Headless CMSTechnical SEO
We wrote an llms.txt for our agency site, here's what we put in it
Answer engines don't crawl like Google. We shipped an llms.txt file that tells AI crawlers exactly who we are, what we build, and how to ...
The Blog
FAQ pages that Answer Engines can actually use
Headless CMSTechnical SEO
Every business website has an FAQ page. Almost none of them are doing anything useful. The typical FAQ is a dumped list with no grouping, no structure, and no thought about how search engines or AI tools consume the content. FAQ content directly answers the questions prospects are asking, the problem isn't the content, it's how it's structured.
The problem
Most FAQ pages are a missed opportunity
Every business website has an FAQ page. Almost none of them are built to do anything useful beyond existing.
The typical FAQ is a single list of questions dumped into a page with no grouping, no structure, and no thought about how the content will be consumed by anything other than a human scrolling through it. It sits in the footer navigation, rarely updated, and does almost nothing for SEO or AI discoverability.
That's a waste. FAQ content is some of the most valuable material on your site — it directly answers the questions your prospects are asking. The problem isn't the content. It's how it's structured.
The architecture
How we model FAQ content in Payload CMS
We treat FAQ as a content architecture decision, not a page template. The structure is designed to serve three audiences at once: humans browsing, search engines indexing, and AI tools extracting answers.
- 01Grouped by topic, not dumped in a list
Our FAQ page organises questions into topic groups, platforms, delivery, integrations, support, pricing. Each group has its own title and description. This matters because a visitor looking for pricing information shouldn't have to scroll past fifteen questions about your tech stack to find it.
- 02Anchor IDs on every group
Each FAQ group gets an anchor ID, which means you can deep-link directly to a specific topic: `/frequently-asked-questions#pricing` or `/frequently-asked-questions#delivery`. This is valuable for sales conversations ("here's the answer to your question"), for internal linking from blog posts and service pages, and for AI tools that cite specific sections of a page.
- 03Accordion UX with content in the DOM
Questions are rendered as accessible accordions, collapsed by default, expandable on click. But the answers are in the HTML whether the accordion is open or not. This is the detail most implementations get wrong: if the answer only loads when the accordion is clicked, search engines and AI tools can't see it. The content needs to be in the DOM from the first render.
- 04FAQ as a CMS collection, not hardcoded content
Individual FAQs are stored as their own collection in Payload, then related to FAQ groups. This means the same question can appear in multiple contexts — on the main FAQ page, on a relevant service page, in a blog post sidebar. The content is authored once and surfaced wherever it's needed.
The SEO layer
FAQPage schema that generates itself
Each FAQ group automatically generates FAQPage structured data. Google uses this for FAQ rich results, those expandable question-and-answer dropdowns that appear directly in search results.
Because the FAQ content is modelled as structured data in the CMS (not free-form rich text), the schema.org markup generates from the content itself. Add a new question to a group, and the FAQPage schema updates automatically. No manual JSON-LD, no developer ticket.
This connects directly to the structured data architecture we've built across the site, every content block type has its own schema generator, and FAQ groups are no exception.
One important detail: Google only accepts one FAQPage per URL. If you have multiple FAQ groups on a single page, they need to be collapsed into a single FAQPage with a combined question array, not one FAQPage per group. We handle this automatically in the merge layer.
The AEO layer
Why FAQ architecture matters for AI tools
This is where FAQ content architecture connects to the llms.txt approach we covered last week.
AI tools answering questions like "how much does a headless CMS build cost in Australia" or "how long does a PayloadCMS project take" are looking for clear, direct answers they can cite. A well-structured FAQ page with grouped topics, anchor IDs, and plain-language answers is exactly what these tools need.
Our FAQ page is referenced directly in our llms.txt file. When an AI tool follows that link, it finds content that's already organised by topic, with each question and answer clearly separated. The anchor IDs mean the tool can cite a specific section rather than just pointing at the whole page.
This isn't about gaming AI tools. It's about making your most useful content findable and extractable by every channel, human, search engine, and language model.
The sales layer
FAQ as a sales tool, not just a support page
Well-structured FAQ content does double duty in sales conversations. When a prospect asks "how does your pricing work" or "what's your typical timeline," you can link them directly to the relevant FAQ group with an anchor URL instead of typing out the answer again.
This is content that works harder because it's structured well, not because more of it exists. The same FAQ page serves organic search visitors, AI tool citations, and direct sales conversations. The investment is in the architecture, not the volume.