A diagram in a course handout should communicate its lesson even when a student cannot see it. Adding a useful description can be awkward when the editable source is missing and only a PDF remains. A new Ally update brings that task into the course accessibility workflow—but choosing the right words still needs human judgement.
What changed on 12 October
Anthology's Ally 2.20 release notes announce PDF image alternative-text fixes for Blackboard LMS, Canvas, D2L Brightspace and Moodle integrations, released to production on 12 October 2026.
Instructors can manually describe PDF images through Ally Instructor Feedback. This option is enabled by default, although administrators can disable it. Saving the changes uploads a replacement course file and overwrites the previous version.
AI-generated suggestions are optional. The AI Alt Text Assistant must be enabled by an administrator; it is off by default. Suggestions consider surrounding PDF content, and an instructor must review and approve each description before it is applied. This is an Ally capability in supported institutional integrations, not a new function in every PDF editor.
Describe the lesson, not just the picture
Consider a hypothetical biology handout showing a flower. “A white flower” may identify the picture, but it does not explain the labelled parts the lesson asks students to compare. Start with the teaching purpose: what information would someone miss without this image?
For a chart or complex diagram, a short description may not carry enough detail. W3C's guidance on complex images distinguishes a brief identifying description from a longer explanation of essential information. Although its examples concern web content, that distinction is useful when planning a PDF's accompanying explanation. Do not squeeze an entire data table into a vague sentence.
A practical course-file checklist
1. Keep a recoverable original. Save an approved backup before changing the course copy. Record which course, handout and revision you are updating so colleagues do not accidentally restore an older file. 2. Start with one representative document. Choose a handout containing both a straightforward image and a more demanding diagram. Ask your accessibility team to review the results before repeating the process across a course. 3. Draft against the learning objective. Check labels, units and relationships against the original. If exact figures matter, provide them in an appropriate accessible explanation or table rather than guessing. 4. Treat an AI suggestion as a draft. Look for invented details, missing distinctions and language that gives away an assessment answer. Follow your institution's approved AI and document-handling policy. 5. Inspect the saved course version. Reopen the replacement through the route students actually use. Check that the description belongs to the correct image and makes sense alongside nearby text. Invite feedback about anything still unclear.
What this does not solve
Image descriptions are only one part of an accessible document. Headings, reading order, tables, language and keyboard or assistive-technology behaviour still deserve assessment. Neither an appealing visual layout nor this single fix establishes overall accessibility or standards conformance.
Do page assembly before the final accessibility review wherever practical. FileStation's Organize PDF can help arrange an ordinary working copy, and PDF Reader supports a visual check. FileStation does not write image alternatives into PDF tags or certify accessibility. Use your institution's specialist remediation and testing tools for that work, then recheck the exact document students receive.