D. Brian Letort, Ph.D.

Governed enterprise AI at global scale.

Operate. Publish. Teach.

Five concurrent enterprise programs
$30M budget
120+ matrixed contributors
$1M+ vendor spend eliminated
CEO-active platform
1,000+ member AI community
5 U.S. Patents
4-paper research program
16 Pluralsight courses
2 books

I build the governed AI platforms, operating models, and data foundations that turn experimental AI into durable enterprise capability. My work spans a $30M portfolio, 120+ contributors, and the rare intersection of operating at public-company scale while publishing a four-paper research program on the systems layer that makes enterprise AI reliable.

Token economics is the new unit economics.

When three thousand dollars of tokens replaces a hundred thousand dollars of vendor work, you are not saving money on AI. You are restructuring what enterprise work costs.

Governance is the velocity layer, not the brake.

Ungoverned AI dies in legal review. Governed AI ends up in the CEO's hands. The platforms I build are used at the top of the house because they are trusted there.

Operate, don't advise.

I do not consult on enterprise AI. I run it. Five concurrent programs. $30M budget. Every major function of a global public company on one governed platform.

Contact See outcomes

Selected Outcomes

What I’ve built.

See all six outcomes

Economic Inversion

I replaced over $1M in recurring vendor-assembled reporting with governed AI pipelines running on a four-figure token footprint.

A single high-complexity analytical deliverable that once cost approximately $100,000 per cycle and took the better part of a quarter to produce now runs on roughly $3,000 of tokens and returns a governed, citable answer in hours. At portfolio scale, the program eliminated more than $1M in recurring vendor spend — at a 1,000–3,000x cost inversion and 156x faster turnaround.

What this provesI do not deliver AI pilots. I deliver verifiable economic inversions at enterprise scale.

Enterprise Operating Surface

I architected and operate the governed AI platform that became the default operating surface for every major business function of a global public company.

Used across legal, finance, HR, operations, revenue, investments, and product — with an active user at the CEO level, compliance approval for high-sensitivity use cases, and a 187% adoption lift inside weeks of a governance-driven rollout.

What this provesI design AI platforms that earn enterprise trust — and produce the adoption numbers that follow from it.

Adoption Flywheel at Enterprise Scale

I built a 1,000+ member internal AI community with 170–178 live attendees per bi-weekly session — one of the most widely attended internal programs at a global public company.

Community is not a soft outcome. It is the lever that turns an AI tool license into a workforce habit. The operating model behind it has produced measurable adoption lifts and seven-of-seven executive business-unit alignments.

What this provesI build the cultural operating system that makes the technical operating system actually get used.

Research

Context Compilation Theory.

Read the research

A four-paper research program on the systems layer between retrieval and reasoning — the governed, measurable, optimizable layer where reliability lives, where provenance is enforced, and where token economics becomes a design surface rather than an accounting line.

Foundational paper