Putting Nuclear History Back to Work

Nuclear organizations rely on an extraordinary amount of documentation. I saw this every day in my 20+ years at the U.S. Nuclear Regulatory Commission as a license reviewer, inspector, and senior manager.

Decades of design and modifications, correspondence, inspections, vendor specs, and internal memos form the backbone of how plants are built, operated, and defended. This body of work represents hard-won engineering judgment and regulatory history. The value shows up when teams can access it fast and use it with confidence.

In nuclear, the difference between having documents and having knowledge is like the difference between owning a box full of gears and chains and seeing a child’s joy on Christmas morning with an assembled bike. All the components exist in both cases, but only one version can be ridden.

When you wish you had document intelligence

Document intelligence shows its worth at moments that require speed, accuracy, and confidence. I’ve seen situations like:

• A URI from an inspection raises a question about whether an activity is permitted. (This is where “we’ve always done it this way” needs receipts.)
• An RAI is launched when you’re up against an exigent amendment deadline. (Precedent beats heroics when the clock is running.)
• A new manufacturing scope of work needs to be evaluated against prior bids. (Get the contract faster when history is your friend.)

In each case, teams are drawing on decisions, analyses, and commitments made over long time horizons. The work depends on understanding what was done before, why it was done, and how it was justified.

History as a working asset

History in nuclear should be treated as a resource and a jump start, not a museum.

Past choices are inputs to present-day decisions. I’ve anchored multiple projects where the real work couldn’t start till I found the right references on the one microfiche reader that still worked in the building. Usable records let teams move faster and with greater confidence. That’s the difference between momentum and molasses.

Document intelligence is about keeping that history operational.

How documents become usable

At a high level, document intelligence treats legacy files as working capital.

This is where nuclear experience pays off. It’s why my team insists on the fundamentals:

• Deconstructing documents into data—text, tables, figures, handwritten notes—to maximize their downstream usefulness.
• Creating metadata for rapid retrieval, without relying on what an assistant typed in (or omitted) 34 years ago.
• Defining the nuclear ontology from first principles—and in a way that meets the needs of the wide variety of nuclear customers.

The result is a body of material that behaves like an organized knowledge base rather than a static archive.

What teams using Gordian start to experience

Everyday work changes. Easy to say, but what does that look like in practice? It’s what I’m seeing right now with our customers:

• They locate relevant information quickly without relying on personal reference collections.
• They assemble coherent, defensible answers grounded in prior plant history.
• They reuse past work intentionally, rather than reconstructing it under time pressure.

Long story short… time spent finding or re-deriving history drops. Confidence in the company’s position improves.

Why this matters now

As fleets age, people leave, and new projects inherit extensive documentary legacies, the ability to work effectively with historical records becomes increasingly important. I’m not there to anchor those NRC projects any more, and it shouldn’t be about me anyway.

Document intelligence keeps prior engineering and regulatory work active in present-day decisions. It allows experience to compound rather than reset. When documents function this way, they support better outcomes across operations, licensing, and oversight.

History stops sitting on the shelf and moves back into service.

If you’re ready to make your past a partner, reach out to me—theresa@everstar.ai.