Chat Frustration Detection via Contextual Text Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Real-time online chat sessions lack effective indicators of mood, leading to misunderstandings and potential frustration due to the absence of facial expressions, voice intonations, and other non-verbal cues, which can delay discussions and create communication barriers.
Innovation Solution
A system and method that identifies text-based signals of potential frustration through a combination of global, local, and user-specific dictionaries, and performs text analytics to determine the cause of frustration, allowing for responsive actions to alleviate or address the issue, such as suggesting rephrasing, translation, or changing communication methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If text analytics analysis is performed to detect frustration in real-time chat sessions, then communication effectiveness is improved, but system complexity and processing time increase
Solution Approach 1:
The text analytics system is divided into multiple specialized dictionaries (global, local, user-specific) that segment the frustration detection task into manageable parts. Each dictionary handles specific aspects of frustration identification, making the overall system more tractable and maintainable while improving detection accuracy.
Solution Approach 2:
The system performs preliminary action by pre-compiling multiple dictionaries of frustration indicators before real-time chat analysis. These dictionaries contain pre-identified text-based signals, patterns, and context hints that are prepared in advance, enabling faster real-time detection without performing full analytics from scratch during chat sessions.
2Measurement precision
If multiple dictionaries and text analytics analysis are used to detect frustration, then detection accuracy is improved, but processing time increases
Solution Approach 1:
Multiple dictionaries containing frustration indicators, patterns, and context hints are compiled in advance before real-time chat analysis. This preliminary preparation enables the system to perform rapid matching during chat sessions without performing full analytics from scratch, maintaining high detection accuracy while reducing processing time.
Solution Approach 2:
The system creates simplified copies of frustration detection rules in the form of structured dictionaries that can be quickly queried. Instead of performing complex analytics on every message, the system uses these pre-processed dictionary copies to rapidly identify frustration signals, balancing accuracy with speed.
Data Source
AI summary
Frustration in online chat communication is prevented in which a text transcript is generated by at least two chat participants, by: (i) identifying in a text transcript a text-based signal listed on a list of text-based signals in a table; (ii) performing a first text analytics analysis on the text transcript to determine whether potential frustration is evidenced by the text transcript, the first text analytics analysis using a context hint provided in the table, the context hint corresponding in the table to the text-based signal; and (iii) responsive to potential frustration being evidenced by the text transcript, taking a responsive action based at least in part upon a potential cause of the potential frustration determined by performing a second text analytics analysis on the text transcript. The context hint is a contextual clue that supports a determination of whether the text-based signal indicates potential frustration as evidenced by the text transcript.