Customer Journey Friction Detection Using LLM-Based Event Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to identify and aggregate electronic customer journey information effectively, leading to fragmented and outdated data, which hinders the identification of friction events, resulting in inefficiencies and missed opportunities to improve customer satisfaction and business operations.
Innovation Solution
A system utilizing a fine-tuned large language model (LLM) to analyze electronic customer journey datasets, identify friction events, and generate user interfaces or alerts based on friction event information, including topics and customer sentiments, to facilitate proactive mitigation of such events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional systems are used to collect customer journey information from various channels, then information can be gathered, but the information becomes fragmented and outdated
Solution Approach 1:
The patent merges multiple data sources (websites, mobile applications, email, telephone calls) into a unified customer journey dataset. The system combines fragmented information from various channels into a single integrated structure, enabling comprehensive analysis while reducing the complexity of managing multiple separate systems.
Solution Approach 2:
The system creates a universal framework that handles multiple types of customer interactions through a single platform. The customer journey analytics system serves multiple functions: data collection, information aggregation, friction event identification, and actionable insight generation, eliminating the need for separate specialized systems for each function.
2Reliability
If friction identification rules are applied to electronic customer journey datasets, then friction events can be identified, but the system requires complex analysis capabilities
Solution Approach 1:
The system employs automated friction identification rules that self-applied to customer journey datasets without requiring manual analysis. The rules automatically detect friction events, classify them into categories, and generate insights, enabling the system to service itself rather than requiring complex manual intervention for each analysis task.
Solution Approach 2:
The patent transforms complex qualitative analysis into quantifiable parameters and rules. By defining friction events through specific measurable criteria (response times, interaction patterns, completion rates), the system converts subjective analysis into objective, automated detection that maintains high accuracy while reducing operational complexity.
3Loss of information
If comprehensive analysis of electronic customer journey information is performed, then friction events can be understood, but the process is slow and reactive
Solution Approach 1:
The system performs preliminary aggregation and organization of customer journey data in real-time as interactions occur. By continuously maintaining updated datasets and pre-processing information, the system prepares analysis-ready data structures that enable rapid friction event detection without requiring time-consuming comprehensive analyses when events occur.
Solution Approach 2:
The system maintains continuous monitoring and analysis of customer journeys rather than performing periodic batch processing. The friction identification rules operate continuously on streaming data, enabling real-time detection and response. This continuous action eliminates the reactive delays associated with periodic analysis while maintaining complete insight into customer experiences.
4Productivity
If businesses react to friction events after they occur, then responses can be made, but opportunities to prevent future events are lost
Solution Approach 1:
The system implements closed-loop feedback by continuously analyzing customer journey data, identifying friction patterns, and generating actionable insights that feed back into business processes. This feedback mechanism enables businesses to not only respond to current friction events but also adjust processes proactively to prevent future occurrences, transforming the traditional reactive model into a proactive prevention system.
Data Source
AI summary
System and methods to identify one or more friction events during an electronic journey of a customer of a business are disclosed. The system obtains a plurality of electronic customer journey datasets. At least one electronic customer journey dataset corresponds to the electronic journey of the customer during a transaction between the customer and the business, and indicates associated transactional events. The system uses friction identification rules to identify the friction events indicated in the electronic customer journey datasets that negatively impact completion of the transaction. Responsive to identifying the friction event, the system generates friction event information, and provides the friction event information to a fine-tuned large language model (LLM). The fine-tuned LLM determines topics and customer sentiments associated with the friction event. The system generates a user interface including the friction event information, and/or an electronic alert based upon the friction event information.