A New Way Forward

Every company is a mind.

It learns, it decides, it develops judgment — hard-won, expensive judgment about its customers, its craft, and itself. And then it forgets. The person who knew leaves. The reason behind a decision dissolves into a recording nobody watches. The same lesson gets paid for twice, then a third time, by people who never knew it had already been learned.

AI was supposed to fix this. It made it faster. Every assistant starts every session knowing nothing about the company it works for. A thousand brilliant sessions a day — and by morning, none of them happened.

Corveil is built on a different premise: what an organization learns should be the one thing it never pays for twice. Every person and every agent working from one living memory. Judgment compounding like capital — and owned like capital. The models everyone runs on are rented; they improve for everyone on the same schedule. A company’s memory is the one advantage that can only come from being that company.

So we build the operating layer where knowledge compounds, work stays grounded, and the organization’s intelligence outlives any session, any tool, any single career. The companies that run this way will be smaller than their ambitions suggest — and greater than their headcount explains.

That is the company we wanted to work for. So we started it.

Origin

Two obsessions. One answer.

Corveil began as two separate obsessions that turned out to be halves of one answer.

Danny and Dustin spent two decades building for institutions that cannot afford to be wrong — defense platforms, security companies, systems trusted with other organizations’ most sensitive work. Fourteen years of building together, across four companies, taught them one discipline above all: trust is architecture. If an organization is going to hand its knowledge to a system, governance cannot be a dashboard added at the end — it has to be the ground the system stands on. So that is what they built: a substrate that could hold a company’s knowledge, govern every use of it, and prove it.

Andy spent the same decades on the opposite problem: how organizations think. Knowledge graphs and language systems in production since 2009 — before the industry had names for this work — then agentic platforms with humans in the loop. By 2026 he had built himself an agentic operating system and was running his entire working life on it: agents that remembered, corrected, and improved, carrying yesterday’s context into tomorrow’s work. It worked well enough to pose a question that wouldn’t leave: if one person’s work can compound like this, what happens when a company’s does?

They met at the world model. One side had built the substrate an enterprise could trust. The other had built the cognition that makes memory act. Neither designed the other’s half. The architecture wasn’t planned — it was discovered by convergence. That convergence is Corveil.

The first customer was this company. Corveil runs on Corveil — the founders work inside the platform every day, and the operating model we sell is the one we already live in. What a company becomes on it is not speculation to us. It is Tuesday.

Founders

We are the customer.

Three founders run this company on the system they sell. The leverage is the evidence.

Danny Gershman

Co-founder & CEO

Danny has spent his career where security failure is not an option: cleared defense environments, federal platforms, and security companies. He led cloud infrastructure and security engineering inside a U.S. Air Force software factory, architected zero-trust systems for classified networks, has served as CISO, advised federal leadership on zero trust, and founded a security-first federal delivery firm. He co-authored Before The Commit and co-hosts its podcast on securing AI-driven development. Corveil asks organizations to trust it with everything they know. Danny’s career is the reason they can.

Dustin Hilgaertner

Co-founder & CTO

Dustin has spent two decades making regulated institutions’ data mean one thing: universal data models for the world’s largest financial firms, compliance systems evaluating thousands of builds a day for a U.S. Air Force software factory, machine learning shipped inside accredited government boundaries, and a streaming platform he led as CTO through its acquisition. He co-originated ModSecOps, a published discipline for securing and governing AI-assisted development. The hardest problem in organizational memory is making heterogeneous knowledge trustworthy at scale. Dustin has been solving exactly that — under audit — his entire career.

Andy Toh

Co-founder & COO

Andy has built the connective tissue between knowledge, language, and platforms since before the industry had names for it: semantic search and knowledge graphs in production at one of the web’s largest publishers, the data-science practice he founded there, platform inventions that became industry standards, and C-suite roles running digital businesses through transformation. Most recently he built production multi-agent systems with humans in the loop — and the personal agentic operating system that became the cognition half of Corveil’s architecture. He operates the company the way the platform intends: through it.

Build the company yours can become.

Sign up as a design partner