Accessibility
This statement is about the Resolutr software: the client portal, the enquiry and intake forms a mediation service publishes, the payment page, and the screens the service's own staff use. It was prepared on 12 September 2026 and describes what we know on that date. It is not about any individual mediation service's own website.
The standard we are working to
We are working to WCAG 2.2 level AA. That target is recorded in the product's own signed architecture decision and in its design system, and it is what the automated checks in our build are written against.
This is not a claim that we meet it. The software is partially conformant: some of it has been assessed and fixed, some of it has not been assessed at all, and the two largest checks — an independent audit and testing with assistive technology — have not happened. They are listed below.
What has been done
- An internal self-audit against the standard above, completed on 23 August 2026, followed by a programme of fixes. It found thirty-five problems. Among them: every text box in the product had lost its keyboard focus outline; forty-four controls on the client-facing screens had no name a screen reader could read, worst on the form where a client types their income and benefits; a failed submit on that form threw away eleven fields the client had typed; and there was no "skip to content" link anywhere.
- Automatic structural checks that run on every change to the software. They read the HTML the application really produces and refuse a build that loses the skip link, the page landmark, the page language, a single first-level heading, a name on a form control or on an icon-only button, a caption or a header on a data table, a working reference between an error message and the field it belongs to, or a large-enough target on a checkbox or radio button.
- A repeatable measurement of the colour contrast of every text and background pair the design uses, in both the light and the dark appearance. Both are recorded, and a change that made any pair fail would fail the build.
What we know is not done
These are stated because a statement that claims more than the evidence supports is worse than none.
No independent audit has been done
Everything below is our own assessment of our own software. No external accessibility specialist has examined it. One is required before launch and has not been commissioned yet.
Nothing has been tested with assistive technology
No part of this software has been used with a screen reader or with speech input as part of this work. A page can pass every automated check and still be unusable with a screen reader, so this is the largest single gap in what we know.
Zoom has been reasoned about, not measured in a real browser
Layout at small and magnified sizes was measured by loading pages in a headless browser at fixed pixel sizes. That is not the same as a person enlarging text or zooming to 200% or 400% in their own browser, which we have not measured.
High-contrast and forced-colours modes were emulated, not run
The software carries rules for forced colours and for a raised-contrast preference, and those rules were checked by emulating the modes. They have not been run under Windows High Contrast on a real machine.
The automatic checks read a named list of pages, not every page
The checks run on every change over a named list of pages chosen to cover the staff, client, administrator and public parts of the software — not all of its pages. They also cannot see colour contrast, whether the focus outline is visible, the order keyboard focus moves in, or anything that needs judgement, such as whether an error message is actually useful.
The payment page is checked, but not as a real payment link opens it
The card payment page is checked automatically with its payment lookup substituted. The same page as it is reached by a live payment link is not, and the card entry field itself — which is supplied by our payment provider — has not been verified.
Two identical unpaid lines on one staff screen read alike
On the mediator pay screen, two unpaid lines for the same person, matter and day at the same rate are announced identically, because the table does not show anything that would tell them apart.
If you cannot use part of this service
No address has been published for this yet. If you are using this service because a mediation service is handling your case, tell that service — they can reach us. If you are a member of staff at a mediation service, tell whoever administers Resolutr for your firm.
If your question is about your own case or your own information rather than about using the software, your mediation service is the right place to ask — the privacy notice explains why.
How this statement was prepared
By the engineers who build the software, from the code and from the automated checks described above. It is a first-party assessment. It is not an independent audit, not a formal conformance report, and not a legal compliance declaration. Where the evidence is a reasoned argument rather than a measurement, we have said so.