LangChain AI Updates: June 3, 2026

1. LangChain Introduces Rubrics for Self-Evaluating Agents

LangChain. LangChain introduced Rubrics for DeepAgents, a feature that lets agents evaluate and self-correct their own output against defined criteria before returning it. It ships as an open-source addition to the DeepAgents architecture aimed at improving output quality on multi-step tasks. Source

2. Case Study: Rippling Goes AI-Native With Deep Agents and LangSmith

LangChain. LangChain published a case study on how Rippling adopted Deep Agents and LangSmith to make every product AI-native within six months. The write-up details how the company leaned on LangSmith for observability and evaluation as it rolled agents across its product suite. Source