🤖 AI Summary
This study addresses the challenge of maintaining effective AI-powered customer service in dynamic enterprise environments, where static assistants quickly become outdated. To overcome this limitation, the authors propose a closed-loop, self-evolving customer service system that enables large language models to adapt autonomously in production without requiring retraining. The approach integrates retrieval-augmented generation (RAG), evolutionary automated prompt engineering, and a modular alignment evaluation framework, complemented by version control and operational guardrails to ensure safe and reliable deployment. Empirical validation through A/B testing demonstrates substantial improvements over baseline systems: a 9.0-percentage-point increase in QA self-service rate, a 4.8-point gain in cancellation self-service rate, and a 30.6-point improvement in routing accuracy.
📝 Abstract
Enterprise support agents operate in rapidly changing environments where policies, product capabilities, and knowledge bases evolve continuously, making static assistants brittle and costly to maintain. We present LinkedIn's self-evolving agentic support system, which integrates retrieval-augmented generation with evolutionary auto-prompting and a modular, production-aligned evaluation framework to enable safe, continuous improvement without retraining foundation models. The system treats prompts, retrieval, and evaluation as a closed-loop, versioned workflow with operational guardrails. Offline simulations and ablations show clear quality gains over vanilla RAG and baseline agents, including reduced hallucinations and improved response completeness. In a two-week user-randomized A/B test on LinkedIn's production support traffic, the integrated self-evolved workflow increased QA self-serve by 9.0 percentage points, cancellation self-serve by 4.8 points, and routing accuracy by 30.6 points. These results demonstrate a practical path to scalable, self-evolving AI agents in real-world enterprise settings.