Watson: A Cognitive Observability Framework for the Reasoning of LLM-Powered Agents
This addresses the problem of monitoring and debugging autonomous LLM agents for developers and users, representing an incremental advancement in observability tools for Agentware systems.
The paper tackles the challenge of opaque reasoning in LLM-powered autonomous agents by introducing Watson, a cognitive observability framework that recovers and inspects implicit reasoning traces without altering agent behavior. It demonstrates practical utility in debugging and correction scenarios on benchmarks like MMLU and SWE-bench-lite, showing actionable insights for improving transparency and reliability.
Large language models (LLMs) are increasingly integrated into autonomous systems, giving rise to a new class of software known as Agentware, where LLM-powered agents perform complex, open-ended tasks in domains such as software engineering, customer service, and data analysis. However, their high autonomy and opaque reasoning processes pose significant challenges for traditional software observability methods. To address this, we introduce the concept of cognitive observability - the ability to recover and inspect the implicit reasoning behind agent decisions. We present Watson, a general-purpose framework for observing the reasoning processes of fast-thinking LLM agents without altering their behavior. Watson retroactively infers reasoning traces using prompt attribution techniques. We evaluate Watson in both manual debugging and automated correction scenarios across the MMLU benchmark and the AutoCodeRover and OpenHands agents on the SWE-bench-lite dataset. In both static and dynamic settings, Watson surfaces actionable reasoning insights and supports targeted interventions, demonstrating its practical utility for improving transparency and reliability in Agentware systems.