← DisclosureTrail For agencies & freelancers

Disclosure Record vs. Provenance Log: What's the Difference?

"We reviewed it and signed off" is not the same claim as "here's exactly what tool made it, and here's the source." Most people conflate these until they need one and only have the other.

The short version A disclosure record proves a human reviewed and approved a piece of work, and when. A provenance log proves what actually produced it — which tool, which version, and what changed along the way. They answer different questions, they get challenged in different situations, and having one does not mean you have the other.

Why this distinction matters

Say a client questions a deliverable six months after you shipped it. There are two very different things they might actually be asking:

These come up in different situations. A disclosure record matters most when trust or process is being questioned — "did anyone check this?" A provenance log matters most when the artifact itself is being questioned — a copyright dispute, a client wanting to edit an AI-generated asset later, or a technical review of exactly how something was built.

Where a disclosure record falls short

A typical disclosure statement asks the reviewer to type in which tool was used. That's a self-reported field, not a verified fact — nothing about a disclosure record independently confirms the tool claim, tracks which version of that tool was active, or keeps the original editable file. For text and written deliverables this rarely matters; the reviewer read the actual output and can vouch for it regardless of tooling specifics.

It matters more for generated visual or media assets shipped as the final deliverable. A generated image or video has failure modes a hand-built file doesn't: it may not be exactly reproducible, the source prompt/seed/version chain can get lost, and there's often no underlying editable file the way there would be with a design file built in layers. If a client later asks "can we get the source file" or "what actually made this," a disclosure record naming a reviewer doesn't answer that.

Where we stand today DisclosureTrail's Pro tier currently builds disclosure records — a persistent, timestamped history of who reviewed what, and when, searchable by client. It does not yet track tool/version metadata or preserve editable source files. That's a real gap for anyone shipping generated visual or media assets as the deliverable itself, and it's on our honest roadmap rather than something we're claiming to already solve.

A simple way to tell which one you need

SituationWhat actually answers it
AI-assisted writing, human edited/reviewed itDisclosure record
Client asks "did someone check this before sending?"Disclosure record
AI-generated image/video shipped as the final assetDisclosure record + provenance log
Client wants to edit or reproduce the asset laterProvenance log
Dispute over what tool/version actually made somethingProvenance log

What to do about it today

If you're doing AI-assisted writing, research, or editing work, a disclosure record covers most of what a client will actually ask for — see what a disclosure statement should contain. If you're shipping generated visual or media assets as the deliverable itself, keep the source file, prompt, and tool/version noted somewhere alongside your disclosure record until purpose-built provenance tooling catches up — a plain note in your project folder is better than nothing.

Start with the record you can build today

DisclosureTrail generates a dated disclosure statement — who reviewed it, when, what tool, what involvement level — in under a minute. Free, no login, nothing stored on our end.

Generate a disclosure statement →