Why a Semantic Layer is Not Enough for AI
A semantic layer tells the AI what your metrics mean. It says nothing about the test accounts, the migration, or the Slack thread that explains the number.
Read article
More articles
Extending Data-Centric AI to Context
Two employees ask the same question and get two different correct answers. That's not a bug. It's a defining property of organizational context.
Read article
The Silent Killer of AI Accuracy: Stale Context
When a metric drops, a human investigates. An AI with stale context confidently tells you the business is down - and misses the actual story.
Read article
The Context Problem: Giving AI the Right Foundation for Data Analysis
The gap between demo and production usually isn't the LLM - it's context. Why starvation and overload both fail, and what dynamic context retrieval looks like.
Read article
Why the Company Brain You're Building for Your CEO Might Be Your Biggest Mistake
The instinct to build a company brain is right. The build is what will break you - a permanent operational dependency with no expiry date.
Read article
MCP Grew Up. The Interesting Problems Just Moved Up the Stack.
The Model Context Protocol went stateless. The protocol stopped pretending it could hold your state - that responsibility got promoted to the application layer.
Read article
How to Build a Context Layer and a Company Brain
What it actually takes to turn a company's scattered knowledge into something an LLM can reliably use - and why the demo is 5% of the work.
Read article