Blog
Notes on data unification
Occasional writing about knowledge graphs, entity resolution, and why enterprise AI programs stall in the same place.
Perspective
Your AI agents are going to be wrong until you build a knowledge graph
Everyone can describe the destination, and everyone can list vendors. The stretch of road in between is the part almost nobody in leadership can name, and it is where I keep watching smart companies stall.
Read
Perspective
Why embedding-based RAG can’t answer the questions enterprises actually ask
Retrieval finds text that resembles your question. Most business questions require computing over records. That gap is why so many AI pilots die between the demo and the audit.
Read
Product
A successful scan is a real test
The most common way integration projects lose their first fortnight is discovering, late, that the service account cannot read the table everyone assumed it could.
ReadQuestions about any of this?
We are happy to have a technical conversation without a slide deck involved.