this forum is an initiative of sapient consulting automation process software gartner hype cycle -...

51
© REDEFINING PERSPECTIVES, 2017

Upload: duongthien

Post on 03-May-2018

225 views

Category:

Documents


2 download

TRANSCRIPT

© R

ED

EF

ININ

G P

ER

SP

EC

TIV

ES

, 2

01

7

This forum is an initiative of Sapient Consulting

SESSION 1

SETTING THE CONTEXT

RAJA RAMANVice President- Technology, Sapient Consulting

Raja is a Vice President at Sapient Consulting, Gurgaon.

He has 18 years of experience in developing and

supporting technology applications for various clients. He

currently leads the full stack practice for Sapient

Consulting. Additionally he works as a technical architect

over various projects.

He has a keen interest in trading and actively pursues

technologies that can enable better trading with interesting

algorithms.

Right here right now !!

SESSION 2

BUILDING THE BUSINESS CASE ON RPA

POOJA MALHOTRASenior Manager and RPA Specialist, Sapient Consulting

Pooja brings close to 13 years of experience delivering

successful and transformational solutions to the energy

and financial industry. Specializing in advanced analytics,

she has helped solve some of the most complex industry

problems through the use of innovative technologies like

Machine Learning, Linear Optimization, RPA and other

advance analytic techniques.

Pooja is a part of the RPA and AI practice leadership team

at Sapient Consulting

$7.5TIn rule based work

Source: McKinsey

4MRobots doing

Office Work by

2021Source: Forrester

50-75%Estimated efficiency

improvementSource: Hackett Group

30-45%60% of all activities have

30% or more activities that

can be automated todaySource: McKinsey

$8.75BProjected market size

of robotic solutions by 2024

Source:Grand View

Find the Talent

Gather the

candidate profile

requirement

Search Job Portal Consolidate data in excel

WHAT IS ROBOTIC PROCESS AUTOMATION?RPA is an automation technology that can mimic the series of human actions taken to perform a task in system. The

software or the bot can capture and interpret existing applications for processing a transaction, manipulating data,

triggering responses and cross application communicating similar to as a human does.

Interactswith system as

human doEfficacies

in processes that are

repetitive,

high frequency and

error prone

Technologythat digitizes the white collar

industry

Automatesany rule based

manual task

NonintrusiveNo- code automation for

any business process

involving any application

Easyto train, clone and

deploy

“Augment the organization’s intellectual and creative potential”

What is Robotic Process Automation?

Benefits

Scalable Speed AccurateEfficiency

Insights

Analytics &

Optimization

Process Efficiency: up to 80%

Operational Cost : up to 40%

Dramatic increase in Speed

Better decision quality

Increased revenue

Increased margins

Faster than human pace

Instant ramp up and down

Productivity Value Add

Not susceptible to fatigue / human errors

Manual Effort Reduction

24/7x365

Provide data not previously available

Capable of optimizing based on historyFaster than

human paceInstant ramp up and down

Not susceptible to fatigue / human errors

Manual Effort Reduction

24/7x365

Provide data not previously available

Capable of optimizing based on history

2 0 1 7

Robotic Automation Process Software

Gartner Hype Cycle - RPA

Why RPA

Market Force

Data

Insight

Technology

High ROI

Fast Implementation

Evolution of Automation

Featu

res a

nd

Eff

icie

ncie

s

Time

~1950’

s

~2000 ~2010 Now

• Platform Agnostic

• Complex Rules

• User Activity Replication

• Unstructured Data Processing

• Optical Character Recognition

• Control Center

• Monitoring , Logging & Audit

• Security & Governance Reporting & Analytics

Robotic Process

Automation+• Natural Language

• Processing and Generation

• Self Scripting

• Bots Cognitive

• Computing Machine

• Learning Chat bots

• Cognitive Virtual

• Assistants

• Voice Assistants

Intelligent Process

Automation+

• Batch Jobs & User Action Automation

• Capture Results

• Rules Engine

• Desktop Recording

• Macros

Process

Automation

TECHNOLOGY

Connects to system, usually through the front end. Focused on high volume and rule-based activities with

digital triggers.

