Client Support System Using NLP for Analogous Request Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Information technology support services face inefficiencies due to dispersed and unstructured information about software issues, leading to time-consuming and unreliable responses, as support associates struggle to find prior related issues, which hampers meeting 'always-on' client requirements.
Innovation Solution
A client system support system that analyzes support request messages using natural language processing to identify analogous requests, allowing for automated responses from a repository, and escalates to expert units when necessary, with responses stored for future reference, thereby streamlining the support process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If support associates manually search for prior related issues in dispersed information sources, then they can provide customized support responses, but the process becomes highly time consuming and unreliable
Solution Approach 1:
The system enables self-service by allowing support associates to quickly find relevant information through automated search functions that query multiple data sources (knowledge base, issue trackers, customer records) simultaneously, reducing manual searching time while maintaining response reliability
Solution Approach 2:
An information retrieval intermediary system is introduced between support associates and dispersed information sources. This intermediary aggregates and indexes information from multiple sources, providing a unified search interface that rapidly delivers relevant prior issues and solutions without requiring associates to manually navigate multiple systems
2Reliability
If support associates spend considerable time searching for prior related issues, then they can ensure accurate responses, but efficiency decreases and 'always-on' requirements cannot be met
Solution Approach 1:
The system performs preliminary actions by pre-indexing and organizing information from multiple sources before support requests arrive. Prior issues, solutions, and customer data are aggregated and made searchable in advance, enabling support associates to quickly retrieve accurate information without time-consuming manual searches during actual support interactions
Solution Approach 2:
A universal search platform is implemented that serves multiple functions: searching knowledge bases, querying issue trackers, accessing customer records, and finding related documentation all through a single interface. This multi-functional approach maintains response accuracy while significantly improving support unit efficiency and enabling 'always-on' service levels
3Adaptability or versatility
If support associates manually collect required information and determine issue nature, then they can provide tailored solutions, but the process becomes tedious and time-consuming
Solution Approach 1:
The system implements feedback mechanisms where support requests automatically trigger searches for relevant prior information, and results are fed back to support associates along with suggested actions. This automated feedback loop reduces manual information collection time while maintaining the ability to provide customized responses based on the specific situation
Data Source
AI summary
In one embodiment, a support request message is received. Further, a check is made to determine whether a support request repository includes an analogous support request message analogous to the support request message. When the analogous support request is included in the support request repository, a response associated with the analogous support request is sent. When the analogous support request is not included in the support request repository, a client system connection associated with the support request message is determined. Furthermore, when the client system connection is established, the support request message, and a plurality of related prior support request messages and associated responses are sent to an escalation support unit. Also, a response received from the escalation support unit, and the support request message, are stored in the support request repository.