What LangChain changed

• Added a new case study link for ServiceNow.
• Removed a case study link for Klarna.
• Added a new video link in the "Learn from teams" section.
• Reorganized the order of existing customer videos and case studies.

LangChain is swapping out customer proof points. They removed a Klarna case study and added ServiceNow. This shows they are focusing on showing how very large, traditional enterprise companies use their tools.

• Added a new "## Deploy" header to the product lifecycle section.
• This header now correctly labels the existing features for Agent Server endpoints and secure sandboxes.

The company is now explicitly highlighting its infrastructure capabilities for running AI agents, moving beyond just building and testing tools. This signals a push to own the full production lifecycle.

• Replaced the feature-based list (Observability, Evaluation, Deployment) with a structured 6-step lifecycle: Build, Test, Deploy, Monitor, Iterate, and Govern.
• Introduced "Govern" as a new core pillar, focusing on cost limits, sensitive data detection, and human-in-the-loop approvals.
• Added "SmithDB" for high-speed trace querying and "LLM Gateway" for spend control.
• Updated open-source download stats from 200M+ to 350M+ monthly downloads.
• Reframed "Fleet" as a natural language building tool for non-technical teams within the "Build" phase.

LangChain is shifting its positioning from a collection of tools to a structured, end-to-end "Agent Engineering Platform." They are emphasizing governance, testing, and iteration to move AI agents from experimental toys to reliable enterprise software.