Skip to main content

PDF Remediator

Campus AI workflow

Pilot

Finds the document-accessibility problems software can catch reliably, and hands a qualified remediator the evidence for the rest.

  • 17+Automated checks9 categories
  • HumanFinal reviewRequired for every document
  • ReviewableEvidence packageFindings and remaining work

Accountability

How this use case is governed

Service owner
TritonAI solutions team and campus accessibility partners
Human oversight
A qualified human validates reading order, semantics, alternatives, and the final accessible document.
Measurement plan
Issues detected, remediation time, residual accessibility findings, and reviewer agreement.
Data boundary
Public description; document handling depends on source data

Measurement period 2026 pilotLast reviewed 2026-07-24

Product media

See the workflow in action

Silent screen recording showing a PDF upload, automated accessibility findings, remediation progress, and the reviewable results.

  • veraPDF Validation
  • PDF/UA Standard
  • Evidence Packs

How it works

How the workflow fits

The problem it addresses, what it actually does, and how far along it is.

Problem

Some of PDF accessibility work is mechanical and checkable. The rest depends on visual and semantic judgments software cannot safely make on its own.

Solution

The workflow looks for common structural and metadata problems, proposes fixes for some of them, and assembles an evidence package a human remediator can work from. It runs 17+ automated checks across 9 categories, covering reading order, tagged content, alternative text, and document structure, so a reviewer can see what was checked and what is still open. It assists expert work. It does not certify conformance.

Current status

Pilot. A finished document still needs a manual accessibility review against the appropriate standards. A staff member with no engineering background built this through the Citizen Developer Program. It runs in the browser, uses no database, and is ready for campus SSO.