Weak citability for AI answer engines

What it detects

Content-type pages (blog, news, docs, etc.) whose prose scores poorly on self-contained, directly-answerable passages.

Why it matters

AI answer engines quote short, self-contained passages that directly answer a question; vague or context-dependent writing ('see our pricing page for details') is unlikely to ever be cited, no matter how accurate it is.

How to fix

- Identify which content passages on the page scored weakly — look for paragraphs that are vague, don't stand alone out of context, or point elsewhere ("see our pricing page for details") instead of answering directly. - Rewrite weak passages as self-contained answers: lead with the direct answer to the implied question, then support it with concrete figures, dates, or named sources (e.g. "Plan A costs $29/mo and includes X, Y, Z. According to our Q1 2024 report, 73% of users on Plan A saw a 40% reduction in setup time."). - Where relevant, add a clear question-style subheading above the passage ("How much does Plan A cost?") so the answer block is easy for an AI search engine to isolate and cite. - Verify: re-read the rewritten passage in isolation (no surrounding context) and confirm it fully answers a plausible question on its own, with at least one concrete statistic or dated fact.

Example

"Plan A costs $29/mo and includes X, Y, Z" is citable on its own; "see our pricing page for details" is not.

References

• llms.txt specification
• Scovant scoring model — citability