$\tau$-Knowledge: Evaluating Conversational Agents over Unstructured Knowledge

TL;DR

This paper introduces $ au$-Knowledge, a comprehensive benchmark for evaluating conversational agents' ability to retrieve and apply unstructured domain knowledge in realistic, long-horizon interactions, highlighting current limitations.

Contribution

It presents a new evaluation framework and domain, $ au$-Banking, for assessing agents' performance in complex knowledge retrieval and policy execution tasks.

Findings

Agents achieve only ~25.5% success rate in the benchmark.

Retrieval accuracy degrades sharply over repeated trials.

Agents struggle with complex, densely linked knowledge bases.

Abstract

Conversational agents are increasingly deployed in knowledge-intensive settings, where correct behavior depends on retrieving and applying domain-specific knowledge from large, proprietary, and unstructured corpora during live interactions with users. Yet most existing benchmarks evaluate retrieval or tool use independently of each other, creating a gap in realistic, fully agentic evaluation over unstructured data in long-horizon interactions. We introduce $\tau$-Knowledge, an extension of $\tau$-Bench for evaluating agents in environments where success depends on coordinating external, natural-language knowledge with tool outputs to produce verifiable, policy-compliant state changes. Our new domain, $\tau$-Banking, models realistic fintech customer support workflows in which agents must navigate roughly 700 interconnected knowledge documents while executing tool-mediated account updates. Across embedding-based retrieval and terminal-based search, even frontier models with high reasoning budgets achieve only $\sim$25.5% pass^1, with reliability degrading sharply over repeated trials. Agents struggle to retrieve the correct documents from densely interlinked knowledge bases and to reason accurately over complex internal policies. Overall, $\tau$-Knowledge provides a realistic testbed for developing agents that integrate unstructured knowledge in human-facing deployments.