Ai driven rpa work fusion kyle hoback cwin18_nyc
1 like557 views
This document discusses how born-analog companies can compete with digital natives through AI-driven automation. It notes that while companies are living longer due to advances in healthcare, company lifespans are decreasing. Most companies spend billions annually on manual data work but old models of addressing this are not working. The document introduces the WorkFusion intelligent automation platform that uses AI and machine learning rather than rules-based automation to digitize entire human roles and processes. It can address common scenarios companies face with existing robotic process automation deployments, business process management software, or machine learning proofs-of-concept. The platform provides a complete solution to automate tasks like mortgage processing that require extracting and mapping unstructured data from documents.
Ai driven rpa work fusion kyle hoback cwin18_nyc
- 1. AI-driven automation: How born-analog companies can compete with digital natives Theo Piletski, Capgemini Kyle Hoback, WorkFusion September 18, 2018
- 2. People are living longer, but companies are dying younger Human lifespanCompany lifespan
- 3. Average tenure of S&P companies is 18 years and falling 33 years lifespan of a S&P500 company in 1965 Projected to go down to ▼12 years
- 4. We know why companies die
- 5. And we know why they dominate
- 6. Company survival is hard, but simple “A company cannot endure without reinventing itself.”
- 7. Most companies are fighting to survive with brute force $130b Spent by on data entry just in US $42b Spent on manually intensive data work by financial services industry alone $250b Global business process outsourcing industry $3t Data processing and data collection work worldwide CONFIDENTIAL
- 8. High Upfront Cost Do IT projects Hire more people High Operating Cost ? But old models aren’t working
- 9. The space between enterprise systems and front-end services is where work happens Apps Productivity Legacy Business Cloud Services Front-office Back- office Supply Chain
- 10. Apps Productivity Legacy Business Cloud Services Front-office Back- office Supply Chain RPA The first generation of RPA promised to automate it
- 11. Apps Productivity Legacy Business Cloud Services Front-office Back- office Supply Chain RPA OCR But it wasn’t a complete solution
- 12. Apps Productivity Legacy Business Cloud Services Front-office Back- office Supply Chain RPA OCR BPM And it needed more…
- 13. Apps Productivity Legacy Business Cloud Services Front-office Back- office Supply Chain RPA OCR BPM BI and more...
- 14. Apps Productivity Legacy Business Cloud Services Front-office Back- office Supply Chain RPA OCR BPM BI ML and more…
- 15. Apps Productivity Legacy Business Cloud Services Front-office Back- office Supply Chain RPA OCR BPM BI ML ITSM and more…
- 16. Apps Productivity Legacy Business Cloud Services Front-office Back- office Supply Chain RPA OCR BPM BI ML ITSM APM No single system to automate COMPLETE human role and more integration...
- 17. Until it created an even bigger mess than before Salesforce SAP Oracle Concur Flexcube LoanIQ Mainframe Mantis Bloomberg TIBCO Mulesoft SOA Pega CRM ERB EBS ABBBY Kofax BPO #1 BPO #2 BizTalk Excel Microstrategy Cognos Custom BI TensorFlow Custom Rage Watson Python SAS R Custom #2 US Account Opening Treasury Services Anti-Money Laundering Mortgage Services Pega App #2 Pega App #3 SFDC BPO #1 BPO #2 India Costa Rica Ireland
- 18. The first wave of RPA is business-as-usual dressed as transformation Transformation program timeline Scale POC 1 month “A few bots” 6 months “We hit a wall” 12 months “Unstructured data” “Exceptions handling” “Bot governance” “Analytics”
- 19. RPA 1.0 is still old-fashioned rules-based software
- 20. Software 1.0: Lots of rules-based code and a little data to automate tasks code data simple task
- 21. Digital natives are using “Software 2.0” – code that learns and improves from data
- 22. Software 2.0 uses data to automate more, faster code codedata data simple task entire role
- 23. AI changes how enterprises digitize operations Train software, don’t code Use screen, don’t integrate Scale real-time, don’t deploy/hire Reduce touchpoints, don’t re-engineer Learn, don’t standardize Before After Engineers coding Business training Licensed software Subscription software Multi-year IT deployments 4 week Agile sprints Six-Sigma operational excellence Digital workforce orchestration Offshoring to BPO In-sourcing with bots
- 24. Apps Productivity Legacy Business Cloud Services Front-office Back- office Supply Chain WorkFusion SPA is built on Software 2.0 and unifies all automation capabilities into one platform Robotics Governance Bot-Human Workflow Intelligent Robots Intelligent Digitization Intelligent Analytics Security
- 25. Apps Services WorkFusion Intelligent Automation AI Platform Robotics Governance Bot-Human Workflow Intelligent Robots Intelligent Digitization Intelligence Analytics Security Differential Privacy AutoQC OCR AutoML Preventative Controls Ops Analytics
- 26. 2 flagship products to start and scale AI-driven automation
- 27. “As RPA matures, analytics will dictate which providers lead the pack. Vendors that provide text analytics and artificial intelligence (AI)...will position themselves to deliver successfully” WorkFusion leading AI-led wave of automation “As RPA matures, analytics will dictate which providers lead the pack. Vendors that provide artificial intelligence (AI)...will position themselves to deliver successfully” RPA 1.0 Incumbents AI RPA 2.0
- 28. Leader in Intelligent Automation CONFIDENTIAL 28 “5 out of 5 in Analytics and AI” Forrester 2018 RPA Magic Wave “Best in capabilities and vision” Everest 2018 FIT Matrix “Choice for organizations dealing with data” Gartner 2017 Market Guide
- 29. But What If…? What if I already have BPM? What if I already have RPA? What if I’ve already done a POC on ML?