E.g.-1: Manual work in existing applications: Contract generated from the trade management systems did not conform to the

requirements for different BUs. RPA bots automated the process of contract generation, based on the rules used by a contract

analyst per BU.

E.g.-2: Repetitive rule-based business processes: Close of business day activities involved in executing and monitoring a

series of processes across multiple systems. RPA automated the entire process.

Intelligent

Process

Automation

Action Automation“fingers on keys / swivel chair”

Analysis Automation“interpret information

e.g. free form comments”

Decision Automation“decision and recommendation”

Learning RobotsSelf-scripting workflows

Includes automated interpretation of information like image processing and natural language processing.

• E.g.-1: Ingestion of information from multiple non-standardized sources: Market data came from different sources in

structured format (data from web, feeds) or unstructured data (brokers, emails, word, pdf attachments). RPA bot extracted and

transformed data into standard information template and keyed in downstream systems.

Automate a decision or make a recommendation to a user

• E.g.-1: Manual review & decisions: Email invoices from a client. The analyst compared the invoiced amount to the amount in a

trading system. If the difference in value was less than a specific dollar amount, the analyst created a cash adjustment. RPA

automated the entire end-to-end process.

Observe patterns, process historical data to train model continuously

• E.g.-1: Identify and learn repetitive tasks: Many users unknowingly work on a number of tasks that they do not realize are

methodical and repetitive. RPA monitored user activities and identified tasks that it can start doing for the user. It automatically

scripted the tasks and requested permission to schedule it on an ongoing basis.

Intelligent Automation Components

TECHNOLOGYIntelligent Automation Components

• RPA with integrated cognitive service from leading vendors like IBM, Microsoft , ABBY ,

• Ingestion of information from non structured sources – Trader gathering all the news on set of

companies/sectors/commodities that he/she is trading

• RPA integrated with deep learning ability

• Process automation opportunity analysis, process deviation and exceptions , process automation

simulations

• Bot leave a trail – key stokes , Screen used , data entered, application accessed , business processes

We are at the beginning of the most

transformative revolution ever

Industry 1.0

Agriculture

Industry 2.0

Industrial

Industry 3.0

Information Age

INSIGHT

Businesses are building connected intelligence experiences

ERADICATE

REPETITIVE TASKS

Liberate employees

from repetitive

knowledge tasks that

require little cognitive

effort beyond 'common

sense,' to drive both

cost savings and

accuracy

Intelligent Process Automation

Virtual Representative

AUTOMATE INSIGHT

GENERATION

Extract previously

unknown insights from

structured and

unstructured data

automatically, to identify

and act on recommend

actions for improved

competitive advantage

Predictive Patterns Causal

Connections

AMPLIFY HUMAN

INTELLIGENCE

Enhance the intelligence

of humans by providing

contextual knowledge

and support, to help

customers and

employees perform tasks

in increasingly simpler

and more effective ways

Expert Assistant Intelligent

Search

Industry Spectrum

Industry

Banking &

Financial Services

Insurance

Healthcare

Manufacturing

Hi-Tech &

Telecom

Energy & Utilities

Function F&A

Accounts receivable,

Account payable,

General ledger

Procurement

Invoice processing

requisition-to-purchase

order

Human resource

Payroll, hiring,

candidate

management

Contact center

Customer Services

Industry-specific

processes

Lost Card Reporting

Fraud Claim Recover

Client Onboarding

Processing Claims

Billing Cycles

Medical Report Analysis

Procurement process

Inventory management

First Call Resolution

IT Infra Services

Counterparty Reconcillation

Customer Invoicing

Potential for RPA

Illustrative processes with

higher potential

Low High

Business Context for Trading

Front Office Back OfficeMiddle Office

Invoicing and Payments

Inventory Valuation

Cash Settlement

Ledger Reconciliation

Accrual Management

Exposure Management

P/L Reporting (YTD, LTD)

