Request a Demo

Insights

Field notes from enterprise private AI.

What we are seeing in the market, what the model layer is actually doing, and what we have learned deploying private AI inside firms across PE, VC, law, audit, and corporate dev. Published when there is something worth saying.

A tug-of-war between public AI models and a private AI — the framing the article reframes.

Field Note

Thinking Tools and Infrastructure

Foundation models and a private RAG are not the same product, and the firms that treat them as competitors end up buying the wrong things in the wrong order. A field note on the reframe — and the hybrid stack that follows from it.

May 8, 20266 min read
Read the story

More from the desk

  • A working data room mid-deal — documents legible at the moment of upload.

    Market Trends

    The Quiet Reshaping of Due Diligence

    The most consequential change in deal work this decade is happening without a press release. Inside the new tempo of diligence, what is shifting, and why most firms underestimate the speed of the curve.

    April 28, 20266 min read
  • Fine-tuning illustration — the practitioner view of model alignment work.

    Practitioner Notes

    Fine-Tuning Without the Hype

    Fine-tuning is the most over-pitched and under-understood capability in financial AI. A practitioner-level look at when it actually moves the needle, when it does not, and what to ask before signing the work order.

    April 14, 20267 min read
  • Citation-grounded answer next to its source page — verification one click from any sentence.

    Field Note

    Citations as Currency

    In production AI for financial services, the line between a useful tool and a liability runs through one feature: whether every answer is grounded in a source the user can verify. A field note on why this stopped being optional.

    March 30, 20265 min read
  • A working production deployment of private AI inside a financial services firm.

    Implementation

    Beyond the Demo: What Production AI in Finance Actually Looks Like

    Demos are easy. Production is the part nobody puts on stage. Lessons from twelve months of deploying private AI inside investment firms — what holds up, what breaks, and what every team underestimates.

    March 12, 20268 min read
  • Private AI versus public model — the architectural fork that defines enterprise deployments.

    Architecture

    Why Enterprises Need a Private AI

    The convenience of typing a question into a public model hides a set of architectural trade-offs that get expensive at enterprise scale. A look at the four places a private AI quietly outperforms going direct: model choice, retrieval economics, document scale, and the permissioning that compliance teams actually require.

    September 18, 20257 min read

Want to be notified when we publish?

Get in touch