CampusMind

FAQ

Quick answers to the questions teams ask most about CampusMind.

Short answers, with links to the longer explanations.

About the product

What does CampusMind actually give me back?

The corrected PDF. Not a report of problems — the fixed file, plus a before-and-after report showing what changed. See Get started.

What file formats does it accept?

PDF only. Export from Word or PowerPoint first; doing so usually carries some structure over, which gives remediation a better starting point than a scan.

Which systems can it pull documents from?

Canvas, Blackboard, Brightspace, and SharePoint. Anything you upload yourself is recorded as source Manual. See Connect a source.

Does it change the original file?

Only if you tell it to. For course documents, the Replace original file toggle is off by default, so the corrected file is uploaded alongside the original. Turning it on overwrites the original in your LMS. The decision is made before the run, not after.

Standards and scoring

What standard is my document checked against?

PDF/UA-1 (ISO 14289-1), using VeraPDF. There are 106 rules and the score is the share that pass, so "106 of 106 rules passed" is 100%.

Does CampusMind check WCAG?

Yes, for the document-level criteria a checker can verify — 1.1.1 Non-text Content, 1.3.1 Info and Relationships, 1.3.2 Meaningful Sequence, 2.4.2 Page Titled, 2.4.6 Headings and Labels and 3.1.1 Language of Page, plus contrast reporting. PDF/UA-1 is the standard that encodes those requirements for PDFs, so validating against it is how they get checked. See Accessibility standards for the criterion-by-criterion table.

How does that relate to Section 508 and ADA Title II?

Both are US legal obligations that reference WCAG — Section 508 incorporates WCAG 2.0 AA, and the ADA Title II rule names WCAG 2.1 AA. The same structural work serves both. See ADA Title II.

What does a 100% score tell me?

That the document passes all 106 machine-checkable rules — a strong, reportable result, and the evidence half of your compliance story. The other half is judgment: a figure can carry alt text that passes the rule and still not convey what the figure is for. Skim the documents that carry an argument, and your score means what it looks like it means.

Why does my document score 99.99% but still show an issue?

The score is a proportion of 106 rules across the whole document, so one unresolved rule on a long file barely moves it. Remaining issues is the useful list — it names what is left, so you can clear it in the editor rather than hunt for it.

How much of the work is automated?

Roughly 80% of common accessibility issues, which is the structural work that would otherwise take a specialist hours per document. The Tag & Caption Editor is where you close the rest, and Revalidate re-scores the file when you do.

Running remediation

Which strategy should I choose?

Preserve Layout when the visual design must not change — print-ready and well-formatted documents. Full Accessibility Reconstruction for scanned documents, textbooks, lecture notes, and complex academic content. Reconstruction is marked Recommended and fixes more, but it's allowed to change the page. See Upload & remediate.

Can I re-run a document with a different strategy?

Yes. Run it again and choose the other strategy; each run creates its own batch, so you can compare the results and keep whichever you prefer.

How long does it take?

Minutes per document rather than seconds, and longer for scanned or complex files — the analytics dashboard reports your own averages per document and per page. Turn on Email notification for large batches instead of watching the screen.

Can I work through a course a bit at a time?

Yes — Select Unrun Files on the course screen picks only documents that haven't been remediated yet, which is what you want when working down a backlog.

What are the confidence thresholds for?

How certain the extraction engine must be before an element counts. Raise them to reduce false elements; lower them to keep more uncertain content.

Which sliders you see depends on the engine: Document Intelligence (the default) has Paragraphs at 0.95 and Figures at 0.99, PaddleX has Table and Formula at 0.50, and Hybrid shows both sets. Change them for a problematic class of document, not as a general setting.

Why do the settings change when I switch strategy?

Because the two strategies don't do the same work.

Full Accessibility Reconstruction shows an Advanced Settings panel — content type, page numbers, document layout, headers and footers — on by default. Preserve Layout shows Advanced Output Formats instead, an experimental toggle for podcast (MP3) and HTML output. The extraction engine and thresholds appear on both. See Upload & remediate.

What should I set "Optimize for Content Type" to?

Whatever the document actually is — Humanities (the default), STEM, Business, Medical, Chemistry, or Other. It matters most for the reconstruction features that need subject knowledge: captions for math, chemistry, and medical content, and table structure correction.

Review and editing

Do I have to review every document?

That's a policy decision for your institution — remediated documents are usable as soon as they finish, so nothing is held up waiting for you. A sensible middle path is to review anything the report marks Review required, plus a sample of the rest.

Is there an approval or sign-off step?

CampusMind uses a review-and-edit model rather than a gated one: you can improve any document at any time, and it stays available throughout. Accountability travels with the Run by record on each batch and the per-document version history, which is what institutions running their own review stage reference against it.

Can I fix something the automation got wrong?

Yes — every tag, heading, and alt text string is editable in the Tag & Caption Editor. Edits apply locally until you save, and saving creates a new version. Revalidate re-runs the PDF/UA check afterwards.

What should I change most often?

Alt text on charts and diagrams. Generated descriptions are literal and describe what's visible, which is the right starting point; you add what the image is for.

Access and administration

Why can't my colleague see my batch?

Because of groups, not roles. A batch is shared with groups when it's uploaded, via Add group. See Set up your team.

What roles are there?

Users, Department Admin, and Admin. Roles govern administration; groups govern document access.

Does CampusMind support SSO?

Yes, if it was set up for your institution. CampusMind federates with Entra ID — other OIDC or SAML providers on request. CampusMind does the setup during onboarding and your IT team maintains it after that, so ask IT for access to the application; an Admin still assigns your groups. Without SSO, people are invited by email and removed by hand, so plan for that in your joiner and leaver process. See Set up your team.

Cost and reporting

How is usage measured?

By AI processing: the analytics dashboard reports token counts and dollar cost, each split into input and output.

Can I attribute cost to a department?

Yes. Filter the dashboard by Groups, Users, or Date range, then use Export CSV. That's the route for chargeback or internal reporting.

How do I show what we've done for an audit?

Two levels. For one document, Download Report saves its before-and-after remediation report. For the institution, Export CSV from the analytics dashboard. The Avg before score and Avg After score pair is the clearest summary of what remediation changed.

Still have a question?

See Troubleshooting for failure states, or Support to reach the team.