The seven points
| # | Point | Scores 2 when |
|---|---|---|
| 1 | Relevant | The title is a real buyer question in the buyer's words, tied to a question we track |
| 2 | Answer first | The first 40 to 80 words answer it plainly and say who, what and where |
| 3 | Specific | Markets, sizes, services, certifications, outcomes and dates replace adjectives |
| 4 | Evidenced | Every claim has a source, a figure or a case; vendor research is labelled |
| 5 | Next question answered | The follow-ups buyers ask next are covered: exclusions, switching, cost drivers, timelines |
| 6 | Sections AI can lift | Each section answers one question and makes sense read on its own |
| 7 | True and current | Facts match the fact register and your listings; the page shows a review date |
The pass mark
11 out of 14, with no zero on points 2, 4 or 7. A page that buries the answer, makes claims it cannot support, or contradicts your approved facts does not ship, however well it scores elsewhere.
Page types we build to the Standard
| Type | Answers | Example (fictional) |
|---|---|---|
| Question page | One specific buyer question | Does cyber insurance cover ransomware payments in Singapore? |
| Fit page | Who it is right for, and who not | Group health for companies of 20 to 200 staff |
| Comparison page | A fair comparison, including when a rival fits better | Broker or direct: which suits an SME? |
| Switching page | Moving from another provider | Changing your broker mid-policy |
| Proof page | A case with sector, problem, change and number | How a regional distributor cut claim turnaround |
| Fact page | Plain, dated facts: services, markets, licences | Who we cover, where, and how |
Why we keep score
We record every page's score so we can see whether higher-scoring pages are cited more. Over time that builds our own evidence for what AI quotes, instead of relying on guesswork.
A broker's cyber page, before and after
6of 14
13of 14
Before: a 600-word page that opens with the firm's history and mentions ransomware once, in paragraph six. Score 6 of 14, with a zero on answer first. After: the title asks the buyer's question, the first 60 words answer it with the conditions that apply, a table covers what is and isn't included, and the page shows its review date. Score 13 of 14.
Frequently asked questions
Is this the same as SEO best practice?
It overlaps, but it is stricter about the opening answer, evidence and accuracy, because AI assistants lift passages rather than rank pages.
Do you apply it to existing pages?
Yes. In an audit we score your important pages and rewrite the ones the Answer Grid says to work on first.