Artificial Intelligence in Action Embedding RPA and AI in Operations

Transcription

1 Artificial Intelligence in Action Embedding RPA and AI in Operations Harsha Kumar President Prodapt Pari Bajpay Senior Vice President of Digital Platform Development CenturyLink 1

2 Prodapt - World s leading robotics implementation partner for CSPs/DSPs 2

3 The New CenturyLink The new CenturyLink, purpose-built to serve the connection needs of the networked world. CenturyLink understands the power of all that is digital. Life is powered by connections and business is connections. Our extensive network reach, significant local presence, and innovative technology solutions combined with an experienced team of employees and relentless focus on the customer experience create a powerful choice for enterprise, government, and wholesale customers globally, and consumers nationally. The new CenturyLink is equipped to serve customers with reliable, flexible and secure connections where, when and how they need them. Approx. 450,000 Route Miles of Fiber Globally 52,500 Employees Globally Approx. 360,000 International Transport Miles International Transport Route Miles are a combination of leased and owned fiber and optical transport connectivity. 100,000 + On-Net buildings 60 + More Than Proforma Annual Revenue $24B Countries and counting (Estimated trailing twelve months, ending December 31, 2017)* *Excluding revenue related to the colocation business sale and including estimated intercompany eliminations and purchase accounting adjustments. Approximately 75% of core revenue will come from business customers and nearly two-thirds of core revenue will come from strategic services CenturyLink. All Rights Reserved. 3

4 What is Robotic Process Automation? Automation technology that significantly reduces manual effort associated with swivel-chairing across systems are software Automated data entry Business Benefits OPEX/Cost Optimization Improved Accuracy (Right First Time) Improved Productivity (24*7 Operations) Robots enable the automation of repetitive rule-based processes mimic interactions of users with multiple systems Multi-system integration Repetitive tasks Process reconciliation Data validation/quality work across functions Processing business rules Service Fulfillment Service Assurance Network Management Billing 4

5 Three main levels of maturity Basic Robotic Process Automation (RPA) Cognitive RPA Intelligent RPA Structured Data and Rules Semi-structured Data + NLP AI/ML Decision Making Automation of repetitive activities without changing the map of existing systems (OCR+ML, NLP, Virtual Assistants, Chatbots) + RPA Machine learning & deep learning-based self-learning bots Routine, repetitive Follow instructions Broad, any suitable process Maturing Lower Weeks Tasks Actions Applications Market Cost Timeframe Complex, cognitive Draw conclusions Narrow, insightful output Emerging Higher Months LOW Maturity Model in Automation HIGH 5

6 Basic RPA (Structured Data and Rules) Based on Prodapt s customer engagements, typical savings in the service fulfillment area (such as order handling) have been in the ~60% range Basic RPA Cognitive RPA Intelligent RPA Various Systems/Apps Customer Address Validation Orders are automatically collated from different source systems Robots automatically validate the addresses for serviceability Robots validate the product codes and associated charges Nothing missed! Orders are keyed in automatically by the robots into OMS with no errors 100% orders processed through RPA - Faster time to market and first time rights No backlogs whatsoever - No more wait time Zero errors induced by robots - Provisioned as ordered! Customer Satisfaction 6

7 Cognitive RPA is applicable across the customer engagement lifecycle Basic RPA Cognitive RPA Intelligent RPA Ex-1: Upgrades/downgrades of services Ex-2: Customer complaints registration and fault diagnosis Ex-3: Provisioning status verification and port status/availability Ex-4: Employee reimbursements/payroll/letter requests Web Static/FAQ Web Dynamic/Interactive FB Messenger WhatsApp Voice. Invoice Support Payment Support Instant Electronic ID&V Marketing Presales Support Select Service Onboard Service Guidance Join Pre-eligibility Check Provisioning/Porting/ Fulfillment Use Invoice Support Service Change Management Personalized Customer Support Pay Engage User Community Interaction Upsell and Cross Sell Personalized Offers, Support and News 7

