How can AI & Automation make your business processes intelligent

1©2019 Mindfields Global. All rights reserved. MindSpeaks Webinar Series
How can AI and Automation make your Business
Processes Intelligent?
MINDSPEAKS
WEBINAR
SERIES
Hosted by:
Mohit Sharma
Managing Director, Mindfields
Presented by:
Praveen Kombial & Jasdeep Singh Kaler
EdgeVerve

About
Mindfields drives innovation and excellence by leveraging disrupting technologies to
optimize business processes that enable its clients to 'Grow for tomorrow’. They are a
vendor-agnostic automation and artificial intelligence services firm. Gartner and HfS
Research has positioned Mindfields as a leading, niche Robotics Process Automation
advisory (RPA) firm and we are globally recognized as a thought leader in this space.
They have ranked among the top 20 fastest growing Australian firms in the Deloitte
Fast 500TM.
EdgeVerve Systems, a wholly-owned subsidiary of Infosys, develops innovative
software products and offers them on premise or as cloud-hosted business platforms.
Their products help businesses develop deeper connections with stakeholders, power
continuous innovation and accelerate growth in the digital world. They power our
clients’ growth in rapidly evolving areas like banking, distributive trade, credit
servicing, customer service and enterprise buying. Today, global corporations across
financial services, insurance, retail and CPG, life sciences, manufacturing, and
telecommunications use EdgeVerve products

Speakers
PRAVEEN KOMBIAL
Global Product Head for
Business
Edgeverve Systems Ltd.
Praveen’s mandate is to leverage Artificial Intelligence
and Emerging technologies to build and launch products
solving for specific business problems for industries
globally. At EdgeVerve, he leads Engineers and Product
managers in building Products addressing industries and
functions such as Financial Services, Insurance,
Consumer Packaged Goods and Supply chains.
Prior to Infosys, Praveen served in various roles
spanning retail finance, corporate leasing and risk
management.
MOHIT SHARMA
Founder and
Managing Director
Mohit has over 20 years of experience working in
Strategy, Corporate Finance, and Risk Management
Solutions for Deloitte, PwC, and EY. Mohit has
authored and published one of the first and most
comprehensive research on Robotics and Process
Automation. Mohit provides advisory services to
leading global financial and non-financial
organization in US, Australia, UK, and India on their
journey for Automation and Artificial Intelligence.
Mohit loves cricket and food, in that order.

Speakers
Jasdeep has spent 22+ years of experience in Product
Development and has led multiple products from idea to
cash, and architected key framework components for
mission critical enterprise class products. He is an
expert in leading large-scale organizational change
management, business transformation and technology
initiatives.
He currently leads multiple products in the areas of
Lending, Insurance and Supply Chain using Blockchain,
AI/ML technologies. His current interest area is how
product startups can be designed for scale.
JASDEEP SINGH KALER
Associate Vice President
& Product Head
Edgeverve Systems Ltd.

Agenda
Session Time
Introduction to webinar topic and speakers 5 minutes
Session 1: Difference between Automation and AI 10 minutes
Session 2: Maturity in Deploying AI in Business 10 minutes
Session 3: Concept of AI Business Applications 10 minutes
Session 4: : Business Applications for Retail Banking and deep dive in one App 15 minutes
Questions and Answers 10 minutes
Polls during the webinar 5 minutes

Session 1: Difference between Automation and AI

What is Artificial Intelligence?
Artificial Intelligence is a revolutionary cluster of technologies and is dependent on the setup of an
existing plethora of smart data. One of the biggest drivers in embracing AI, is that is becoming a
differentiator in how we can do business. An organisation can decide whether to delegate mundane
tasks to AI or utilise the existing workforce to focus on them instead of adding value to high end
decision making. A sophisticated AI project will go through these phases for optimal efficiency gains:
D I S C O V E R P R E D I C T L E A R NV A L I D A T E
A C T I O N

