FAQPage is gone. Q&A content isn't. Mark it up like this.

FAQ content didn't get killed in May 2026 — only the rich-result schema reward did. Q&A is still one of the highest-utility content patterns on the web: it matches the way users phrase questions, it gives AI engines clear extraction units, and it gives screen reader users a predictable structure.

The mistake the SEO industry made was treating FAQPage JSON-LD as the goal instead of as one optional encoding. The actual goal is content that reads as questions and answers — and that goal is better served by semantic HTML than by JSON-LD, because semantic HTML benefits accessibility, AI extraction, and search simultaneously.

Joost de Valk's recommendation in May 2026 captured the shift: keep FAQ content where it's genuinely useful, drop FAQPage schema from pages where Q&A isn't the primary purpose, and use semantic HTML — the details and summary elements, plus question-style headings — as the primary representation. This article is the implementation guide for that recommendation.

We'll cover three patterns — collapsible disclosure, heading-based prose, and definition-list — plus when to keep JSON-LD, when to drop it, and how to test that screen readers and AI engines see what you intend.

The default pattern. Use the details and summary elements when you have a list of Q&A pairs and the page is dense enough that users will want to scan. Screen readers announce these elements natively, the markup is self-contained, and you don't need a single line of JavaScript or ARIA.

Wrap the whole list in a section with an accessible name. The summary holds the question. The body of the details holds the answer — and it can contain anything: paragraphs, code, lists, even further nested details. The open state can be controlled via the open attribute on the first item if you want it expanded by default.

The one gotcha: don't put interactive elements inside the summary (links, buttons). Modern browsers handle it but old assistive tech can choke. Keep the summary text-only and put any links inside the body.

When answers are long, when context matters, and when the FAQ is really an essay disguised as Q&A, the heading-based pattern beats the disclosure widget. Each question becomes an h3 (or h2, depending on outline), each answer is a paragraph or several. The page reads top to bottom, the table of contents is automatic, and search engines have an easier time identifying the answer span.

This is the right pattern for help center articles, expert Q&A interviews, and any FAQ where the answer benefits from being immediately visible. It's also the pattern AI engines tend to extract from most cleanly — the question heading becomes the anchor, the prose underneath is the answer span.

If you want both an overview and depth, combine the patterns: a disclosure block at the top for fast scanning, followed by heading-based sections for the long answers. Keep the same question text in both places so the page reads consistently.