RTA Projects All Articles
Design

Accessibility as a Creative Constraint: What Changes When WCAG Shows Up at Kickoff

By RTA Projects Design
Accessibility as a Creative Constraint: What Changes When WCAG Shows Up at Kickoff

Photo: Software: Open Observatory of Network Interference Screenshot: VulcanSphere, BSD, via Wikimedia Commons

Somewhere in your organization, someone has a spreadsheet with a compliance date on it, and that spreadsheet is quietly doing the work of a design decision.

For public entities in the US, the deadline math has already shifted. The Department of Justice's 2024 rule under Title II of the ADA requires state and local governments to meet WCAG 2.1 Level AA — April 2026 for the larger ones, April 2027 for the rest. That first date has come and gone. The second is close enough to touch. And plenty of private-sector teams are reading the room and doing the same work anyway, because demand letters don't usually wait for a rulemaking.

The problem isn't the deadline. It's the shape of the response: a scurry in Q1, an overlay bought in Q2, and a remediation sprint that touches four hundred pages and fixes maybe a third of them.

Accessibility doesn't work like that. It's not a cleanup task. It's a design brief — and a good one, the kind a sharp creative director would have written anyway.

What the rule actually asks for

Short version: content that can be perceived and operated, by everyone, built to a published standard. WCAG 2.1 AA covers the stuff you already argue about behind closed doors — contrast, focus states, keyboard paths, labels, error messages, headings that mean something, captions, time limits, motion that can be paused.

Read that list again. Nothing in it is exotic. Contrast is a visual hierarchy conversation. Focus states are an interaction design conversation. Headings are an information architecture conversation. Error messages are a UX writing conversation. If your team fights about those things, you're already doing accessibility work. You're just not calling it that, and you're probably not checking it.

The overlay detour

Automated overlay widgets promise a shortcut: one script, a compliance badge, an icon in the corner. They're popular for a reason. They're also a bad bet.

Most overlays address a fraction of the standard. They can't fix a heading structure that's wrong, a form that doesn't tell a screen reader which field failed, or a component that traps focus. Some interfere with assistive technology users already have configured. And from a design standpoint, an overlay solves nothing — the underlying page is still exactly what it was. Bolting a widget onto an inaccessible site is like adding a ramp to a building whose front door is welded shut.

You still have to do the design work. The overlay just delays the bill.

Where the constraint makes the design better

Here's the part that gets lost in compliance conversations: these constraints are generative. They produce better decisions.

Take hierarchy. If someone navigating by headings needs to understand your page, those headings have to be real, sequential, and descriptive. Which means you can't have nine things styled as H2 because a designer liked the weight. You have to decide what actually matters most. That's not a technical fix — that's editorial clarity, and it makes the page easier for everyone.

Take forms. Labeling every field, describing errors in plain language, announcing state changes — that work forces you to design error states properly instead of screenshotting the happy path and shipping. Most conversion problems live in exactly those states.

Take motion. If animations must be pausable and respect reduced-motion preferences, teams stop using movement as decoration and start using it for emphasis. Usually that means less of it. Usually the page gets calmer.

Take navigation. Keyboard operability demands a logical tab order, which demands a logical reading order, which means someone had to make a decision about sequence. Sequence decisions are design decisions.

Component calls get cheap when they're early

The least expensive accessibility work happens at the component layer, before there's anything to remediate. Decide once — how our modals trap focus, how our nav behaves at 200% zoom, what our minimum contrast ratio is on brand color pairings, what inline validation looks like — and the decision propagates.

The expensive version is what most teams do: ship, audit, discover the same broken component on sixty pages, then argue about whether to fix the component or patch the pages.

Design systems are the leverage point. If your tokens include contrast-checked color pairs and your components ship with accessible states baked in, remediation turns into maintenance instead of a project.

Content is where the real backlog lives

Nobody warns you about this part. A large share of accessibility debt isn't code — it's content. Alt text that reads image1-final.jpg. PDFs scanned as images. Headings used as font sizes. Link text that says 'click here' eleven times on one page. Video with no captions.

That's editorial work, and it's usually bigger than the dev ticket. Which is why the smart move is to fold it into the content workflow instead of retrofitting it: alt text authored by whoever writes the page, captions required before a video goes live, a standing rule that any PDF needs a real HTML equivalent.

If you're already behind

Triage works. Here's a reasonable order.

Inventory your templates and your top tasks — the things people actually come to do — rather than every page equally. Fix templates, not pages; a shared header repaired once beats a header repaired four hundred times. Prioritize task completion: can someone apply, pay, register, find a location, file a form? Those flows go first. Bring in real assistive technology testing rather than relying on an automated scanner, because automated tools catch a slice of issues and miss the ones that matter most. And write down what you fixed, because the second pass is always faster than the first.

Put it in the kickoff

One line in the brief, every time: this project meets WCAG 2.1 AA, and we test with keyboard and screen reader before launch.

That's it. Not a scramble in March. A constraint from day one, doing what good constraints do — narrowing the options and forcing better ones.