De-mystifying Robotic Process Automation

De-mystifying Robotic Process Automation

• 1.
  www.nicsa.org De-Mystifying Robotic Process Automation 2017NICSA Midwest Regional Meeting #NICSAMWRM THANK YOU TO OUR SPONSORS! Platinum Sponsor
• 2.
  www.nicsa.org Agenda Welcome • Lisa Shea,Co-Chair, NICSA Midwest Regional Committee Panelists • Amber Krueger, Moderator, US Bancorp Fund Services, LLC • John Sjosten, Senior Manager, Deloitte & Touche LLP • Randy Guy, Chief Technology Officer, FIS Global • Andy Curtis, Data Analytics Manager, Northern Trust Q&A Roundtable Discussions #NICSAMWRM
• 3.
  www.nicsa.org Robotics & Cognitive Intelligence JohnSjosten, Senior Manager, Deloitte & Touche LLP #NICSAMWRM
• 4.
  www.nicsa.org Digitization of whitecollar jobs via RPA & CI, and advances in data science have sparked the Business 4.0 revolution We are on the cusp of “Business 4.0” BPM Systems Early RPA Early cognitive Widespread RPA Widespread cognitive Ubiquitous global horizontal MLPs Business 4.0 • This revolution redefines what it means to be a professional • RPA has commenced deployment in most large businesses • RPA & Cognitive will be ubiquitous in business by 2020 • Horizontal machine learning platforms (MLPs) become ubiquitous by 2025 Industrial revolution 1-3 4.0 2017 Within 10 years $5bn 2020 RPA market 1 $31bn 2019 Cognitive spending 2 2nd most important strategic priority 3.09 2.59 4.75 3.9 4.14 4.5 5.12 2.05 3.36 3.91 4.24 4.27 4.89 5.05 GBS model Geo. scope Analytics… Increased func.… Func. Proc.… Automation Contin.… Today In ten yearsSource: Deloitte Global Shared Service Survey, 2015 Interest in automation is increasing at a rapid rate 1 http://www.transparencymarketresearch.com/pressrelease/it-robotic-automation-market.htm 2 http://www.idc.com/getdoc.jsp?containerId=prUS41072216 3 https://www.slideshare.net/AIFrontiers/jeff-dean-trends-and-developments-in-deep-learning-research/8 Growing use of Deep Learning at Google3 What’s changed: Convergence of 20+ years of AI research, Cloud Computing, Big Data and increased computing power Copyright © 2017 Deloitte Development LLC. All rights reserved. #NICSAMWRM
• 5.
  www.nicsa.org Robotic and cognitiveautomation solutions can increase capacity and extend the capabilities of organizations Criteria  Rules-based, standard, repeatable processes  Structured / digital inputs and outputs  Human interaction with multiple systems  Limited decision-making or interpretation Criteria  Unstructured inputs and outputs, including documents, forms, handwriting, audio, etc.  Customized, context-sensitive information  Feedback through training or data to improve algorithms over time (machine learning) Sample Use Cases  Opening e-mails and attachments  Accessing web/enterprise applications  Moving files and folders  Extracting structured data from documents  Filling in and validating forms  Aggregating, validating/reconciling, transforming and calculating data Sample Use Cases  Generating insights based on customer activity  Developing fact-sheets and investment summaries  Improving the effectiveness of electronic communications monitoring  Guidance through complicated workflow with the use of intelligent agents  Machine-learning-based exception management and root-cause analysis Cognitive Intelligence Robotic Process Automation Software that can be configured to undertake rules-based tasks, replicating human action Algorithms which can interpret, learn and communicate, replicating human thought Robotic Process Automation (RPA) Cognitive Intelligence (CI) By leveraging these capabilities, firms have begun to realize significant improvements leading to enhanced quality, reduced cost and efficiency gains Copyright © 2017 Deloitte Development LLC. All rights reserved. #NICSAMWRM
• 6.
  www.nicsa.org Demo: Data Management& Reporting What we will see  Using RPA to automate the FR-Y9C report production process using Excel, internal document management systems and databases  Automating the data acquisition, reconciliation, aggregation, transformation and validation activities, and using tools to generate exception reports and audit logs  Eliminating up to 80% of manual processing, enabling analysts to focus solely on managing exceptions and performing quality control checks Securities Lending Security Master & Pricing Corporate Actions Regulatory Reporting Automating the report production process for financial and non-financial reporting Aggregating overnight loan reports from external counterparties and balancing vs. stock record Identifying discrepancies between data providers or internal sources and generating exception reports Reconciling across pre- and post-payment data Where this applies Copyright © 2017 Deloitte Development LLC. All rights reserved. #NICSAMWRM
• 7.
