Measuring Agents in Production – ICML

A new study, Measuring Agents in Production (MAP), based on 20 case studies and 86 surveyed practitioners, reveals that most deployed LLM-based agents use simple, controllable methods: 68% execute at most 10 steps before human intervention, 70% rely on prompting off-the-shelf models, and 74% depend primarily on human evaluation. Reliability remains the top challenge, addressed through systems-level design.

Computer Science Computers and Society Submitted on 2 Dec 2025 v1 https://arxiv.org/abs/2512.04123v1 , last revised 4 Jun 2026 this version, v4 Title:Measuring Agents in Production View PDF /pdf/2512.04123 Abstract:LLM-based agents already operate in production across many industries, yet we lack an understanding of what technical methods make deployments successful. We present the first systematic study of Measuring Agents in Production, MAP, using first-hand data from agent developers. We conducted 20 case studies via in-depth interviews and surveyed 86 deployed systems practitioners across 26 domains. We investigate why organizations build agents, how they build them, how they evaluate them, and their top development challenges. Our study finds that production agents are built using simple, controllable approaches: 68% execute at most 10 steps before human intervention, 70% rely on prompting off-the-shelf models instead of weight tuning, and 74% depend primarily on human evaluation. Reliability consistent correct behavior over time remains the top development challenge, which practitioners currently address through systems-level design. MAP documents the current state of production agents, providing the research community with visibility into deployment realities and underexplored research avenues. Submission history From: Melissa Pan view email /show-email/a2a079d2/2512.04123 Tue, 2 Dec 2025 16:45:10 UTC 337 KB v1 /abs/2512.04123v1 Fri, 30 Jan 2026 22:21:00 UTC 345 KB v2 /abs/2512.04123v2 Tue, 3 Feb 2026 18:06:26 UTC 345 KB v3 /abs/2512.04123v3 v4 Thu, 4 Jun 2026 19:57:38 UTC 340 KB Current browse context: cs.CY References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .