Status : Verified
| Personal Name | Narvaja, Armand Angelo S. |
|---|---|
| Resource Title | Knowledge-2-resolution: a support troubleshooting playbook utilizing AI initial triage |
| Date Issued | 5 January 2026 |
| Abstract | This study aims to assess and develop an existing troubleshooting playbook of the organization for support engineers that uses manual keyword search as its initial course of investigation, integrating the AI-assisted search as initial part of troubleshooting process with human-decision skills and expertise. Working with support engineers drawn from multiple product teams, the research began assessing the sentiments of the users using the existing process to map how they find and acquire knowledge on their day-to-day productivity. These engineers helped in co-designing the structure of the playbook, and initiated a controlled pilot deployment using tagging, surveys, interviews, and Net Promoter Score comparisons. The evaluation focused on practical outcomes (average resolution time, trends on how AI improved the percentage of resolved cases, and how support engineers adopted the use of the tool. The result shows positive user sentiments after the AI process implementation, support engineers reported faster starts on their investigation and few repetitive searches. While the NPS distribution moved towards greater share for the AI-assisted method, the actual reactions are still parted, some support engineers welcomed the AI as learning improvement and time-saver, while others preferred manual keyword searching for control and contextual accuracy. The qualitative data revealed common sentiments that Gen AI provided too vague and sometimes irrelevant suggestions that still need human validation and review. The study concluded that AI can materially improve initial triage process through the average resolution time cases were resolved and its consistent increase of volume on cases resolved each week when implemented as a troubleshooting partner rather than replacement to support engineers. The trust and confidence from users of utilizing the tool shows huge improvement as well overtime. Recommendations include continuous user-feedback loops, tagging of cases, and mo |
| Degree Course | Master of Technology Management |
| Language | English |
| Keyword | Playbook; Application support; Troubleshooting playbook |
| Material Type | Thesis/Dissertation |
Preliminary Pages
123.13 Kb
Category : C - Confidential information of a third-party is embedded.
Access Permission : Limited Access