PAGER: Proactive Monitoring Agent for Enterprise AI Assistant
Junior Garcia (New York University), Sujan Dutta (Rochester Institute of Technology), Pranav Umakant Pujar (Adobe), Sai Sree Harsha (Adobe), Dan Luo (Adobe), Nikhil Vasudeva (Adobe), Bikas Saha (Adobe), Pritom Baruah (Adobe), Yunyao Li (Adobe)
Engineering & Operations
Summary
A proactive monitoring agent that statistically models historical system errors to surface potential failures before they impact enterprise AI assistant users.
Description
Enterprise AI assistants embedded in customer data platforms typically operate reactively, addressing failures only after they occur. We present PAGER, a proactive monitoring agent that augments an enterprise AI assistant by statistically modeling historical system errors and surfacing potential failures before they impact users. Grounded in a formative study with five domain experts, PAGER addresses the need for proactive error remediation with a multi-turn conversational interface, predictive models coupled with natural language explanations, and interactive visualizations to support rapid troubleshooting. A technical evaluation shows that random forest models achieve F1-scores of 67.8% and 57.5% across two critical error prediction stages, substantially outperforming baselines. In a within-subject user study, participants using PAGER completed error investigation tasks faster, reported significantly higher ease of use and confidence in their findings, and achieved greater task correctness. These findings demonstrate that proactive AI support improves user confidence and decision quality in enterprise environments.