Difference between Automation and Artificial Intelligence
Automation is making a software that is capable of doing things automatically — with or without human
intervention. Artificial Intelligence, however, is a science and engineering of making intelligent
machines, and eventually supersede human behavior and intelligence.
Automation can or can not be based on Artificial Intelligence.
Particulars Artificial Intelligence Automation
Mimics Human behaviour Human actions
Learning capabilities Self learning Predefined
Design Complex Simple
Tools/Platform Mostly cloud based/API integrations Software and web based
Maturity level Early stages Advanced stages

Difference between Machine Learning and Artificial Intelligence
Similar to the digital wave of the early 90’s, the intelligent wave of automation and machine learning is
revolutionising the industry. Machine learning is technically a branch of AI, but it's more specific than
the overall concept.
Deep Learning
Machine
Learning
Artificial
Intelligence
AI is the all-encompassing idea that initially came
about, then followed by Machine Learning that thrived
later, and lastly Deep Learning that is promising to
escalate the advances of AI to another level. The intent
of Machine Learning is to enable machines to learn by
themselves using the provided data and make accurate
predictions.

Drivers of Artificial Intelligence
Worldwide Spending on Cognitive
and Artificial Intelligence Systems
Forecast to Reach $77.6 Billion in
2022
C a p i t a l
A true artificially-intelligent system is
one that can learn on its own. There
are many like the DeepMind, that are
progressing towards that
A p p l i c a t i o n s
With the availability of the plethora of
data, it helps the organisations in
analyzing this and in drawing
meaningful insights from the same
D a t a
Talent is learning not only in brick and
mortar universities, but catching up
online and on e-learning websites
T a l e n t
AI with ethical responsibility is driving
the new trends, and is not just about
making rules to govern intelligent
machines but the humans too
R e s p o n s i b i l i t y
By 2030, global GDP could be $15.7
trillion as a result of AI, making it the
biggest commercial opportunity in
today’s economy.
G r o w t h

EVOLUTION OF AUTOMATION
11
• Highly rule-based
processes
• Large transaction and
processing volumes
Routine and Repetitive
• Requires RPA
configuration to
incorporate decisions
• Processes are
moderately complex
Judgement Based
o • Incorporates advanced
analytical capabilities
• Process unstructured
data from multiple
environments
Analytical Based
• Rise of self learning and
cognitive capabilities
• Automatic detection and
resolution of faults and
errors
AI Based
NOW FUTUREPAST

Session 2 : Maturity in Deploying AI in Business

Infosys AI Research report “Leadership in the Age of AI”
*Source: Autonomous Next
Financial sector will see $1 Trillion in savings from AI Implementation by 2030*
Interestingly the maturity score seems similar across industries……

There are challenges in AI adoption………
Functional, Regional,
Corporate Silos
Lack of Data, Expertise &
Infrastructure To Make AI
Effective
Keeping Up With AI Tech
Explosion
Measuring Business
Outcomes of AI
Governance and Regulatory
Oversight
Lack of Scalability and
Explainability of AI Based
Decisions

Internal Systems & Applications External 3rd Party Systems Social Feeds & Environmental Data
Sources of
Organizational Truth
Knowledge &
Reasoning System
AI Platform
AI Adoption In Financial Services - Reality
Results
We have
implemented a
Customer
Service Chat bot
Me too
We did a pilot in
one geography
We are not sure
what to do
beyond the POC
Challenges
ROI Scalability Compliance Measurable Benefits Technology Obsolescence
Data Science Expertise Existing IT Investments IT and Functional Silos Complex Infra Needs
Our Chatbot is
decommissioned
because of
security issues
BIG DATA OCR NLP CHATBOT ML RPA

The Ying & Yang Of AI For Business
Purposeful AI
Enhance
Experience
Drive
Decisions
++ Revenue
- - Costs

