Introducing Remilo: LMS Accessibility That Fixes Course Content at Source

Claire Watson
Marketing Manager

AHG 2026 was our first Accessing Higher Ground. Between Table 24 and Chris’s session, we saw only a slice of the programme. So we read all 156 published session abstracts to see what the community was saying as a whole.
Here are the five themes that stood out, with our working shown. If you think we’ve got something wrong, we’d like to hear it.
How we did this
We ran every published session abstract through an AI-assisted discourse analysis. It coded each one for topics, concerns and recommendations. Then we checked every theme and quote against the original abstracts by hand.
There are limits. Abstracts describe what presenters planned to say, not what they said in the room. Counts depend on how we defined each theme. And theme 4 is our own reading across several abstracts, so treat it as a prompt for discussion.
1. AI is welcome, but only with a human in the loop
64 of 156 sessions (41%) involved AI. 37 of those (58%) also raised human review, quality or risk. No session we read presented AI as a replacement for skilled staff.
The shared position is “AI drafts, a person decides.” That’s a sensible place for the field to be.
Still open: what does good review look like at scale, when one person is checking hundreds of documents a term?
2. The mood is fatigue, not urgency
After the deadline shifted, UTSA described the feeling on the ground as “frustration, fatigue, uncertainty.” Many sessions were less about racing a date and more about defensibility and keeping the work going over the long term. [CHECK: add one or two concrete examples from the abstracts, then delete this note.]
Still open: how do you keep momentum when the external pressure eases but the backlog doesn’t?
3. Teams are building their own toolkits
Several workshops taught staff to remediate with prompt libraries for general-purpose tools such as ChatGPT, Gemini and Copilot, and with specialist tools like DAISY’s Fido. Several universities described building their own remediation pipelines.
Practitioners are resourceful, and they want control over their own process. That came through clearly.
Still open: where does a trained person with a good prompt library hold up, and where does it start to struggle? Consistency across a whole course, complex content and time spent reviewing all came up.
4. Maths is still unsettled
We saw three different approaches taught for making equations accessible. Each one raises its own question:
- Describe it. When is a description enough for a student working through the maths?
- Extract structure. How do you handle legacy files that weren’t authored that way?
- Rebuild the page. Is it sustainable across a full course or department?
No session we read settled when a description is enough and when a student needs the structure. This is our reading across abstracts, not a tested finding, and it’s the question we most want to hear your view on.
5. Buyers want proof and are wary of scores
Procurement sessions said the same thing in different ways: a vendor’s own conformance report (ACR/VPAT) is a claim, not evidence.
A University of Washington session warned that accessibility scores can have “negative side-effects”. Our own reading is that a score can reward what’s easy to measure over what matters to a student, or signal “done” too early.
Still open: if not a score, what evidence would you trust when choosing a tool, a vendor or your own process?
Tell us what we missed
This is a first pass, and it will have gaps. If you presented at AHG and think we’ve misread your session, or you saw something in the room that the abstracts didn’t capture, get in touch. We’ll update this post and credit corrections if you’d like.
About Remilo. Remilo finds and fixes accessibility issues in course materials, right inside your LMS, including equations, charts and scanned slides. It’s built by Estendio, an accessibility tech company in Glasgow and the team behind Tailo and Present Pal. If you’re weighing up tools, our remediation checklist works for any of them, including ours.