- 30. What if I already have BPM? CONFIDENTIAL 30 2. SPA can be called by BPM2 3. SPA can be called by BPM mid-process3 4. SPA can function as BPM4 Only RPA software with complete automation capabilities 1. SPA can call BPM1
- 31. What if I already have RPA? Transformation program timeline Scale POC 1 month “A few bots” 6 months “We hit a wall” 12 months “Unstructured data” “Exceptions handling” “Bot governance” “Analytics” Options: 1. Ecosystem of Automation Software 2. Single Automation Platform
- 32. What if I’ve already done a POC on Machine Learning? Typical POC gets here 1. Initial Model 2. Execute -able “In Production” gets here 9. Next Model(s)… 4. Handle Exceptions 5. Monitor Quality 6. Audit Data 7. New Data 8. Re- training 3. Connected Real Production is here And, on to next projects Options: 1. Ecosystem: different software for each step 2. WorkFusion SPA: packaged solution for each step
- 33. Three Common Scenarios 1. Robots Exist But Cannot Scale – robots without much value 2. Disparate Businesses or Processes – a lot of different things 3. Existing BPO Customers – lean but still manual approaches
- 34. Adding Cognitive Bots into basic-RPA landscape for Existing British multinational BFS company Existing Bot was extracting information from an Excel file and entering it into SAP Non-excel information must be extracted and mapped manually by employees The information then converted into specific web formats for processing RPA RPA could handle structured inputs only and couldn’t handle exceptions Cognitive RPA WF Models trained to map the data correctly and verify the information Bots input and map verified data for processing The results Accuracy Rate for RPA Accuracy Rate for ML 90% vs 100% Automated 90% of all completely manual process Reduced transaction processing time from 18 minutes to 3 minutes WorkSpace Exceptions that cannot be mapped correctly are sent to employees
- 35. Automating a wide array of LOBs and Internal Processes for Insurance Customer Sales Claims* Core work focused on intake and adjudication of claims that require human decisioning. Core work focuses on sales operations such as set up of new business, adding additional policy holder to groups, and executing benefit changes. Operations Core work focuses on policy holder support such as: • Deleting / adding dependents • Verification of Death, and • Appointments • Fulfillment delivery Federated LOBs focusing on reactive and transactional work through manual labor force under increasing volume demand. Siloed LOB specific implementations of Automation resulting in limited reusability and limited opportunities of to implement operational changes to respond to new growth and parallel operational efficiency requests due to customer experience constraints. This is not an exhaustive list of what these businesses do. Internal Processes* Internal processes such as travel approvals, expense reporting, internal and external Invoicing, Accounts Payable/Receivable, Compliance Workforce Orchestration Workforce Orchestration and Productivity measurement. Departmental and Individual SLA and KPI tracking Other * Utilizing “Niche” LOB-specific Automation
- 36. Claims Processing for Existing BPO Customer Client was processing 2M+ claims and is estimating 1M+ claims per year to be added Client prefers an onsite (higher cost) delivery solution for the BPO scope Processing takes place across multiple teams and at least 8 systems Problems to AddressClaim Requests are submitted for processing Capgemini optimizes the existing claims process Review, adjudication and payment must be also entered into multiple systems manually Automate the extraction of data from claims payment form Automate the data entry multiple systems Lean but Still Manual approach that can be automated Capgemini optimizes the process using ESOAR methodology Data from multiple client systems is moved to Capgemini BPO Solution for Insurance There are still some COTS systems team has to navigate across uses that can not be moved WF Platform BPM to Drive further process Optimization Over 35% Efficiency gains by leveraging WF RPA and Cognitive automation (vs 19% for RPA Only platform) Leverage WorkFusion reporting and visibility capabilities for SLAs/Metrics Achieve a high degree of financial payment accuracy and reduce Claims over-payments
- 38. Apps Services WorkFusion Intelligent Automation AI Platform Robotics Governance Bot-Human Workflow Intelligent Robots Intelligent Digitization Intelligence Analytics Security Differential Privacy AutoQC OCR AutoML Preventative Controls Ops Analytics
- 39. Mortgage Processing requires collecting data from documents, but it’s not a template-driven OCR problem
- 40. Solving Mortgage Processing requires the combination of capabilities
- 41. Building the process is simplified and more cost effective within one environment
- 42. Interacting with the UI is an important aspect
- 43. But, the true impact comes from working with the mortgage documents themselves
- 44. And, having the auditability and governance of this process (and others)
- 45. With the ability to understand the health and impact of all processes being automated
- 46. Thank You