Session 3 : Concept of AI Business Applications

So what are Business Apps?
Business Apps are bite sized,
purpose built Apps
that leverages AI to solve
specific business problems
using structured data from
existing systems of records and
unstructured data from
internal & external data
sources
Each of the Business Apps
come with prebuilt data
models and are quick to
deploy and easy to use.
Value from these Apps
manifests in one/ more of
the following ways:
Driving Decisions Enhance experience Relevant and meaningful
Process Improvements
Improved Business
Metrics

What will our Apps consist of
Data Mgmt. Technologies
Advanced Machine Learning Experiential AI- Bots, Image
Recognition, Speech Analytics
Workflows Responsive User Interfaces Dashboards

Our Strategy is to create Business Applications, that will serve as systems of differentiation and innovation for
Enterprises, by leveraging RPA and AI Capabilities(Nia)
Sourcing&
Procurement
Promotion Effectiveness
Provides comprehensive insights into promotion performance for
identification of performance drivers, leveraging ML techniques, and
enables predictive modeling to create optimized promotions
Perfect Order Measurement
Delivers insights and intelligence to identify factors influencing
Perfect Orders, predictive assessment of risk fulfilment and
proactive alerting for enabling enterprises to meet customer’s delivery
expectations
Channel Data Management
Enables enterprises with efficient and reliable Omni channel
data capture and algorithmic data harmonization through its
data as a service platform
Fraud Detection and Prevention
Leverages Infosys Nia Advanced Machine Learning platform, to detect and prevent
fraud in credit card transactions & insurance claims, thereby minimizing fraud losses
and reducing false positives for the enterprise.
CollectEdge
A data driven intelligence platform powered by machine learning
that helps our clients reduce delinquency rates and improve
recoveries, stay compliant and lower operational cost through
an intuitive experience.
Guided Buying
Enables intelligent routing of procurement requests to appropriate buying
channels and preferred suppliers, demand consolidation for improved savings
and spend classification at source for accurate insights, by leveraging AI/ML
Category Intelligence
Empowers product category teams to proactively manage categories and take best-
in-class decisions powered by data and insights. It transforms sourcing and
procurement decision making through interlinked internal and external data,
embedded analytics, real-time insights, action prompts & expert recommendations
Procurement Insights
Delivers business relevant procurement insights, through automation
of data management, ML-based spend classification, predictive
spend analytics and guided opportunity analysis
Procurement Assistant
Comprehensive AI-based conversational assistant that simplifies
procurement process, while ensuring process compliance and
enhancing user efficiency
Provenance
Nia Provenance is an asset & industry agnostic, blockchain based application that lets
businesses establish traceability of the product, transparency of the processes and
trust-worthiness of the information across the lifecycle of assets (or) products
LoanFulfilment
Leverages the power of AI in driving efficiency and enhancing experience through application enrichment,
automated document verification, personalized risk scoring and to improve of loan processing cycle time
Business Apps - Snapshot

Why Spend Analytics?

Nia Procurement Insights - Solution Overview
Nia Procurement Insights delivers business relevant insights by automating data management activities, classifying spend into right categories
using Machine Learning techniques, generating predictive spend analytics and highlighting opportunities to optimize spend and mitigate risks.
Modules
Efficient Data
Management
AI/ML Spend
Classification
Advanced Spend
Analytics
Supported by Future Proof Big Data Stack
Transactions Master Data Taxonomy
Supplier Market Rules Taxonomy
Right Visibility to Spend
Savings Opportunity Analysis
Root Causes of Non-Compliant Spend
Demand Forecasting and Planning
Approach
Client Base
Knowledge Base

Nia Procurement Insights - Value Proposition
Accelerated Insights for Informed Decisions
Simplify Data
• Cleanse and Unify Data
• Normalize Data
• Enrich Data
Gain Visibility
• Classify Spend
• Derive KPIs
• Configure Dashboards
Identify Opportunities
• Consolidate Vendors
• Aggregate Demand
• Reduce Price
Amplify Savings
• Insights to Actions
• Forecast Demand
• Plan Budget
50-60%
Productivity Gain
8-10%
Savings Opportunity
10-12% Impact in
Vendor Normalization

