Answer block format for crypto pages: the module pattern AI engines actually reward
AI engines extract content in 50-150 word chunks called answer modules. Pages built around modules outperform long-form prose by 3.1x in citation rate. Here is the exact format, where to place modules and how to retrofit existing pages.
What an answer module is
An answer module is a self-contained block of content that answers a specific question without requiring context from the rest of the page. AI engines extract these modules independently. One page with 8 well-formed modules can earn 8 distinct citations for 8 different queries.
The format: an H2 phrased as a question or noun phrase that matches a real query, followed by 50 to 150 words that resolve the question completely. No transitions like "In the previous section we discussed..." No setup like "Before we get to the answer let us first..." No filler.
The module is the unit of optimization, not the page. This is the single biggest shift in content strategy from classic SEO to AEO.
The structural template
Every answer module follows the same structure. Three elements, in order.
## [Question or noun phrase matching search intent]
**[40-60 word direct answer in plain prose.]** The first sentence states the answer. The next 1-2 sentences add the most important supporting fact or qualifier.
[Optional 50-90 word extension paragraph with specific numbers, mechanisms or examples. This is what gets quoted when the AI engine wants to add detail.]The bold first sentence is critical. AI engines weight the opening sentence of a section heavily. If the answer is hidden in paragraph 3 the engine misses it. Lead with the answer.
Where to put modules on a token page
A typical token detail page needs 8 to 12 modules. The specific questions vary by vertical but the structure is consistent. For a generic token:
- What is [token] and what does it do
- Where does [token] trade and at what volume
- How is [token] priced (mechanism, not the price itself)
- What is the supply schedule of [token]
- How does [token] governance work
- Is [token] audited and by whom
- How is [token] different from [closest competitor]
- How do I hold [token] safely
- What are the risks of holding [token]
- How is [token] taxed (general not jurisdictional)
- Where can I learn more about [token]
Where to put modules on a DEX or CEX page
Different category, different questions. For an exchange (DEX or CEX):
- What is [exchange] and how does it work
- How much volume does [exchange] handle
- What chains and tokens does [exchange] support
- What are [exchange] fees
- How does [exchange] handle custody and security
- How is [exchange] different from [competitor]
- What is the [exchange] token (if applicable)
- How do I withdraw from [exchange]
- Is [exchange] available in [region]
- How is [exchange] regulated
Retrofit pattern for existing pages
Most crypto sites have long-form pages that need restructuring not replacement. The retrofit:
Step 1. Read your existing page. List the 8 to 12 distinct facts a buyer might want. Write each as a question.
Step 2. Find the existing prose that answers each question. Extract it. Tighten it. Lead with the answer sentence.
Step 3. Insert H2 question headers above each extracted block. Reorder if needed so logical flow holds.
Step 4. Add FAQPage schema with the same Q&A pairs. The schema reinforces what the visible content already shows.
Step 5. Delete the transitional prose between modules. The page is shorter after the retrofit. That is correct. AI engines prefer dense modules over padded prose.
The 3.1x citation lift, measured
We tested this on 24 token pages across 6 protocols. 12 pages were left in long-form prose format. 12 were retrofitted to the module pattern. Same content, same total facts. Different structure.
Over 60 days the module-format pages earned 3.1x more citations across ChatGPT, Perplexity and Gemini combined. The lift was largest in Perplexity (4.2x), smallest in Gemini (1.9x), with ChatGPT in the middle (2.8x).
The cost of the retrofit was 90 minutes per page. The ROI is unambiguous for any project that gets meaningful traffic from AI engines or wants to.