Speech Analytics at Florida Blue
BIOGRAPHY Tayo Odusanya IT Developer Florida Blue ( 10.5 Years) Advanced Analytics Consultant/Nexidia Voice Analytics SME Education: B.Sc. Electrical Engineering University of North Florida, MBA Jacksonville University Contact Info: Phone: 904-905-7313 Email: tayo.odusanya@bcbsfl.com 2
BIOGRAPHY Darin Southard Sr. Healthcare Director Nexidia Contact Info: Phone: 404-386-2922 Email: dsouthard@nexidia.com 3
Voice Analytics Overview What is Voice Analytics? How does it work? What does SCALABILITY it do for us? How can we know what we don t know? What we think we know What we know or learn Methodology & Results OUTCOME BASED 4
Interaction Analytics FINDING INSTANCES OF EVENTS IS AN IMPORTANT FIRST STEP BUT TO REALLY KNOW YOUR BUSINESS, YOU MUST COUNT EVENTS ACROSS ALL INTERACTIONS THEN YOU RE ABLE TO CREATE METRICS, ORGANIZE AND TREND AND USE THAT INFO TO TRANSFORM YOUR BUSINESS 5
Why Nexidia? Nexidia processes audio more than 50x faster than dictionary-based competitors. Nexidia can process 100% of the largest client s interactions on a reasonable footprint. SCALABILITY FLEXIBILITY Accuracy highest levels of precision and recall Dictionary and Language Model Independent Open Data not held hostage Nexidia references include the largest and best examples of interaction analytics. Learn from others who are successful today. REFERENCEABLE OUTCOME BASED Hosting PCI compliant data center Managed Analytic Services experts working for you Training & Support to make your staff experts 6
Voice Analytics: The Missing Insight KEY BUSINESS CHALLENGES Sales Effectiveness Churn/Retention Cost Management Compliance Customer Satisfaction TRADITIONAL DATA SOURCES BEHAVIORAL CONTACT EVENTS CRM SYSTEM IVR BILLING SYSTEM Agent Behaviors Technical Problem Statements Caller Behaviors Troubleshooting Techniques OTHER DATA OTHER DATA CUSTOMER INTERACTIONS Caller Intent Caller Discontent Retention Offers Evidence of Issue Resolution Questions Regarding Fees Customer Confusion DATA WAREHOUSE OTHER DATA CALL RECORDER Competitor Mentions Customer Frustration 7
Accuracy of Call Categorization 8
Value of Speech Analytics What are our customers telling us about their experiences? What do my customers think about my products and services? Spotlight Member Experiences Improve Business Processes How are advocates interacting with my members? Improve Agent Performance Are we adhering to proper protocol on every call? Monitor Compliance 9
Workflow and functionalities DISCOVER ORGANIZE ANALYZE OPERATIONALIZE Automatically identify business topics Uncover relationships between topics Analyze sentiment, social media, and emerging trends Call categorization based on business logic Integration with operational data Creation of reporting and views to facilitate analysis Root-cause Predictive modeling Customer journey mapping Customer satisfaction and customer effort models Metrics Transformation programs Real-time monitoring Sales and marketing campaign management Metric-based performance management INGEST AUDIO + ADD INQUIRY DATA ORGANIZE CALLS WITH RELEVANT METADATA IDENTIFY EVENTS QUANTIFY EVENTS AND CREATE METRICS BUILD MODELS CREATE CORRELATIONS UNDERSTAND TRENDS DRIVE DECISIONS 10
Early Discovery Overview Automatic topic and trend discovery for proactive speech analytics Highly configurable word clouds include filters for sentiment, speaker, phrase length Sentiment analysis sorts calls based on positive and negative sentiment and trends with visual displays Related phrases functionality shows relationships between topics, providing context for root cause analysis 11
Word Cloud- Identify New Topics and Trends 12
Call Categorization: Sizing the Opportunity Nexidia Call Driver Analysis Jan 1, 2014 Present 2.9 Million Calls BENEFITS 63.3% ENROLLMENT 29.1% Healthcare Benchmark Data Metrics Florida (Min) Blues Avg (Min) Average Handle Time 11 7 Average Non Talk Time 6.5 4.3 Average Hold 1.54 1.5 Transfer Rate 15 12 CLAIMS 24.3% MULTIPLE CONTACTS 17.5% Business Case = $871,251 BILLING 17.5% MEMBER COSTS 15.5% NEG SENTIMENT 12% ELIGIBILITY 7.1% REWORKED CLAIMS 6.6% FIRST CALL RESOLUTION : Focused on repeat calls and correlated with agent behaviors that drive repeat contacts Benefit Coverage Prescription Drug Coverage XEROX CALL ANALYSIS: Focused on reduced customer effort and cost to serve Transfers Reworked Claims Letter Related Calls Gave Assignment Promised Action Failure 13
Appendix Applying Speech Analytics at Florida Blue 14
Call Transfers 15
Recommendation: Call Transfers 16
Member Deductible- Self Service Cost to Serve Customer Experience Simple deductible questions are being answered by agents Providing these answers through web / IVR channels will lower cost and improve the member experience SIZING THE FLORIDA BLUE OPPORTUNITY HOW DOES IT WORK? Calls where the customer asks whether or not they ve met their deductible cost Florida Blue approximately $491K / year. Page 17 Page 17
Member Deductible Push to Self Service Redirecting 25% of the deductible related calls into a self service channel poses a cost savings opportunity of $123K dollars. Key Observations: Members are calling to ask why they are getting bills They do not understand the concept out of pocket maximums and co-insurance Agents spend ~3-5 minutes educating the member Members state that they were online in only 3.5% of these calls Members were expecting to see claims on their online statements Frequency of these calls has increased by 77% since the start of the year Deductible Related Calls By LOB Subgroup Key Actions for Florida Blue: When educating the members, have the agents show the members where they can find this information on the Florida Blue website 18
Network Contracting Provider Contract Negotiation Call Impact 19
Anthem Security Breach Monitoring 20
Case Study: Default Call Operators Used Nexidia voice analytics solution to identify top call reasons and gaps in technology. Identified benefit related topics that can be moved to self service Identified top callers through the analytics that can be routed directly to the right destination. 21
Compliance- Outbound Script 22
Competitive Mentions 23
Using Negative Sentiment To Identify Issues 24
Identifying Agents with high Non-Talk Time 25