Level AA · WCAG 2.1

1.3.5 Identify Input Purpose

Fields asking for the user’s own information have to declare what they collect, so browsers and assistive tools can fill them in.

Partly automated Some failures are detectable, others need judgement.

How it fails in practice

A checkout with a dozen fields and not one autocomplete attribute, so every address has to be typed again by hand.

How to fix it

Add the right autocomplete token to each field: given-name, email, postal-code, cc-number. It is a one-line change per field and it helps every user, not only the ones the criterion is written for.

At a glance

Level
Level AA — the level EN 301 549 requires, and therefore the level the BFSG and the EAA are measured against.
Principle
Perceivable
Guideline
1.3 Adaptable
Added in
WCAG 2.1
EN 301 549
9.1.3.5
Automated check
Partly automated

Read the official definition at W3C

Partly automated

How far an automated check gets

A tool finds some failures here and cannot see the rest. A clean automated result means the checks that exist found nothing — not that the criterion is met. This is the gap that makes a green scan look like an accessible site, and it is where our agents and a human reviewer do the remaining work.

A wrong or missing token on a field that obviously wants personal data is detectable. Whether a custom field should have carried one is a judgement about your form.

Find out which criteria your site fails

We scan up to 100 pages against WCAG 2.2 AA, run the agent protocol on one key journey and send you the report — free, once, no strings.

Get a free scan See how agents test

Report within one working day.