FAQ pages are usually poor content design. AI does not change that.
FAQ pages are one of the most persistent content design failures in government websites. They look helpful. They feel productive to build. And with AI-generated summaries and answer engines now a real factor in how people find information, there is a growing claim that FAQ pages improve search visibility and machine retrieval.
That claim has some basis. But it mostly misses the point.
Why FAQ pages exist
They are easy to produce. A subject matter expert lists the questions they get asked most often. A content team formats the answers into an accordion. The result looks comprehensive.
Nobody had to design anything. They just collected questions.
That is the problem. A well-designed page answers users' questions within the flow of the content. If the answer belongs in the main content, it should be there. If it belongs on another page, there should be a clear link. A FAQ section usually means neither decision was made, and the content got parked somewhere in between.
What FAQ pages actually reveal about your content design
FAQs are question-shaped, not task-shaped. Users typically want to do something: apply, check eligibility, understand what happens next, find out whether they qualify. A list of questions organised around what an organisation assumes people ask does not map to that.
They also increase cognitive load. A user has to scan through a list of possible questions, identify which one sounds closest to their situation, and then interpret the answer. That is harder than following clear headings through well-structured content.
Critical information hides in FAQ sections. Eligibility criteria, deadlines, exceptions, and process detail frequently end up buried in expandable sections instead of being surfaced where people need them. And FAQs are easy to add and hard to govern. Teams keep appending questions without reviewing duplication, overlap, or whether the answer is already published elsewhere.
This is a pattern I see consistently. FAQ sections are one of the most reliable markers of a page that has not been properly designed. The questions may be real. The need may be genuine. But the format was a substitute for structure, not a solution to a user need.
What AI changes — and what it does not
There is a reasonable case that question-and-answer formatting helps with machine retrieval. FAQ-style phrasing can mirror natural language queries. Clean question-answer pairs are easier for language models to parse, extract, and surface in AI overviews or direct answer boxes.
That is true. It does not follow that the solution is more FAQ pages.
AI does not make poor information architecture retrievable. It changes the retrieval environment, not the user need.
A page that buries critical information in a poorly labelled accordion will still fail the users who visit it directly. Making it slightly more parseable for a language model does not fix the underlying problem. And if users reach content through an AI summary, the question becomes whether the source content is accurate, maintained, and structured well enough to be cited reliably — not whether it was written in Q&A format.
The organisations that benefit most from AI-enabled search are those whose content is clearly structured, task-based, and well maintained. That describes the opposite of most FAQ pages.
What to do instead
The goal is not to optimise for a format. It is to make answers findable, usable, and maintainable.
In practice, that means:
Task-based pages with question-led headings where they add clarity. "Who can apply?" is a better heading than "Eligibility." It works for users navigating the page and for machines parsing the structure.
Critical information embedded in the main content. Costs, timelines, eligibility, exceptions, and next steps belong in the body of the page, not tucked into a separate section.
Short Q&A components only for genuine edge cases. A tight, governed set of questions at the end of a task page can work when the questions are truly bounded and would disrupt the main flow. Not as a default.
When a small FAQ section might still be justified
Support contexts with a genuinely bounded, stable set of user questions are probably the clearest case. Even then, each question is a maintenance commitment. Who reviews them? Who retires the ones that are no longer accurate?
Without clear governance, a "small" FAQ becomes a dumping ground within 12 months.
The problem is not that users have questions
The problem is treating a list of questions as a content strategy.
If your FAQ section exists because users are confused by the content around it, the answer is to fix the content. If it exists because the information architecture is unresolved, the FAQ will not resolve it. And if it exists to optimise for AI retrieval, know that AI rewards clear, structured, maintained content — which a well-designed page delivers without the accordion.