Using EXCITEMENT textual inference platform to extend Almawaves - PowerPoint PPT Presentation

P r o v i d i n g S o l u t i o n s Using EXCITEMENT textual inference platform to extend Almawave’s products capabilities Adriana Farina, Researcher and Analyst Programmer Iride Lab, Almawave, Rome Symposium on Semantic Text Processing Bar-Ilan University Nov. 18-19 2014

About us Iride Customer Centric Suite Agenda Iride Claim Textual Entailment Technologies in Iride Claim Question and Answer

Easier Front-End Activities Speech Analytics Information Search Optimization Active Complains Management Contact Classification Management Marketing Campaign Automation Social Media Analytics

Understanding of Customer Dissatisfaction reasons and prompt resolution to improve Customer Satisfaction and reduce churn rate. It identifies claim related contents, by understanding, analyzing and classifying meanings according to business vistas and needs .

(1/3) SEMANTIC ANALYSIS OF CLAIMS Different vistas for multiple business need purposes.

(1/3) SEMANTIC ANALYSIS OF CLAIMS Different vistas for multiple business need purposes. IDENTIFICATION OF CUSTOMER DISSATISFACTION Causes and critical emerging trends are detected, ensuring the achievement of an exhaustive and multi-channel vision of Customer complaint and communication.

(2/3) PRIORITY LEVEL ANALYSIS AND CLAIM AGGREGATION Iride Claim can prioritize claims according to implicit complaint criticality.

(2/3) PRIORITY LEVEL ANALYSIS AND CLAIM AGGREGATION Iride Claim can prioritize claims according to implicit complaint criticality. CHURN REDUCTION Risk evaluation of analyzed claims and consequent routing to Back-Office systems to properly address complaints, issues and monitoring outcomes.

(3/3) IDENTIFICATION OF KEY CRITICAL ISSUES Efficiency improvement by assigning issues to a specific step of the product/service life-cycle.

Textual entailment technology integration Iride Claim is the Iride Suite product that most of all can benefit from the results of a Recognition of Textual Entailment (RTE) engine and in particular from the availability of the entailment graphs produced on the basis of the underlying data.

Exploiting textual entailment Iride Claim will be more effective in helping business analysts by showing a complete resume of complaints by means of the entailment graph, where we have complaints “clustered” by area and with several levels of granularity regarding the details about them.

RTE engine integration RTE engine Entailment Graphs Data (users ’ claims)

RTE engine role Showing a complete summary of complaints

RTE engine role Showing a complete summary of complaints Making it easier to find similar/related complaints

RTE engine role Showing a complete summary of complaints Making it easier to find similar/related complaints Improving the search experience by producing more expressive annotations for documents

RTE engine role Showing a complete summary of complaints Making it easier to find similar/related complaints Improving the search experience by producing more expressive annotations for documents Developing new strategies for query expansion

RTE engine role Showing a complete summary of complaints Making it easier to find similar/related complaints Improving the search experience by producing more expressive annotations for documents Developing new strategies for query expansion Developing new strategies for results navigation

Thank you