140 likes | 260 Views
Amdocs Intelligent Decision Automation. AIDA and Guided Interaction Advisor to be announced Feb 14, 2011. Amdocs – Bill Guinn, Mike Lurye, Susan MacCall December 7, 2010. Overview for Gartner . The Leader in Customer Experience Systems Innovation. $3.0B in revenue $410M operating income
E N D
Amdocs Intelligent Decision Automation AIDA and Guided Interaction Advisor to be announced Feb 14, 2011 Amdocs – Bill Guinn, Mike Lurye, Susan MacCall December 7, 2010 Overview for Gartner
The Leader in Customer Experience Systems Innovation • $3.0B in revenue • $410M operating income • $1.4B in cash • More than 19,000 employees • Over 60 countries A Unique Business Model Service Provider Industry Focus Leading BSS, OSS & Service Delivery Products Strategic Services TM Forum’s “Industry Leadership” award. (June 2010) Highest Possible Overall Rating in BSS Scorecard and OSS Scorecard Reports. (Dec 2009, Jan 2010) Leading Telecom Operations Management Systems Vendor Worldwide. (May 2010)
What Is Amdocs Intelligent Decision Automation (AIDA)? • Amdocs Intelligent Decision Automation (AIDA) is a closed-loop, self-learning system that lets you… • See what happens, when it happens • Understand what it means to your business • Take action andenforce business policy – automatically, intelligently and in business real-time See • By uniquely combining technologies AIDA is able to better understand and predict the behavior of customers, providing individualized treatment to achieve specific business targets: • Increasing RPU by selling them the right products • Decreasing churn by treating customers based on their specific likes, needs, and recent issues • Reducing cost of operations by proactively preventing issues and optimizing cross-channel care functions Think Act
Requirements • Massive scale, billions of events/day -> 100M customers • Business real time infrencing and decisioning (100ms response time) • Uncoupled integration – open world model – • M:M:M : atomic ->abstraction : structured and unstructured data • Cohesive past, present, and real time view World View • True user managed and defined schema and decision factors (concepts) • User defined policy, decisions, and the logic to manage concepts • Closed loop learning: decision->actions->effect->learning • Low cost .05 customer/year
AIDA Runtime Architecture Think Act See Actions Decision Engine Events Application Server - Space Based Container Amdocs Event Collector Container Semantic Inference Engine andBusiness Rules Amdocs Integration Framework Event Ingestion Events Scheduled Events Bayesian Belief Network Revenue Mgt Ordering CRM CRM Other System Other System Other Systems “Sesame” Performance cache Network Business Intelligence Other System Other Systems RDF Triple Store DB Data Sources Any internal or external system Structured or Unstructured Operational Systems Any operational system
How the AIDA Inference Engine Works Semantic Concept Model The model defines the concepts including the high level business concepts Inference Engine andBusiness Rules Machine learning trained models The model contains the relationship between concepts including the dependencies BBN probabilistic model Recommendation model ( Future) Other Machine learning algorithms Inference • When an event occurs the event handler rule fires for that event • Evaluates the event message • Evaluates the existing ontology • Determines which semantic instances to create or update Bayesian Belief Network When any data changes, the inference engine fires in a “When - Then” style of computing, updating all “Automatic” concepts. Custom concept rules are fired if necessary. This creates a chain of updates When a “on demand” concept is needed the inference engine finds and computes all of the dependant concepts Business policy Rules Semantic Concept Model Event Create semantic concepts from events Machine learning Concept Define a custom business concept Action A set of custom rules actioning business policy When a concept is dependent on “machine learned” information the inference engine manages the invocation and timing of interfacing Defines concepts and meaning Defines relationships between concepts Semantic model capabilities reasoning
Composer Demo Application Filter Map CONSTANTS + BUSINESS ENTITIES & CONCEPTS AIDA Business Composer Library TAG PATH RELATIONSHIP MAP DEPENDENCY MAP Search Map Customer ACTIONS PROCEDURES EVENTS ENTITIES Entity Name Customer Acquisition Date Status Status History Type Call Frequency Churn Propensity Collection Risk Payment Pattern Customer Accounts Customer Action Acquisition Date Customer Bill Customer Accounts Status Customer Devices Customer Bill Status History Customer Problems + + + Type Customer Problems Customer Description. The role played by an Individual or Organization in a business relationship with the service provider in which they intend to buy, buy, or receive products or services from the service p - Customer Call Frequency - Interactions Churn Propensity Collection Risk Subscriptions Payment Pattern
