SecMate: Multi-Agent Adaptive Cybersecurity Troubleshooting with Tri-Context Personalization

2026-04-29

AI Summary

This study addresses the low resolution rates and poor user experience in traditional cybersecurity troubleshooting, which stems from the lack of personalized adaptation to the triadic context of devices, users, and services. To overcome this limitation, the authors propose the first multi-agent virtual support system that integrates these three contextual dimensions. The system leverages lightweight local diagnostics, large language models, implicit user profiling, and context-aware recommendation mechanisms to enable adaptive fault resolution. Experimental results demonstrate that the approach elevates the correct resolution rate from approximately 50% to over 90%, achieves an MRR@1 of 0.75, and is perceived by users as a viable substitute for conventional IT support at a price significantly lower than human labor costs, thereby confirming both its effectiveness and practicality.

Abstract

Recent advances in large language models and agentic frameworks have enabled virtual customer assistants (VCAs) for complex support. We present SecMate, a multi-agent VCA for cybersecurity troubleshooting that integrates device, user, and service specificity from conversational and device-level signals. Device specificity is provided by a lightweight local diagnostic utility, while user specificity relies on implicit proficiency inference and profile-aware troubleshooting. Service specificity is achieved through a proactive, context-aware recommender. We evaluate SecMate in a controlled study with 144 participants and 711 conversations. Device-level evidence increased correct resolutions from about 50% to over 90% relative to an LLM-only baseline, while step-by-step guidance improved pleasantness and reduced user burden. The recommender achieved high relevance (MRR@1=0.75), and participants showed strong willingness to substitute human IT support at costs well below human benchmarks. We release the full code base and a richly annotated dataset to support reproducible research on adaptive VCAs.

Problem

• cybersecurity troubleshooting
• multi-agent system
• context personalization
• virtual customer assistant
• device-user-service specificity

Innovation

• multi-agent
• tri-context personalization
• cybersecurity troubleshooting
• implicit proficiency inference
• context-aware recommender