LangChain founder: what enterprise AI lacks isn't better models, it's a knowledge factory
hwchase17 · x · 2026-11-22
LangChain founder Harrison Chase amplified his article "How to build a knowledge factory," arguing that what makes AI work for a company is not better models or more context, but better knowledge of how the company actually operates.
A knowledge factory is the system that collects, keeps updated, and supplies this knowledge to agents. It holds three types:
- Skills: how your team delivers, e.g. writing a report in your house style;
- Domain knowledge: how the business works — products, customers, competitors, what KPIs and metrics mean;
- Data access guidance: which sources to trust for current data and how to pull from them.
This knowledge usually lives only in your best people's heads or is scattered across documents and chats. Since manual upkeep doesn't scale, the factory automates it in two parts: agents get tools to flag anything missing or wrong, forming a continuous learning loop.
Chase notes that LangChain's internal agents (GTM & content agents, LangSmith Engine) already run learning loops with these pieces.
More from coding & agent
- Ransomware gangs wired Claude Opus and Kimi into attacks: 32 MCP servers, $4.5K AI-driven ops — cyb3rops · 2026-11-22
- ArcKit v6.25.0 brings 190 architecture governance Agent Skills to Google Antigravity — mcraddock · 2026-11-22
- A 6-step recipe to build your own knowledge factory with Obsidian plus a cloud agent — hwchase17 · 2026-11-22
- Claude played three roles in the Riemann Hypothesis proof: ideation, verification, explanation — Prestigious_Bag_5349 · 2026-11-22
- Developer asks how to track full agent journeys across docs, CLI and MCP — nguyenfamjj · 2026-11-22
- t3code runs 8 sub-agents on the same brief in parallel with one sentence — jasonkneen · 2026-11-22