Composer Demo Application Filter Map CONSTANTS + BUSINESS ENTITIES & CONCEPTS AIDA Business Composer Library TAG PATH RELATIONSHIP MAP DEPENDENCY MAP Search Map Customer ACTIONS PROCEDURES EVENTS ENTITIES Entity Name Customer Acquisition Date Status Status History Customer Accounts Type Customer Bill Call Frequency + Customer Problems Churn Propensity - Customer - Collection Risk Interactions Customer Payment Pattern Payment Pattern Subscriptions Customer Accounts VIEW EDIT DEPLOY Customer Action Acquisition Date Customer Bill Customer Payment Pattern Status Customer Devices LAST MODIFIED: 02/12/2010 MODIFIED BY: Greg Sloan VERSION: 1.3 COMPUTED:Automatically Status History Customer Problems DESCRIPTION: Determine the payment pattern for the customer by evaluating payments determining good payer, bad payer, worsening payer or improving payer + + Type Customer Description. The role played by an Individual or Organization in a business relationship with the service provider in which they intend to buy, buy, or receive products or services from the service p Call Frequency Find all of the Previous Bills Within 6 months 1. Churn Propensity Collection Risk Payment Pattern 2. If all bills have Payment Timeliness equal to early then Payment Pattern is good payer 3. If all bills have Payment Timeliness equal to late otherwise then Payment Pattern is bad payer 4. If at least 75% of Earlier Bills are early or on time otherwise and later bills are late then Payment Pattern is worsening payer 5. If At least 50% of Earlier Bills Late are and all later bills are On time or early then Payment Pattern is Improving payer
The Applications Virtual Agent (GA in September) Guided Interaction Advisor (GA in May)
A pre-built Ontology and rule set in AIDA designed to address key issues in call center interaction Amdocs Guided Interaction Advisor Anticipates reason for the customer interaction Then Automates access to the required information and Guidesthe flow of action and decision making Business Benefits Eliminates system and agent diagnosis time Provides consistent and efficient call handling Increases agent and customer satisfaction Anticipated benefits based on 100K actual accounts assessment: AHT reduction of 10-15% FCR improvement of 10-15% CSR training time reduction of 15-20% Guided Interaction Advisor (GIA)
and and and and and and and and How Guided Interaction Advisor Reduces AHTExample * In addition to displaying the bill, GIA returns multiple “next” actions such as dispute charge, adjust the fee, make payment depending on high level concepts and policy GIA Looks at events including bill sent event Finds abnormal fee – Returned Check Interactions Probability that the call is motivated by fee vs. other issues Compute High-level concepts LOGIC ZOOM: LOGIC ZOOM: VIEW VIEW EDIT EDIT DEPLOY DEPLOY simple simple advanced advanced Examine company policy for treating this customer Late Abnormal Fee Policy VERSION: 1.3 VERSION: 1.3 OWNER: User Name OWNER: User Name Dollars MODIFIED BY: User Name MODIFIED BY: User Name DESCRIPTION: Loremipsum dolor sit amet, consecteturadipiscingelit. Sedcondimentum ante euturpisplaceratporttitor. Maecenas vitae lectus non justomollisrutrum at condimentum mi. [Edit] DESCRIPTION: Loremipsum dolor sit amet, consecteturadipiscingelit. Sedcondimentum ante euturpisplaceratporttitor. Maecenas vitae lectus non justomollisrutrum at condimentum mi. [Edit] 1. 1. Determine actions for each probable issues If If Customer has a late Fee Customer has a late Fee and and Customer Payment Pattern Customer Payment Pattern is is Good Payer Good Payer or or Improving Payer Improving Payer and and Customer Value Customer Value is not is not Low Low 3 3 then then Credit Fee Credit Fee . . Presents contextually relevant info, streamlined action Other Amount Received Amount Received Partial Payment Partial Payment Late Fee Late Fee Payment Payment Credit Fee Credit Fee Method Method Customer Payment Pattern Customer Payment Pattern Payments Policy action rules CIM actions Customer Bill Motivation for call • Abnormal Fee 47% • Pay Bill 23% • Device Exchange 12% • Customer Cancel 5% • Bill Due Date is Oct 1 • Third Party Charges are normal • Charges are not typical • Bill has charges for fees. • Bill has abnormal fee • fee is “Returned check” • Tax amount is normal * Abnormal fee Display bill Highlight fee on bill Script message Prepare one click action Highlight one click action Pay Bill Educate free on-line pay Prepare one click pay Assess Probability Customer Graph High level Concepts • High value customer • Long time customer • Improving payer • Contract expiring Guided Interaction Advisor
Guided Interaction Advisor Predictions & Actions Events collected in real time Transformed into a connected graph of business concepts Plan Overage Bill greater than last month Prorates Interactions Roaming charges Orders Third Party Charges Bills Abnormal fee Payments Rate increases Charge dispute Collections First bill Charge dispute Past due amount Customer Pay bill Probability of call motivation Customer Cancellation Pay instructions Reactivation Individual Device activation Device Activated Device education Device heartbeat Device exchange Device Lost Subscriptions Device not working Chronology of events Device changes Subjective "good payer" Patterns "always pays 2 days late" Trends “improving payer" Geospatial “within 5 miles of the tower" Time “within 5 minutes of an outage" Probability “probably will call about the bill" Absence of occurrence “missed payment” Highly User extensible … Device resume Extensible Service data not working Service text not working Service voice not working Use Case Extensions
Presenting Insight to CSR Process opens relevant screen for reference and action Prediction on reason for the call – ranked by probability Recommended actions – based on best ROI Prioritized Recommended treatment and script