Legal
Accessibility
What we have built, what was tested, what failed that test, and what is still not confirmed.
1. Our target
We aim to meet the Web Content Accessibility Guidelines (WCAG) 2.2 at Level AA. We do not currently claim to meet it.
2. Built and verified by test
Each item below was checked by driving the running site in a browser, not by reading the code.
- A skip link to the main content is the first focusable element on every page, and the main landmark now takes focus when it is used.
- Visible focus indicators on interactive elements, with a higher-contrast variant on dark backgrounds, including the interior page banners.
- The service drawer and the tool dialog keep keyboard focus inside them while open and mark the rest of the page inert. Those two, the assistant panel and the mobile menu all close on Escape and return focus to the control that opened them. The assistant is a corner panel rather than a dialog, so it deliberately does not trap focus and does not make the page inert.
- Elements that are visually hidden are also removed from the tab order.
- The two self-assessment tools present each question as a labelled group of radio options, move focus into the dialog body after each step, and write the result into a region that is already marked live.
- Touch targets of at least 44 by 44 pixels on coarse pointer devices for the navigation, the footer navigation links, the assistant quick replies, the tool options and the dialog close buttons. Links that sit inline inside a sentence, such as the email addresses in the footer address block, are sized by the text around them; WCAG 2.2 exempts inline links, and we are not claiming 44 pixels for those.
- Assistant replies and assistant status are announced through live regions.
- Animation, including the hero, is disabled when the visitor's system requests reduced motion.
- No page scrolls horizontally at 320 pixels wide or at 400 per cent zoom, checked with the live monitor boards loaded as well as empty.
- Text and background contrast meets the 4.5:1 minimum for normal text throughout, including the smallest labels.
3. Not confirmed
Stated in these words: NOT YET INDEPENDENTLY VERIFIED.
- No accessibility audit has been carried out by an external firm or by a certified auditor.
- The site has not been tested with screen readers by people who use them daily. Structural correctness has been checked; lived usability has not.
- No formal WCAG 2.2 AA conformance claim is made, and none should be inferred from this page.
Open defects we know about and have not yet closed. Each was found by independent audit and is listed here rather than left for a visitor to discover:
- The breadcrumb link at the top of interior pages measures 41 pixels wide against a 44 pixel target. Its height was corrected; its width was not.
- With JavaScript disabled, the service list, the standards explorer, the two self-assessment tools, the live monitor boards and the assistant do not work. A notice now appears in place of each, and controls that would do nothing are hidden rather than left visibly dead, but the content itself is not yet served as static HTML.
Defects reported in an earlier version of this page have since been fixed and re-tested by an independent auditor: the focus trap on the classifier dialog, the assistant panel making the page inert, and the monitor board tabs and filters losing keyboard focus. They are recorded here rather than quietly deleted.
4. Known limitations
The Live Monitors page reproduces content published by third parties. We control how it is presented but not how it is written, so headline wording, abbreviations and link text are outside our control.
The Contact page currently offers email only. A form is planned.
5. What version 1.0 of this page got wrong
We publish this because a firm that sells assurance does not get to quietly correct its own record.
Version 1.0 stated as tested fact that dialogs, drawers and the assistant panel trapped focus. They did not. There was no focus-trapping code anywhere on the site, the background remained fully tabbable behind every dialog, and the dialogs additionally told assistive technology the background was inert when it was not. It also stated that the assistant closed on Escape, which it did not, and that touch targets met 44 pixels, which several did not because the rule was written above the rules it needed to override. It further implied the smallest text had been brought above the contrast minimum when six text and background pairs were still below it.
Those claims were found by an independent adversarial audit commissioned for this site, were reproduced by measurement, and have been fixed. The lesson recorded here is the one we would put to a client: a statement that distinguishes verified from unverified is only worth anything if the verified half was actually verified.
6. Feedback
If you hit a barrier on this site, tell us at [email protected] and describe what happened and what you were trying to do. We will respond and, where we can, fix it.