University of Michigan

Building a Google Slides extension to cut accessibility remediation time by ~80%

Role

Product Designer

Context

Self-led project

Timeline

Jan - Apr 2026

Tools

Problem

The University of Michigan faced a federal ADA Title II compliance deadline requiring all student-facing course content to meet WCAG 2.1 AA standards.

Sociology faculty had tons of existing documents that weren't accessible with no realistic way to manually fix them.

The ADA Title II rule set a deadline for every student-facing course file.

Discovery

I was remediating slides alongside faculty every week and started tracking how long it actually took.

The math didn't work. A single faculty member would need 30+ hours just to remediate one semester of slides, on top of their typical teaching load.

A 40-slide lecture deck averaged ~90 minutes of manual fixes, and many professors had 30+ decks per course.

Faculty were overwhelmed.

Professors understood why accessibility mattered but couldn't justify the time cost against their teaching responsibilities.

The bottleneck was simple, repetitive fixes.

Many flagged issues were predictable and mechanical such as missing slide titles or insufficient contrast ratios.

The existing tool, Grackle, could flag issues but couldn't fix them, and each flagged issue had to be corrected by hand.

Grackle flagged accessibility issues, but every fix still had to be made by hand.

First Attempt

My initial solution was a standalone web app.

Faculty would upload a Google Slides link, the tool would scan and remediate, and they'd get a fixed version back.

It worked technically, but it was the wrong modality. It pulled professors out of their workflow, added steps, and felt disconnected from how they previously prepared lectures.

The web app worked, but faculty had to leave Slides to use it.

Pivot

I rebuilt the tool as a Google Apps Script extension via a sidebar panel that lives directly inside Google Slides.

The decision to move into the native environment changed adoption from “a new tool I have to remember to use” to “a button that’s already there.” I used Claude Code to research existing remediation tools and build both versions. Moving into Slides, and choosing which fixes to automate and which to leave to faculty, were my calls.

How It Works

The extension mirrors Grackle's audit parameters so faculty see familiar issue categories, then adds an auto-fix layer on top.

Google Apps Script extension Variants

The extension runs inside Slides and fixes the common issues in place.

Impact

A 45-slide deck that previously took ~90 minutes to remediate now takes ~20 minutes, with most of that time spent on issues that require human judgment.

"I went from dreading this to doing it between meetings."

— Sociology faculty member

Handoff

I wrapped up the project in April 2026, packed up the files and notes, and handed it off to the person taking it over.

I left them two things to pick up:

Security approval

The extension only runs on personal Google accounts for now. Approval from LSA Tech Services would let any faculty member install it on their university account.

U-M Maizey Integration

U-M’s own AI tool, Maizey, could write alt text and slide titles, covering the slowest fixes without outside API costs or privacy worries.

Reflection

The biggest lesson was that a viable solution can still be the wrong shape.

The web app solved the right problem in the wrong place. Rebuilding as a native extension was a design decision about where faculty attention actually lived. The best tool is the one that doesn't ask users to change their behavior.

Next project

PlaySmartScaling a coaching platform to 35+ schools and 300+ athletes.Sept 2024 - Apr 2025