IVR Journey Mapping via Embedding Vector Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current IVR systems face challenges in automatically identifying customer journey contact reasons due to the need for manual rule definition and analysis of extensive IVR sessions, which is time-consuming and limited to situations with a complete category list, failing to address partial or non-existent category lists.
Innovation Solution
A method that builds a directed graph from sample IVR journeys, filters non-informative menus, concatenates menu prompts and user responses, calculates similarity scores with category names, and maps journeys to categories, enabling unsupervised categorization without prior knowledge of the IVR system logic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual rule definition is used to identify contact reasons, then categorization accuracy can be maintained, but the process requires substantial manual labor and takes several weeks
Solution Approach 1:
The system performs self-service by automatically analyzing IVR session data, menu structures, and customer interactions to generate contact reason categories without requiring manual analyst intervention. The automated system processes and categorizes journeys independently, eliminating the need for human analysts to manually review sessions and define rules.
Solution Approach 2:
The patent replaces the mechanical manual process of rule definition with an automated computational system that uses algorithms to analyze IVR data patterns, menu navigation paths, and interaction sequences. This substitution transforms the manual analytical process into an automated information processing system.
2Measurement precision
If complete category list is available, then accurate mapping can be achieved, but the system cannot address situations with partial or non-existent category lists
Solution Approach 1:
The system achieves universality by designing a categorization framework that functions effectively whether a complete category list is provided, a partial list is provided, or no category list exists. The same automated analysis engine adapts its operation based on the availability of category definitions, making the system versatile across different deployment scenarios.
Solution Approach 2:
The system exhibits dynamic behavior by adjusting its categorization strategy based on the completeness of the provided category list. When categories are available, it maps to predefined categories; when categories are partial or missing, it dynamically generates or suggests categories based on observed interaction patterns.
3Loss of information
If hundreds of IVR sessions are reviewed manually, then common patterns can be identified, but the process becomes extremely time-consuming and labor-intensive
Solution Approach 1:
The patent replaces manual pattern recognition with automated computational analysis that processes IVR session data, menu navigation paths, and customer responses using algorithms. This substitution enables the system to analyze vast numbers of sessions rapidly, identifying patterns that would be impractical for human analysts to detect manually.
Solution Approach 2:
The system performs excessive action by analyzing far more IVR sessions than the minimum needed for manual pattern identification. While manual analysis might review hundreds of sessions, the automated system can efficiently process thousands or millions of sessions, ensuring comprehensive pattern detection and more robust categorization.
Data Source
AI summary
Systems and methods of mapping a customer journey in an interactive voice response (IVR) system to a contact reason from a contact reasons list: receive an IVR log comprising a plurality of customer journey entries, wherein each customer journey entry comprises a sequence of one or more menu identifiers; generate an embedding vector for each menu identifier; filter one or more menu identifiers from a menu identifier list, wherein the menu identifier list comprises all menu identifiers contained in the IVR log; cluster one or more remaining menu identifiers from the menu identifier list into one or more clusters, based on the embedding vector of each menu identifier; map each cluster to a contact reason; and create a rule that categorizes a newly received IVR sequence based on a cooccurrence of at least one menu identifier in the newly received IVR sequence and in a given cluster.