Credit Risk Management

Risk Metrics Reporting

VaR reporting

New Product Approval

Portfolio Management

Performance Management

Broker Reconciliation

Intelligent Process Automation

Partial Automation

Nominations & SchedulingFee/Expenses Management

Inventory Reconciliation

Tax Management

Exception ManagementException Management Risk Policy

Compliance Reporting

Market Risk Management

Write on/ Off

Monitor DOA

Corporate Earning

Deal Amendments

Position Management

Asset Management

Retail Demand

Counterparty Reporting

Account close process

Netting Statement

Pre-Trade / Trading

Not Automated

Physical Deal Capture

Hedging and Risk Management Imbalance Valuation

Controls Management

Reference Data Management

Contracts/Agreement

Trader Checkout Estimate for Actual

ISO Settlements

Financial Deal Capture

Logistics Deal Capture

Confirmations

Process can be partially automated to reduce

human effort and for value added insights

Nature of business process requires human

interaction and decision making based on subjective factors

Process can be automated using

intelligent rules and self learning algorithms

Client Case Studies

90%Reduction in duration

80%Reduction in manual labor

ZeroError

0Day lag to resolve errors

Trade ValidationTrading data from the in-house trading system is compared against the data downloaded from the clearing broker website.

Based on predefined rules the trade between the trading system and the broker data is validated. For the mismatched trades

the mail is sent to the traders and the snapshots of broker data is archived for audit.

Client Case Studies

90%Reduction in duration

90%Reduction in manual labor

95%Reduction in errors

3Xmarket coverage

Cargo Market Data IntelligenceBrokers send the fixture market intelligence data to the commercial analyst via email. Based on rules defined

per broker this data is then interpreted, enriched and entered into the market data intelligence management

system.

Client Case Studies

85%Reduction in duration

70%Reduction in manual labor

90%Reduction in errors

1Standardized process

(5 processes to 1)

Contract GenerationContract analyst review and validate the deal and contracts generated out of the ETRM system based on predefined rules.

The contract language is amended based on certain rules .

Thinking Points

• Legacy IT estate ,costly to support and maintain ,making significant investments in transforming or replacing their underlying platform not

viable

• Complex IT integration projects requiring lengthier implementation cycles

• Bridging the gaps between the Business Application features and business processes requirements

• Data entry and the data rekeying jobs with its automated tools and techniques

• Repetitive jobs of data assembling, formatting tasks that use a set of rules to follow the process

• Every type of computer-aided process that uses a set of protocols for its operations

• Regulatory reporting , reducing the risk of inaccurate data and with improvised analytics and higher data accuracy

Lifecycle

Define Process Implement

Productionize

Delivery Approach

Infrastructure, Security and Governance

Operational Support

SESSION 3

IMPLEMENTING RPA AND TECHNOLOGY BACKBONE

ARUN NAGPALSenior Manager and RPA Delivery Lead, Sapient Consulting

Arun brings over 14 years of experience in Business

and IT consulting industry for Fortune 500 companies

with a focus on services, telecom and energy

domains. He has a proven track record in building and

leading diverse cross-functional teams in process

improvement, business-technology alignment and IT

system delivery programs for clients across the

world.

Arun is currently part of AI practice team and is

responsible for managing and delivering RPA

solutions.

Robotic Process Automation – What’s Next

Hello WorldMy first personal bot and my love for pizzas

BOT gathers information required to order a pizza

from user

BOT displays the gathered information to

User and asks for confirmation

If confirmed, BOT orders pizza via RPA

BOT Displays order price once order is done.

