When to use this prompt
When a service page, blog post, or pillar page lacks FAQ schema and you need to add 5 to 10 questions that read naturally and match the page’s existing claims. FAQs are one of the highest-leverage content additions for AI search because they map directly to FAQPage schema and they answer buyer questions in the format LLMs prefer to extract.
Run this on every priority page that does not already have an FAQ section. Run it again if you add a new section to an existing page.
The prompt
<role>Content strategist specializing in FAQ design for SEO and AI search.</role>
<task>Generate 7 FAQs grounded entirely in the page content below, formatted as YAML ready to drop into FAQPage schema.</task>
<inputs>
<page_content>
[PASTE FULL PAGE TEXT]
</page_content>
</inputs>
<instructions>
1. Each question must be one a real buyer in this category would actually type or speak. No marketing-internal questions like "What are the benefits?".
2. Each answer must be 2 to 4 sentences and 50 to 90 words.
3. Each answer's first sentence must directly answer the question. No preamble.
4. The 7 questions together must cover, in this order:
1. Definition
2. Comparison to a similar but distinct category
3. Pricing or value
4. Prerequisites or who it is for
5. A common objection
6. An edge case
7. How to start
5. Every claim in every answer must be grounded in the page content. Do not invent facts. If the page does not address one of the seven topics, write the question and put "Not addressed on the page" as the answer rather than fabricating one.
6. Do not output anything outside the YAML block.
</instructions>
<output_format>
faqs:
- question: "..."
answer: "..."
- question: "..."
answer: "..."
(continue to 7)
</output_format>How it works
The seven-question framework (definition, comparison, pricing, prerequisites, objections, edge cases, how-to-start) is the buyer journey in compressed form. Most FAQ sections are weak because they only cover the first three. The latter four are where AI engines find the specific, citable answers most pages do not provide.
The “Not addressed on the page” fallback is a 2025 hallucination guard. Frontier models will fabricate plausible answers if you do not give them an explicit out. This phrasing also surfaces the gap so you can decide whether to add coverage to the page.
The YAML output format drops directly into Astro content frontmatter or any FAQPage schema generator without copy-paste cleanup.
Example output
faqs:
- question: "What is product analytics software?"
answer: "Product analytics software is a system that tracks user behavior inside a digital product, measures engagement and retention, and surfaces patterns that explain why users convert, churn, or get stuck. It centralizes event data in a single warehouse, supports cohort and funnel analysis, and lets product teams test changes against real usage. Modern platforms add AI-assisted insights and integrations with data warehouses and CRMs."
- question: "How is product analytics different from web analytics?"
answer: "Web analytics measures top-of-funnel traffic patterns: pageviews, sessions, traffic sources. Product analytics measures what users do inside the product after they sign up. Most teams need both, but product analytics is what answers retention and feature-adoption questions. If your priority is marketing acquisition, web analytics is the right fit. If your priority is understanding why users stay or leave, product analytics is the right tier."Variations
- Schema-first version: Add a constraint that each answer must be exactly 50 to 75 words. Tightens the output for snippet eligibility.
- Voice search version: Reframe questions in the way they would be spoken aloud, not typed. Useful for voice assistant optimization.
- Buyer-stage targeted: Generate three FAQs each for top-of-funnel, mid-funnel, and bottom-of-funnel queries instead of the seven-question framework.