  www.nicsa.org Robotics, Machine Learning, CognitiveComputing Andy Curtis, Data Analytics Manager, Northern Trust #NICSAMWRM
• 8.
  www.nicsa.org WHAT DOES ARTIFICIALINTELLIGENCE DO? Value to Financial Services Industry Smart Automation (Learnt patterns) AI Automation (Cognitive ability) AICapabilityLevels Basic Automation (Rules based) Research PoV & Prototyping Production Rollout • Document analysis – auto trade capture • Transaction reconciliations • Intelligent wealth advisory assistant • Compliance surveillance & reporting • Portfolio analytics and management • Advanced client personalisation Industry research and adoption is gaining pace as AI moves to mainstream #NICSAMWRM
• 9.
  www.nicsa.org WHAT ARE THEOBJECTIVES OF AUTOMATION? #NICSAMWRM
• 10.
  www.nicsa.org WHAT ARE SMARTMACHINES? Smart Machine Automation AdvancedBasic • Technology is making great strides in Smart Machines and Artificial Intelligence • The Fourth Industrial Revolution is underway: • Self-driving vehicles • IBM Watson Jeopardy champion • Google DeepMind Go champion • Still a long way from true artificial intelligent machine (i.e. the Terminator) • Basic process automation is the simplest form of smart machine automation #NICSAMWRM
• 11.
  www.nicsa.org ROBOTIC PROCESS AUTOMATION RoboticProcess Automation (RPA) refers to area of configuring “software” or “robots” to capture and interpret existing applications in order to perform a repeatable set of activities in an automated fashion. Features • Virtual workers • Faster setup • Uses existing systems/applications Benefits • Quick wins and faster ROI • Reduced risk and error rates • Respond faster to business peaks and troughs • Enable users to do other cognitive tasks #NICSAMWRM
• 12.
  www.nicsa.org COGNITIVE + RPADOCUMENT INJECTION  Documents arrive to centralized ops center  First several documents are unrecognized and must be “trained” via ML (classify + snap name/value)  After training documents are auto classified, text and meta data extracted  Passed on to RPA for business process completion #NICSAMWRM
• 13.
  www.nicsa.org ROBOTIC PROCESS AUTOMATION WhyAutomation Projects Fail And How To Avoid This? Randy Guy, Chief Technology Officer, Asset Management
• 14.
  www.nicsa.org KEY THOUGHTS Think strategic andnot tactical RPA investment continues to grow “One size does not fit all” – Don’t lock into a single technology 14
• 15.
  www.nicsa.org 15 Top 5Reasons RPA Projects Fail Buying a tool in place of creating a strategy Not setting up the proper organizational structures Run as a project instead of a program Process being automated is not well understood Bringing in an RPA tool and ignoring what is in place
• 16.
  www.nicsa.org Organization Initial ExperienceNew Plan Global telecom organization Automate Account Set-up, Operations and Closing • RPA tool redundant with other in-house tools. • Took many months to automate 2 FTEs • Tool did not have the cognitive processing required • Reviewed the process – got savings through streaming the process • Implementing cognitive solutions, which will allow them to automate up to 80% of the process Global card provider Automate Dispute processing • IT purchased an RPA solutions – No prior business buy-in • Tried to automate everything with workflows given by the business, but not working as a team with the business – did not automate the proper flows. • Set up the proper organization and realized savings of 1.1M. Fitness product provider Automate fulfilment process • Did not have expertise in automation – partner used for fulfilment processing did not have the expertise in cognitive solutions • Got stuck with a process they felt could be automated, but did not know how • New partner: 30% discount on service • Took on the risk of automation savings • Partial automation solutions utilizing cognitive solutions for processing emails/ e- forms via NLP. • Single screen interface for unifying multiple applications and screens • Leveraging BOTs to do the actual processing. Examples of RPA Failures and Turnarounds 16 #NICSAMWRM
• 17.
  www.nicsa.org Expected cost basedon project complexity Complexity Cost BASIC e.g. Update form from 1 app to another HIGH Judgement-driven processes, e.g. disputes management, document management, KYC, etc. MODERATE Standardized, unstructured inputs: e.g. Name, address change form processing $40K – $100K $100K – $175K $300K – $500K Illustrative costs include software, resources and infrastructure costs. 17 #NICSAMWRM
• 18.
  www.nicsa.org Questions & Roundtable Breakout Sessions #NICSAMWRM