Marginalia Partners helps independent professional firms ship real AI capabilities — without the hype, the lock-in, or the fear that the technology is here to replace anyone.
The firms that run our economy are getting sold AI by people who've never worked inside one.
The name is deliberate. Marginalia are the notes written in the margin — the practitioner's annotations that make a text actually usable. That's the work: not grand transformation decks, but the specific, marginal, high-leverage changes that make AI useful inside a real practice on a real Tuesday.
Independent professional firms are skeptical of AI for good reason. They've been pitched by vendors who don't understand their work, sold tools that never get adopted, and told their people are about to be replaced. None of that has earned their trust — and none of it has shipped them anything that works.
We do the opposite. We sit inside the firm, find the one or two places AI genuinely creates leverage, ship them in weeks, and leave the firm able to do it again. The judgment your clients pay for stays with your people. The repeatable work around it gets faster.
Our method has a name because it's repeatable. Our approach takes a firm from "we should look at AI" to a capability running in production — three deliberate moves, each earning the next.
A free 30-minute readiness diagnostic and a 1-page assessment you keep — your opportunities, your hidden risk, your plan.
A fixed-scope, four-week Sprint that ships one real AI capability at a time into your real workflow. Ship-or-it's-on-us.
Implementation and optional continuity that leaves your firm able to keep shipping — on its own terms, without us.
William founded Marginalia Partners after years as a product manager in banking — a world where the gap between an impressive demo and a tool people actually use is measured in real money. He started Marginalia because he kept seeing the same gap open up for independent firms: enormous AI promise, almost no practical guidance from anyone who'd done the work inside a practice like theirs.
He works hands-on inside every engagement. The writing in Resources comes from real work in real firms — first-hand experience, not aggregated theory. That's the whole basis of the practice: guidance you can trust because it's written by the person doing the building.
Jonathan is a security and risk advisor with ten years helping companies build trust and pass enterprise security review. For firms whose entire business runs on client trust, he's the person who makes sure adopting AI strengthens that trust rather than putting it at risk.
He partners with founders and their teams to embed governance directly into the product and the way it's built — and shows firms how to adopt AI securely and at scale using the standards that actually matter: the NIST AI Risk Management Framework, ISO 42001, and OWASP's controls for large language models.
Thirty minutes, a 1-page assessment you keep, and an honest read on whether AI is worth it for your firm right now — including a clear "wait a quarter" if that's the truth.