Technologies at PlayMy first personal bot and my love for pizzas

UI

• HTML 5

• JavaScript

• CSS3

• AngularJS

Voice Processing

• Google Speech Recognition

Text Processing

• Google Dialogflow

Service

• Java Spring

Action Automation

• UiPath- RPA

Design: 0.5 Person Days

Implementation: 2 Person Days

Technologies Fueling the Robotic Process Automation

Process

mining

Visualization

Data

warehouse

Natural Language

Generation

Reporting

Process Analytics

Dynamic/Elastic

Search

BPMS

Text Analytics Natural Language

Processing

OCR /ICR

Machine Learning

Chatbots

Process Automation

Data IntegrationRecord Replays

OrchestratorCloud

Virtualization

Workflow Manager

Process Control

RPA Vendor Landscape

(Sou

rce: Forrester W

ave™

: Ro

bo

tic Pro

cess Au

tom

atio

n, Q

1 ’1

7)

Current Offering:• Bot development and core functions

• Control room, reporting, and Resilience

• RPA analytics

• Architecture

• Breadth of use case

• Deployment, governance, and security

Market Presence:• Installed customer base

• Revenue growth

• Services

• Number of employees

• Technology partners

Strategy:• Vision

• Innovation Approach

• Planned Enhancements

• Execution Strategy

Business Context for Trading

Front Office Back OfficeMiddle Office

Invoicing and Payments

Inventory Valuation

Cash Settlement

Ledger Reconciliation

Accrual Management

Exposure Management

P/L Reporting (YTD, LTD)

Credit Risk Management

Risk Metrics Reporting

VaR reporting

New Product Approval

Portfolio Management

Performance Management

Broker Reconciliation

Intelligent Process Automation

Partial Automation

Nominations & SchedulingFee/Expenses Management

Inventory Reconciliation

Tax Management

Exception ManagementException Management Risk Policy

Compliance Reporting

Market Risk Management

Write on/ Off

Monitor DOA

Corporate Earning

Deal Amendments

Position Management

Asset Management

Retail Demand

Counterparty Reporting

Account close process

Netting Statement

Pre-Trade / Trading

Not Automated

Physical Deal Capture

Hedging and Risk Management Imbalance Valuation

Controls Management

Reference Data Management

Contracts/Agreement

Trader Checkout Estimate for Actual

ISO Settlements

Financial Deal Capture

Logistics Deal Capture

Confirmations

Process can be partially automated to reduce

human effort and for value added insights

Nature of business process requires human

interaction and decision making based on subjective factors

Process can be automated using

intelligent rules and self learning algorithms

RPA Journey

Initialize Industrialize Adoption

RPA Journey

Initialize Industrialize Adoption

• Select one/two business processes based on automation

potential and impact

• Narrow down RPA Vendors

• Work with Business/IT to develop a POC

• Validate against expected benefits

Pilot

Explore

RPA Journey

Initialize Industrialize Adoption

• Stakeholder orientation and problem assessment

• Assess, rank and select business processes

based on automation potential and impact

• Finalize long term Vendor

• Create Automation Roadmap

• Define governance and control framework

• Perform process optimization

• Create detail designs for scalability and

flexibility

• Design infrastructure requirements

• Design for Security considerations

• Build and test process

(configurations, train the BOT)

• Create deployment plan

• Create and execute change

management plan

• Solution deployment and End of Life

Support

• Performance measurement

• Key Learning and inclusions

Strategy and Assessmentv Plan, Design and Architecture Implementation Deployment

Governance

Implement

RPA Journey

Initialize Industrialize Adoption

Ongoing Support & Considerations

• Ongoing support for RPA applications

• Continuous learning and incorporation

• Follow the S Curve prioritization Framework (point

of diminishing returns)

Low Value High

Lo

wH

igh

Co

mp

lexit

y

Priority 1Priority 2Priority 3

Steady State

RPA Success FactorsSKILL STRUCTURE

Business Application Knowledge• Application functionality / logic

• Application architecture

• Interfaces (e.g. biztalk, file based)

