Recursive self-improving customer service agents adapt and evolve automatically
CARE: Grounded Recursive Self-Improvement for Production Customer-Service Agents
Artificial Intelligence
Summary
Customer service agents powered by AI need to keep getting better as policies and situations change. The authors created CARE, a system that helps these agents learn from their mistakes after they start working, using human checks to confirm errors. When an agent makes a wrong move, the system figures out if the problem was the AI model itself or other tools, fixes the issues, and retrains the model to avoid repeating the same mistake. CARE also tests new versions carefully before fully replacing the old ones to ensure reliable improvements over time.
What this means in practice
- •For customer service bot developers: Update chatbots progressively using verified failure reports to improve accuracy without stopping service.
- •For business process automation teams: Maintain automated workflows that adapt to changing policies by integrating mixed historical failures for continuous model evolution.
Authors
Yanbei Jiang, Chao Lei, Rena Gao, Xiaochen Zhao, Yuchen Ge, Haohang Liu, Runze Nie, Qitian Chen, Qi An, Haolin Song, Yufei Sun, Yihao Ding, Jey Han Lau, Krista A. Ehinger, Xingjun Ma, Weijia Chen, Ruiyuan Wu, Zhen Luo, Ke Xing, Yongjun Yin
Abstract
Recursive self-improvement (RSI) in production agents involves transforming accumulated experience into reliable model updates while adapting to evolving business policies. We present CARE (Customer-service Agent Retrospective and Evolve), a grounded RSI framework that operationalizes this process in customer service. Its goal is to replace a frontier proprietary incumbent with an in-house mid-sized model that continues learning after deployment. After a one-time cold start on governed production traces, the current model serves production traffic alongside the incumbent, and its own failures drive the next update. A 24/7 inspection process establishes failures through evidence-blinded, independently reviewed judgments. Attribution separates model errors from defects in Skills, tools, or the harness, routing component failures to repair. Model-attributable trajectories are truncated at the first erroneous decision and paired with verified repairs. On-policy self-distillation (OPSD) then updates the model using the same model as teacher, with the diagnosis and repair supplied as privileged information unavailable to the student. To support continued evolution as business requirements change, training mixes failures from different Skill/SOP versions, each judged against the policy in force when it occurred. Candidates run an executable simulation test compiled from the latest SOP and then frozen, followed by offline inspection of the resulting traces. Promotion requires a combined simulation-and-inspection score at least at parity with the deployed version, confirmation through an online comparison with the incumbent, and human approval.