# On Evaluating and Improving Conversational Agents in Production

arXiv: https://arxiv.org/abs/2609.32092 · HTML: https://mlsanity.com/paper/2609.32092/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Social simulation and multi-agent debate, one of 16 directions in LLM agents and tool use.
- Builds on 45 of its 59 references that are in the map.
- First posted on arXiv 25 Sep 2026.

## Placed in

- LLM agents and tool use / [Social simulation and multi-agent debate](https://mlsanity.com/agents/social-simulation-and-multi-agent-debate/)

## Builds on (most cited first, 30 of 45)

- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Self-Refine: Iterative Refinement with Self-Feedback](https://mlsanity.com/paper/2303.17651/) (2023)
- [τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains](https://mlsanity.com/paper/2406.12045/) (2024)
- [Large Language Models are not Fair Evaluators](https://mlsanity.com/paper/2305.17926/) (2023)
- [The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery](https://mlsanity.com/paper/2408.06292/) (2024)
- [Large Language Models Are Human-Level Prompt Engineers](https://mlsanity.com/paper/2211.01910/) (2022)
- [Beyond Accuracy: Behavioral Testing of NLP models with CheckList](https://mlsanity.com/paper/2005.04118/) (2020)
- [Large Language Models as Optimizers](https://mlsanity.com/paper/2309.03409/) (2023)
- [DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines](https://mlsanity.com/paper/2310.03714/) (2023)
- [LLM Evaluators Recognize and Favor Their Own Generations](https://mlsanity.com/paper/2404.13076/) (2024)
- [Large Language Models Can Self-Improve](https://mlsanity.com/paper/2210.11610/) (2022)
- [Why Do Multi-Agent LLM Systems Fail?](https://mlsanity.com/paper/2503.13657/) (2025)
- [The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search](https://mlsanity.com/paper/2504.08066/) (2025)
- [State of What Art? A Call for Multi-Prompt LLM Evaluation](https://mlsanity.com/paper/2401.00595/) (2023)
- [ToolSandbox: A Stateful, Conversational, Interactive Evaluation Benchmark for LLM Tool Use Capabilities](https://mlsanity.com/paper/2408.04682/) (2024)
- [LLMs instead of Human Judges? A Large Scale Empirical Study across 20 NLP Evaluation Tasks](https://mlsanity.com/paper/2406.18403/) (2024)
- [Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systems](https://mlsanity.com/paper/2505.00212/) (2025)
- [Estimating means of bounded random variables by betting](https://mlsanity.com/paper/2010.09686/) (2020)
- [Game-theoretic statistics and safe anytime-valid inference](https://mlsanity.com/paper/2210.01948/) (2022)
- [Who Validates the Validators? Aligning LLM-Assisted Evaluation of LLM Outputs with Human Preferences](https://mlsanity.com/paper/2404.12272/) (2024)
- [Large language models can accurately predict searcher preferences](https://mlsanity.com/paper/2309.10621/) (2023)
- [Self-Harness: Harnesses That Improve Themselves](https://mlsanity.com/paper/2606.09498/) (2026)
- [Safe Testing](https://mlsanity.com/paper/1906.07801/) (2019)
- [Perspectives on Large Language Models for Relevance Judgment](https://mlsanity.com/paper/2304.09161/) (2023)
- [EvalLM: Interactive Evaluation of Large Language Model Prompts on User-Defined Criteria](https://mlsanity.com/paper/2309.13633/) (2023)
- [Style Over Substance: Evaluation Biases for Large Language Models](https://mlsanity.com/paper/2307.03025/) (2023)
- [TREC CAsT 2019: The Conversational Assistance Track Overview](https://mlsanity.com/paper/2003.13624/) (2020)
- [SimulatorArena: Are User Simulators Reliable Proxies for Multi-Turn Evaluation of AI Assistants?](https://mlsanity.com/paper/2510.05444/) (2025)

## Abstract (arXiv)

We present a framework for evaluating and improving a large-scale, multi-agent shopping assistant in production, and report lessons from its use. Offline evaluation of such a system faces three obstacles. (i) A logged conversation cannot be replayed against a modified system, because a different response changes every turn that follows. (ii) The unchanged system itself varies from run to run. Its LLM components are stochastic, and in product search the available products, their prices, and the customer's personalization signals change. (iii) Aggregate quality scores combine distinct behaviors, so they show that quality has changed but not which behavior caused the change. Our framework addresses each obstacle in turn. For a reported behavior, an Evaluation Harness generates targeted assertions and a fixed cohort of customer scenarios. It then reproduces the behavior in a local instance of the assistant through grounded user simulation. Instead of replaying the log, the simulator writes new customer turns conditioned on the recorded messages and context. Repeated runs of the unchanged system form a stored baseline. An Improvement Orchestrator turns the assertion results into hypotheses, implements each as an isolated modification, and compares it with the baseline using paired percentile bootstrap intervals over scenario-level differences. When an investigation ends, the harness may propose revisions to future evaluations, subject to human approval and without altering past decisions. We report production investigations with this framework. Assertion profiles showed which positions of a product carousel a failure affected, and repeated runs distinguished a real improvement from run-to-run fluctuation. Audits of the evaluation itself found a judge that lacked the evidence it needed and a model setting that was configured but not applied.