8 NLP powered BOT identifies a customer request, understands the emotions, offers right solutions & also triggers backend processes through RPA for quick implementation This is ridiculous I had asked the customer care rep to put a notification if I am about to exceed 20GB; why isn t this happening? I apologize Mr. Kumar Let me look into it Thanks for your patience I noticed that you had this conversation on 4/26 but we did not update your account in time for this change to be effective I cannot pay for anything more than 20 GB data!! Period!! I am sorry for this inconvenience I can create an adjustment to your account to cancel additional charges. Shall I go ahead and do that? Basic RPA Cognitive RPA Intelligent RPA Bot automatically senses the tone of the customer and associates it with displeasure uses an apologetic tone to pacify the customer Bot has the ability to understand the context that the customer is referring to a past conversation and had asked for a specific update to be done Bot again accesses the required systems at the backend to provide accurate information to the user with minimal wait time Bot proposes a corrective measure based on the intelligence as to which step will solve the issue faced by the customer - also understands the customer s queries about payment Bot has the ability to implement corrective actions as well by triggering RPA bots in the backend to make the necessary changes The bot uses episodic memory to understand past actions and also the EQ ontology to understand human emotions while responding 8

9 Intelligent Robotic Process Automation Enable BOTs to make decisions similar to those made by human experts Basic RPA Cognitive RPA Intelligent RPA 9

10 CenturyLink Digital Transformation Vision 01 Customer Experience Shift to left DevOps 06 Shift towards digital media with realtime transactions and consolidate portals to an integrated omnichannel experience. Product and services available through online store of the future. 02 Customer Care Cognitive enabled agents. Exceptionbased handling with smaller, higher skilled centers. 03 Digital Business Model Market place of digital products with real-time or near real-time fulfillment & simplified experience for traditional products. Customer Care of the future 02 Customer Experience 03 Digital Business Model Shift to left - DevOps 06 Big Data & Analytics Driven 05 Personalized Experience Reduce risk, cycle time and improve speed to market with DevOps-based software development. Personalized Experience 05 Personalized experience and targeted bundles and solutions through social/mobile channels by leveraging customer segmentation analysis. Big Data and Analytics-driven 04 Leverage big data analytics, deep learning, and AI/ML to drive experience, improve efficiency, and increase revenue. Digital Platform Self-service Omnichannel API Strategy Social Integration Data Analytics Automation Cognitive/AI 10

11 RPA Approach: Centralized & Federated BOT Development Factory Model Stakeholders from Telecom operator Benefits Centralized or De-centralized CoEs Business 1 Business 2 Joint RPA Team Business 1 Business 2 Business 3 Joint RPA Team Demand Management + Generation RPA Benefits Realization IT-Head Responsible for run & maintenance of IT/Infrastructure Delivery/Program Head This core team will build additional team (developers, testers, controllers, etc.) based on RPA project s roadmap and requirements Continuous Improvement Head of RPA CoE Chair of the governance board from customer end RPA Core Team Process Consultants Delivery Steering + Training/Skill Building Business Unit Head Responsible for program management Technical Architects Begin with centralized model and move into federated model for scale Standardize governance and ROI measurements Identify and share best practices to optimize investment Leverage RPA as legacy systems wrappers for systems modernization effort RPA provides quick operational benefits allowing for long term strategic platform work 11

12 Bringing it all together to provide better service at lower costs Technician Customer Employee Omni Channel IVR Chat SMS Self Help IoT (Alexa) Social Media NLP Actions for issue resolution Chatbots Cognitive/ML Auto Resolution API Layer (back-office automation) RPA (back-office automation) Legacy Applications Microservices Benefits Reduce order and care call volume Provide seamless multichannel experience Eliminate tech calls to back office and improve MTTR Storage of customer interactions information for future analytics Natural language processing (NLP) engine provides context, intent and sentiment identification to carry forward meaningful conversation with customer, technician or employee 12

13 How you can get started? Start with a process study to determine automation potential and execute a PoT (Proof of Technology) High Sample Chart Automation potential review factors Process standardization Technical feasibility Number of exceptions Step 1: Business Case/ROI Report Transaction time Number of distinct screens Security - thin client, special access Throughput/ volume Input and output format Environment - reusability, test data etc. Automation Potential Low Assignment Service Provisioning Long-term Action Sales Commission Require Change Billing Trouble Ticket Closure Toll-free Assignment Effort Immediate Directory Listing New Employee Onboarding High Process Size 13

14 Harsha Kumar President Thank You Pari Bajpay, Senior Vice President of Digital Platform Development 14