AgentGuard: Early-Warning and Routing for Predictable Agentic
AI on Azure

This project introduces AgentGuard, a monitoring and routing system designed to improve reliability and cost-efficiency in Azure-based agent workflows. By analysing early trajectory signals—such as reasoning patterns and tool usage—within the first 10–30% of an agent’s execution, AgentGuard predicts task outcomes and dynamically halts or re-routes agents to prevent failures. Calibrated with ADeLe for demand-aware evaluation, the system aims to optimise the Reliability–Latency–Cost frontier across diverse benchmarks, including software engineering and cybersecurity tasks. Expected outcomes include integrated Azure routing capabilities, diagnostic insights into team composition (optimized on agent psychometrics, collective agent behavior/ intelligence, neural evolutions, and evolutionary dynamics), and practical strategies for building more predictable, trustworthy agentic AI.

This research is conducted via The Agentic AI Research and Innovation (AARI) Initiative which focuses on the next frontier of agentic systems through Grand Challenges with the academic community and Microsoft Research.