• Master data

Support and Maintenance• Centralized controls

• Support for changes to application and business

processes

IT & Risk Control• Strategy for reliability, scalability and business

continuity

• Standards for security, risk ,data integrity and quality

COE• Strategy and vision

• Governance and Prioritization framework

• Delivery and service models

Business Expertise• Industry understanding

• Business process optimization with required

controls and visibility

• Business scenario and exception

management

RPA and Cognitive Expertise• Right blend of RPA and AI Expertise

• Best in class OCR and ICR knowledge for hand written

and machine generated character recognition

• NLP and Text Analytics Expertise

RPA is a journey and NOT a project

Align experience to users’ needs,

environment and journey…design the

service model of the future and execute in

agile way to get the organization to future

focusing on top priorities in alignment to

the corporate strategy and with constant

feedback thus delivering high impact and

quality automations

Inspiring vision with a sprint ahead on the most

compelling solutions

Redefine the future relationship between

technology and people, giving the uninteresting

deskilled jobs to the bots and augmenting the

creative potential

Variants: 36/Prime, Spark

Create an internal capability to test out innovation

and transformation opportunities; rapid

prototyping and market testing to quickly find the

path to greatest value and differentiation

Capability to process vast amounts of data, learn

& adapt on the go and automate complex

business workflows autonomously , dramatically

elevating the business value delivered.

North Star Vision Service DesignInnovation

[Hub, Center, COE, …]

Case Studies

Market Intelligence processing

Read market intel

information from broker

(pdf, excel, email, html)

Transform unstructured data

to structured

Apply Business rules and

data enrichmentMap data with IT system

Send notifications post data

update

Benefits:

• Analyst are now saving 14 hours daily which help them in

managing fixtures more effectively

• Real time market data updates across globe

• Processing of 300+ broker files monthly

• Handling of 50+ variations of Broker Fixture reports

• Short implementation cycles :

• Phase 1 – 5 weeks

• Phase 2 – 4 Weeks

Key Features

• Unstructured to structured data conversion

• Predict incomplete market intel based on historical trend

• Intelligent data mapping

Movement Ticket processing

Benefits

• Schedulers are now saving 6-8 hours daily which is

helping BOL information to reach downstream systems

almost real time.

• Short implementation cycle of 10 weeks

Key Features

• Extraction of movement ticket info from image based

pdf (text as well as OCR)

• Data mapping with reference data

• Intelligent data mapping

Read movement ticket

information from shipper

(pdf, html)

Parse image based BOL

using OCR

Apply Business rules and

data enrichmentMap data with IT system

Send notifications post data

update

Contracts Transformation and Processing

Search for New Contracts

generated by ApplicationExtract information from DB

and application

Apply business rules on

Contract Document

Save Transformed Contract

Document for review by

Contract Analyst

Document

Transformation

Check for Emails from

Contract AnalystRead email and process

contracts sent by analyst

Send Contract

to Counterparty

Capture Delivery

Notification, process &

Update status in Application

Communication

Management

Benefits:

• Transformation of 6000 Contracts yearly (which were earlier done manually)

will be done by BOT

• Saves 700 person days of effort yearly spent on contract transformation

• Implementation cycle of 16 weeks

Key Features

• Smartly handles transformation of contracts on multiple variations

• Establishing effective and reliable communication mechanism with third party

via emails along with maintaining communication history and acting on

acknowledgement

• Smart Data Mapping mechanism

Volume to weight

conversion data

UAT!!!

Werner vogels: “everything fails all the time”

RPA Final value

Volume to weight

conversion data

Macros / Formulae

Prod

Werner vogels: “everything fails all the time”

RPA couldn’t understand macros and formulae System crashed !!!

Werner vogels: “everything fails all the time”

• Exception handling needs to be top notch

• There should be a kill switch

• Test with prod data for UAT

• Scope needs to be clearly documented and managed tightly with business

• Team structure should have right balance of skills – business, technology and management

• Excel may have volume limitations

Takeaways

• This isn’t rocket science – download, learn and use !!

• We’re at the very beginning – now is the time to start riding on the wave!

• Start conversations in your teams today – about use cases, pilots

• Automation is a mindset – carry it everywhere you go!

Thank You