The automated pass checked 36 pages on bekee.com against WCAG 2.2 AA and found no automated violations. Mechanical checks (contrast, ARIA, structure, alt presence, label presence) all passed.
The AI judgment pass flagged 23 items across 14 of 36 pages (39%): heading, link purpose, page title. Each is a wording or clarity judgment, not an automated failure; confirm before acting. Of those, 3 are an active action item; 20 you have reviewed and accepted, grouped under "Noted, not action items" below.
The link text 'something this morning' does not describe the destination or content. A screen reader user encountering this link out of context cannot tell what it leads to. The surrounding context identifies the sender (Justin Welsh) but still leaves the subject of the linked content unclear.
Suggested: Justin Welsh's newsletter issue on there being no map (opens in a new tab)
The page title leads with what reads as an article headline rather than a clear description of the page's topic or purpose combined with the site name. The number '230' is unexplained in the title, making it opaque to someone who has not yet opened the page. A user scanning browser tabs or history cannot immediately tell this is a notes/blog article on a web accessibility auditing site.
Suggested: What a Free Scan Can and Cannot Tell You - bekee gibson
The link text 'that post is here' does not identify the destination or subject of the linked post. Someone encountering this link in context would need to read the surrounding sentence carefully to infer the topic, and the phrase 'here' is the classic ambiguous link pattern.
Suggested: measuring the right pixels: the bug I found in my own scanner
These images look like charts, graphs, diagrams, screenshots, or maps based on their filename or alt text. No automated tool can measure contrast inside an image. Confirm that parts required to understand each graphic have at least a 3:1 contrast ratio against adjacent colors, and that meaning is not conveyed by color alone. Plain photographs are exempt from the contrast requirement.
Inline SVGs can carry information through color: charts, diagrams, map regions, status icons. Confirm that meaning is also conveyed by shape, text, label, or pattern, and that graphical parts required for understanding have at least a 3:1 contrast ratio against adjacent colors. Decorative SVGs (aria-hidden) were excluded.
Submit the form with required fields left blank and confirm that required fields and any error states are indicated by more than color (asterisk, icon, label text, border thickness). Any color cue used must also meet contrast and must never be the only signal.
The same destination href is reached by links with different visible text on different pages (and 11 more). Screen-reader users who navigate by label expect a consistent name for the same destination.
Suggested: Pick one label per destination and apply it everywhere. The clearest, most descriptive variant is usually the right choice; update the rest to match.
Other scanners set aside 9 color-contrast checks across 3 pages as "needs manual review." The reason it gave: the element holds symbols or punctuation, not words. We measured every one against the rendered page at desktop (1280×720) and mobile (390×844), worst case: 9 pass, 0 fail. Nothing failed, so there is nothing here to act on.
Accessibility is not a certificate you earn once and hang on the wall. It is a continuum. Sites change, content gets added, and the WCAG guidelines themselves keep evolving; a site that passes today needs another look as it grows.
This audit covers what automated and AI-assisted testing can evaluate, which is a meaningful share of the WCAG 2.2 AA criteria. It does not include testing with assistive technology by the people who depend on it, such as screen-reader users; that work takes specialized expertise and equipment beyond this engagement, and I will always tell you plainly where a site reaches the limit of what I can verify. No audit, from anyone, certifies a site as one hundred percent accessible. The honest goal is steady, measurable improvement.