Session 4 : Business Applications for Retail Banking and deep
dive in App

Business Applications
for Retail Banking #BuiltOnAI
Risk
Modelling
OCR and
Document
Classification
NLP and Text
Analytics
Behavior and
Sentiment
Analysis
Chatbots and
AI based
collaboration
AI based
Decision
Engine
EdgeVerve: AI-enabled Business Applications suite Retail Banking

Consumer
loan or lineof
credit
All payments
are current
Consumer
fails to make
required
payment
Loan Origination Loan Servicing Collections Recovery/Write-Off
0-30
days
30-60
days
60-90
days
90-120
days
Delinquency
Recovery
Sold to debt buyer
Repossessions
Payment due date
reminder
Payment is invoiced
Validate payment
method
Reminder
letters/email
sent to the
borrower
Creditor
sends more
remindersto
the borrower
Creditor sends
warning/
notice e-mail,
SMS, calls to
the borrower
Account
assigned to
internal
collection
agent
Collector
attempts to
rehabilitate/
collect from
the borrower
Account charged-
off by the lender
Assigned to
external collections
Legal Action
Interventions/
Actions
Noearly warningmechanism
to predict/identify delinquent
accounts
Risk segmentation of accounts
basedonrudimentary
statistical modelis inefficient
Inefficientcollectionprocesses
withhighoperatingcostand
call investments
Typical Challenges
Typical Debt Collection Process

CollectEdge - Solution Overview
Customer Insights Engine
Generates 360 degree
view of customer to
identify patterns and
predict behavior
Prediction Engine
Contextualized
Suggestions and
Predictions. Self
Optimizes with Time
CollectEdge
KEY FEATURES
Servicing Collection Recovery
Risk Bucket Prediction
AI Based Prioritized
Queue
Suggestion on
Preventive Measures
AI based
Prioritization
AI Based Resolution
Strategies
Account Level
Personality Insights
Predict Channel and
Time
AI Based Personalized
Communication
Roll Rate Prediction
Over the top solution
providing intelligence to
existing collection process
Drives collection strategies
through its exhaustive
feature set
Not a loan servicing or
core collection platform
Individual ML model for
each of the features
illustrated
Not a data science
platform
Predefined data model
for each loan product
enables easy onboarding
Inbuilt automated data
engineering and model
build for quick value
realization
Simple operationalization
Entire data engineering
and data science is
abstracted with focus on
business outcome

Data Transformation Predictive InsightsFeature Engineering Model BuildData Analysis
PDP Plots
Multicolinearity
Extensible Domain
Data Models
Automated
Transformation Pipeline Pre created Domain
Features
Scatter Plot Matrix
Best in class ML
models like CNN,
RF, GBT, LR etc…
Auto Model/ Auto
tuning
Real time explainability
into individual
predictions made by
the ML model
Deep Feature
Synthesis
Automation In CollectEdge
80% of Data Science effort 20% of Data Science effort

AI models comparison
Limited variables set
Manual effort and time
consuming
Accuracy highly dependent on
skills of modeler
Smart Search for high accuracy
models
Higher accuracy with less
dependency on skill
Time consuming, requiring
scarce resources
More variables, better view of
borrower financials
Explainability vs accuracy
challenge
Operationalization is
challenging
Automated data science for
rapid value delivery
Borrower Propensity +
Behavioural Patterns + Financials
Demystify the ML model with
inbuilt explainability
Simplified operationalization &
self optimization
Inherent Focus on Compliance
Realize value in weeks Monthly
subscription
Statistical Models
Bespoke ML Build
Packaged Business Applications overcomes the challenges presented by
statistical models and bespoke ML builds

How can AI & Automation make your business processes intelligent

More Related Content

What's hot

Similar to How can AI & Automation make your business processes intelligent

More from Mindfields Global

Recently uploaded

How can AI & Automation make your business processes intelligent