AI rewrote your document.
Can you prove what changed and who approved it?

An evidence SDK that connects document-level AI edits with human review, approvals, final versions, storage, and sharing events.

document_review_session
Export
Service Agreement
v7 draft
4.2 Limitation of Liability
AI-written change
AI CHANGE
Clause 4.2 rewritten to clarify liability scope.
AI

A file update is not an evidence trail

Storage logs can show that a document was modified. They cannot prove which AI-written change was reviewed, who approved it, what entered the final version, or where that version was shared.

Typical storage log

				
					file: contract_v7.docx
event: modified
user: @ReviewBot
time: 2026-06-30 14:32
				
			

What the log cannot answer

What exactly did AI change?

Who reviewed or approved it?

Did it enter the final version?

Where was the final version shared?

AI edits collapse into "file modified"

AI-generated text and revisions can become part of the document while the evidence is reduced to a file-level event.

Review sits outside the content trail

Review and approval records may exist, but often stay disconnected from the AI-assisted changes they cover.

Final files lose the decision path

The final document shows the result, but not which AI-assisted changes were accepted, rejected, or carried forward.

Traceability is becoming a product requirement

Storage logs can show that a document was modified. They cannot prove which AI-written change was reviewed, who approved it, what entered the final version, or where that version was shared.

EU AI Act, Article 12

logging and traceability

NIST AI RMF

trustworthy AI risk management

An evidence SDK
that connects the editor and the workspace

Capture AI edits, human review, and approval actions inside the document experience, then link them to final versions, storage, sharing, and audit events outside the file.

Capture editor-level actions

Record AI edits, human revisions, review decisions, and approval actions from the document experience.

Link workspace events

Connect each evidence record to final versions, storage, sharing, and workflow activity across your stack.

Export an audit-ready trail

Show what changed, who reviewed or approved it, and what reached the final document.

AI Document Evidence SDK animated evidence layer diagram AI and human document events flow into an external evidence layer, connected to workflow and storage evidence. Editor-level actions Workspace-level events AI AI changes generated or revised text USER Human edits written or modified by users OK Review & approval reviewed or approved VER Version history final document versions DMS Storage events saved, moved, retained OUT Sharing events shared externally LOG Audit metadata actor, time, source Evidence SDK outside the document file

Built to fit your document stack

Designed for document editors, AI review tools, Microsoft 365, DMS, CLM, and approval workflows.

Core capabilities for AI document evidence

AI Change Evidence

Capture AI-written, rewritten, inserted, or removed content from the document editing experience

Human Review & Approval Proof

Link each AI-assisted change to the people who reviewed, edited, approved, or rejected it

Version & Workspace Link

Connect evidence records to document versions, storage events, sharing activity, and workflow systems

File-Independent Evidence

Keep structured evidence outside the document file, without relying on Track Changes or file metadata alone

Tamper-Evident Records

Create signed, append-only evidence records designed to detect after-the-fact edits

Audit Export & Integrations

Export evidence trails or connect them to audit, compliance, and workflow systems

Built for teams adding AI to document workflows

If AI is now writing, reviewing, or revising business documents in your product or organization,
this evidence layer is designed for you.

AI document & agent products

Products that add AI writing, review, or redlining to business documents

Legal tech & contract review

Contract review, redlining, and legal workflow products adopting AI assistance

Document workflow platforms

Document management, collaboration, storage, CLM, and approval platforms

Compliance & audit programs

Governance, risk, audit, and records teams responsible for document oversight

Building AI document workflows?

Share how your team reviews, approves, and audits AI-assisted document changes. We’re using early feedback to shape the Evidence SDK.