns in
Machine reasoning methods supported by interconnected data. Methods such as path
semantic technology standards such as RDFS, analysis to determine the shortest distance
OWL and SHACL allow our solution to derive between nodes, connectivity analysis to
new facts from data. Additionally, advanced AI determine weak/strong nodes, community
machine learning techniques such as Graph analysis to find groups of related nodes and
Convolutional Networks (GCN) are applied to centrality analysis to identify the most
graph data to predict new links and learn influential nodes in the graph are applied to the
complex graph patterns. underlying graph.

This component deals with data distribution, harmonization,

integration and storage at scale. Linked Data standards such as
HTTP, URIs and RDF data model are used to represent data in a
single interchangeable format that is both understandable by
machines and humans.
Additionally, multi-model graph databases are incorporated to
support multiple data models against a single, integrated backend.
Knowledge Graphs for Financial Services 13

Conclusion

Knowledge is every company's most valuable asset, In this report, we shortly touched upon some
however it is scattered across different systems and example use cases of Knowledge Graphs for
human minds. This decentralized nature of financial services. Knowledge Graphs are already
knowledge makes it hard to grasp and increases the in use in industries such as public sector,
challenges for organizations. The key to integrating healthcare & life sciences, energy, resources &
knowledge efficiently among various systems and industrial sector, technology & media sector, etc.
human users is to provide knowledge representation Some other prominent use cases can be found in
and reasoning in a machine-readable form. the following domains:

Creating a Knowledge Graph with semantic - Personalized health

description of information context allows users to - Logistics & supply chain management
access a machine-readable representation of - Industry 4.0 in automation & manufacturing*
complex interdependencies that form a real-world - IoT data integration & management
model of the knowledge domain. - 360° view of criminals cases
- Network / IT/ cloud resource optimization and
maintenance
- Traffic management
- Geospatial analytics
- Performing grid and network quality of service
- Preventing cyber attacks

*If you are interested to know more about

applications of Knowledge Graphs in industry 4.0,
you can read the following Deloitte article:

Wisdom of Enterprise
Knowledge Graphs:
The path to collective
intelligence within
your company
Knowledge Graphs for Financial Services 14

Contact
Contact

Dr. Ali Khalili Janvier Jessurun

Junior Manager Risk Advisory Senior Manager Risk Advisory
[email protected] [email protected]
+31650074011 +31683339436

Stuti Oberoi Marko van Zwam

Senior Consultant Risk Advisory Partner Risk Advisory
[email protected] [email protected]
+31650055356 +31621272904

Stuti Oberoi
Marco van Zwam
Eminence Lead
Sponsoring Partner

Janvier Jessurun
Business Lead
Ali Khalili
Technical Lead

Alexander Easton
Data Scientist
Ilayda Karatas
Jeroen Bijman
Business Analyst
Data Engineer
Marit Beerepoot
Data Scientist

Sara Fernandez
Data Scientist

The Knowledge Graph Team

Knowledge Graphs for Financial Services 15
Knowledge Graphs for Financial Services 16

Deloitte refers to one or more of Deloitte Touche Tohmatsu Limited ("DTTL"), its global
network of member firms and their related entities. DTTL (also referred to as "Deloitte
Global") and each of its member firms are legally separate and independent entities. DTTL
does not provide services to clients. Please see www.deloitte.nl/about to learn more.

Deloitte is a leading global provider of audit and assurance, consulting, financial advisory, risk
advisory, tax and related services. Our network of member firms in more than 150 countries
and territories serves four out of five Fortune Global 500® companies. Learn how Deloitte’s
approximately 286,000 people make an impact that matters at www.deloitte.nl.

This communication contains general information only, and none of Deloitte Touche
Tohmatsu Limited, its member firms or their related entities (collectively, the “Deloitte
network”) is, by means of this communication, rendering professional advice or services.
Before making any decision or taking any action that may affect your finances or your
business, you should consult a qualified professional adviser. No entity in the Deloitte
network shall be responsible for any loss whatsoever sustained by any person who relies on
this communication.

© 2020 Deloitte The Netherlands