Does ADA Title II apply to your Confluence knowledge base?
If you work at a public university, a school district, a city, a transit agency or a public library, the answer is probably yes — and the date is closer than most teams think.
The rule, in one paragraph
In 2024 the US Department of Justice issued a final rule under Title II of the Americans with Disabilities Act setting WCAG 2.1 level AA as the technical standard for web content and mobile apps provided by state and local government entities. The compliance dates were subsequently extended by a year: entities serving populations of 50,000 or more have until 26 April 2027, and those serving fewer have until 26 April 2028.
Who counts as a covered entity
Title II covers state and local government and their instrumentalities. In practice that means city and county governments, public school districts, public colleges and universities, transit agencies, public libraries, and special district governments. It also reaches content provided on behalf of those entities — which is where contractors and shared service providers get pulled in.
Where Confluence sits
The rule is about web content the entity provides or makes available. It is not limited to the marketing site. Three Confluence patterns commonly fall inside it:
- Public spaces. Anonymous access turned on for a space makes it web content the entity provides.
- Service desk help centres. A Jira Service Management portal with a knowledge base is a public service interface.
- Content published out of Confluence. Exported, themed or embedded into a public site — the source of the accessibility problem is still the page.
Internal-only wikis are a different question. The rule's scope is public-facing content, but employment obligations under Title I and, for many institutions, Section 504 and state law reach internal materials too. If a staff member who uses a screen reader cannot use the internal onboarding wiki, the legal theory is different but the outcome is the same.
The exceptions, and why they help less than they look
The rule includes limited exceptions — archived web content, pre-existing conventional electronic documents, third-party content not posted by the entity, and individualised password-protected documents. Two things to note. First, most of them are narrow and conditional. Archived content must actually be marked as archived and kept for reference, not simply old. Second, an exception is a defence you have to be able to demonstrate, which means knowing what is in each space and why. That inventory work is the same work as measuring conformance.
How to find out where you stand
- List your public-facing spaces. In Confluence Cloud, check which spaces have anonymous access and which service desks expose a knowledge base.
- Measure one of them. Not a sample of five pages — the whole space, so the number means something. Expect the failures to be concentrated in a few checks.
- Separate mechanical from judgement. Alt text, headings, tables, contrast and link text are mechanical and can be fixed at scale. Reading order, media alternatives and meaning need a person.
- Keep dated evidence. A conformance report showing what was checked, when, and what was found is what a review asks for — and what shows good-faith progress if you are not finished.
The realistic timeline
Remediating a large space is not a sprint you can run in April 2027. If a space has 5,000 pages and 40% have at least one level A failure, that is 2,000 pages to touch. At even five minutes each, with the pages spread across dozens of owners, the project needs a year of steady work and a way to stop new pages from adding to the pile.
Which is the actual argument for starting now: not the deadline, but the throughput.