How to Choose an Accessibility Monitoring Tool

The accessibility tool market has exploded. There are now dozens of scanners, and they all claim to make your site compliant. Choosing the right one matters because the tool you pick determines what you see, what you miss, and how your team responds.

This is a framework for making that decision.

Step 1: understand what automated scanning can and cannot do

Before evaluating any tool, calibrate your expectations.

The best automated scanners detect 30-50% of WCAG issues. The rest require human judgement - understanding whether alt text is meaningful, whether a page makes sense to a screen reader user, whether cognitive load is reasonable.

No tool catches everything. Any vendor claiming "full WCAG compliance" through automation alone is misleading you. The FTC fined accessiBe $1 million in 2025 for exactly this kind of claim.

What automation does well:

Detecting missing alt text, labels, and landmarks
Checking colour contrast ratios
Identifying ARIA errors and misuse
Finding structural HTML problems
Keyboard focus testing (if the tool does real browser testing)

What automation cannot do:

Judge whether alt text is meaningful
Evaluate the reading order for screen readers
Assess cognitive accessibility
Test complex interactive patterns in context
Determine if error messages are helpful

Step 2: ask about the scanning engine

This is the single most important technical question.

Most scanners use one engine - typically axe-core (84 rules) or the WAVE engine. Some use Lighthouse's accessibility audit (roughly 50 checks).

The limitation of any single engine is coverage. axe-core is excellent at what it checks, but there are entire categories of issues it doesn't cover: keyboard traps, focus obscured by sticky elements, non-text contrast, touch target sizes, alt text quality beyond presence.

Questions to ask:

Which engine(s) does the tool use?
How many rules does it check?
Does it do real browser testing (Playwright, Puppeteer) or DOM-only analysis?
Does it check WCAG 2.2 criteria or only 2.1?
Can it detect keyboard navigation issues?

Multi-engine scanners that deduplicate results across engines will catch more real issues than any single engine.

Step 3: evaluate the workflow, not just the scan

Finding issues is step one. Fixing them is where the real value is.

Look for:

Severity ratings - Are issues prioritised by impact? Can your team focus on critical issues first?
Fix guidance - Does the tool tell you what to change? Code examples? Before and after comparisons?
Tracking over time - Can you see your score trend? Know when regressions happen?
Team features - Can you assign issues to developers? Comment? Track progress?
Integration - Does it connect to your existing tools (Jira, Slack, CI/CD)?

Step 4: check the pricing model

Accessibility tool pricing varies dramatically:

Per-page - You pay based on how many pages are scanned. Can get expensive for large sites.
Per-site - Flat fee per domain. More predictable.
Per-seat - Charged per team member. Watch for this hidden cost.
Tiered - Feature tiers with different capabilities at each level.

The most important thing is understanding what your actual monthly cost will be when you're scanning the number of sites and pages you need.

Step 5: run a real comparison

Don't trust marketing pages. Scan your own site with 2-3 tools and compare the results.

You'll quickly see:

Which tool finds more issues
Which provides more actionable guidance
Which interface your team prefers
Whether the pricing matches the value

Lumi offers a free scan with no signup required, and a live comparison tool at lumi.livana.io/compare that lets you see the difference between single-engine and multi-engine scanning on your own site.

Summary

The best accessibility monitoring tool is the one your team will actually use consistently. Prioritise:

Engine coverage (more engines means fewer blind spots)
Actionable fix guidance (not just issue lists)
Workflow integration (it needs to fit how your team works)
Honest about limitations (any tool claiming 100% coverage is lying)
Price-to-value ratio (enterprise features shouldn't require enterprise pricing)