Courage Horizon

Guide

Will an FAQ Page Actually Get You Mentioned by ChatGPT?

Courage Horizon · August 2026


Adding an FAQ page to your site does nothing for ChatGPT mentions unless the questions on it are the ones your buyers actually type before they buy, and most FAQ pages aren’t built that way. A typical FAQ answers shipping windows, return policies and accepted payment methods, because those are the questions support tickets bring in. An AI assistant answering a buyer question is looking for something narrower: a sentence that states the answer to a specific situation, in language close to how the question was actually asked. Call that a buyer-intent question, as distinct from a support question. An FAQ built from support tickets misses it almost entirely, however well it’s formatted.

What makes an FAQ answerable by an AI assistant, and what makes it invisible?

An AI assistant needs a passage it can lift, not a page it has to interpret. If an FAQ answer opens with the specific answer in its first sentence, rather than backing into it, that answer becomes citable on its own. If it opens with a preamble, “Great question, here’s everything you need to know about…”, before reaching the actual answer three sentences later, the model has to do the extraction work itself, and it usually cites whoever didn’t make it do that. This is exactly the gap worth measuring first: can an assistant lift a usable answer about you at all.

Does FAQ schema markup make the difference by itself?

Probably not much. FAQ schema, the structured data markup that used to earn a business a rich snippet in Google’s results, is often assumed to help with AI mentions too. In August 2023, Google’s own Search Central blog announced that FAQ rich results would be shown “only for well-known, authoritative government and health websites”, pulling that treatment away from ordinary commercial sites almost entirely. Google made that change based on what looked right in its results page, nothing to do with AI assistants, but it’s a clear sign the schema tag itself was never doing the work. The sentence underneath it was. A language model reads the words on the page when it forms an answer. It has no obvious reason to treat a schema-tagged question as more trustworthy than the same words written as an ordinary heading and paragraph.

Which questions should actually be on the page?

The ones a buyer types right before they choose, not the ones a company assumes are wanted. “Does this work for a sole trader in the UK” is a buyer-intent question. “What are your business hours” rarely is, unless the business is a restaurant and someone’s asking whether it’s open right now. The test is whether a stranger would type that exact sentence into a chat window while deciding. If nobody would, it can stay on the page for humans. It just isn’t doing anything for an AI assistant. An FAQ built the right way sits alongside other site fixes, like a properly structured llms.txt file, rather than replacing them.

How do you find out which questions your FAQ is missing?

Most businesses don’t know which buyer questions they’re missing until they watch an AI assistant answer around them, citing everyone except them for a question they never thought to write down. When we do this work for a business, the FAQ is written directly from the specific buyer questions the check found the business never answering, rather than from a generic template pulled from the industry. What changes the FAQ is not better writing. It’s knowing the actual question first.

What should you actually do with an FAQ, then?

Treat it as a delivery format, not a strategy. Format was never the problem. It’s one place, among several, to publish direct answers to buyer-intent questions, and it works under two conditions: the question is one a real buyer typed, and the answer opens with the answer itself rather than working up to it. Get both right and the page design barely matters. Get either wrong and no amount of markup or structure fixes it. For the broader question of how to get your website recommended by ChatGPT at all, an FAQ is one tactic among several, not the whole strategy.

Common questions

Should I delete my FAQ page if it isn’t generating AI mentions? No, not on that basis alone. An FAQ built from support questions still serves customers already on your site looking for shipping or returns information. It just isn’t the lever for AI mentions unless you also add the buyer-intent questions it’s currently missing.

How do I find out which buyer questions to add? Check your category in the assistants and keep the specific questions where you are not appearing, then write the FAQ from that list. It is one of the jobs we set up for a business, with the check repeated on a schedule.

Does the order of questions on the page matter? Less than the content of each answer. What matters is that each question and its answer could be lifted as a standalone pair, with the answer in the first sentence, not where that pair sits on the page.

Sources

  • Google, “Changes to HowTo and FAQ rich results”, Google Search Central Blog (2023): developers.google.com

Courage Horizon sets up AI agents for UK small businesses: software that does one job, such as chasing unpaid invoices, and asks you before it sends anything. Being found when customers ask ChatGPT is part of that work. Get a free report on your own business.

All guides