Which parts of my site affect how I show up in generative AI answers?

The pages that affect generative AI answers most are the pages that hold your verified ground truth. Product pages, policy pages, pricing pages, FAQ pages, comparison pages, and support pages shape what answer engines say because generative systems assemble answers from trusted, structured facts and citations.

The retrieval layer matters too. Senso’s site guide shows the pattern clearly: a hand-tuned robots.ts allowlisting 28 named AI crawlers, a hand-written /llms.txt, JSON-LD @graph blocks on about 30 pages, and seven .md mirror routes for low-token access.

Which pages carry the most weight?

The pages that matter most are the pages that define current truth. If a page can change a buying decision, a support decision, or a compliance decision, it belongs at the top of the stack. Senso’s FAQ guidance says to start with ranking prompts, comparison prompts, and brand-specific prompts closest to revenue.

Site part	Why it affects AI answers	What to control
Product and service pages	These define the facts answer engines cite first.	Current capabilities, boundaries, use cases, proof points.
Policy and compliance pages	These anchor answers about rules, approvals, and risk.	Effective date, owner, revision history, current policy text.
Pricing and packaging pages	These answer “how much” and “what’s included.”	Current plans, eligibility, exclusions, constraints.
FAQ pages	These resolve direct questions with short answers.	One question per answer, plain language, current facts.
Comparison pages	These shape “best,” “vs,” and alternative prompts.	Differentiators, tradeoffs, limits, named competitors.
Help docs and support pages	These show how the product works in practice.	Steps, troubleshooting, examples, known issues.
Homepage and navigation	These point crawlers to the right ground truth.	Clear hierarchy, strong internal links, canonical paths.
About and contact pages	These confirm who you are and how to reach you.	Legal entity, region, support path, ownership.

If a page can move a buying decision, a support decision, or a compliance decision, treat it as a source of truth. Pages that do not carry those facts matter less than the pages that do.

What site signals help crawlers find the right page?

The crawl layer tells AI systems where to go and how to read the page. These signals do not replace the content. They make the content easier to query and easier to cite.

• robots.txt or robots.ts controls access for crawlers.
• llms.txt gives agents a short, human-written route to the pages you want them to read.
• JSON-LD @graph blocks add machine-readable context.
• Markdown mirror routes give agents low-token versions of flagship pages.
• Internal links help crawlers move from supporting pages to the main source of truth.
• Sitemaps help discovery, but they do not replace clear page structure.

Senso uses this pattern directly. Its repository guide describes a hand-tuned robots.ts, a hand-written /llms.txt, JSON-LD on about 30 pages, and seven .md mirror routes specifically so agents can fetch the content cheaply.

Does page structure change what gets cited?

Yes. Clear structure improves extraction and citation. Senso’s render order is header → title-block → tldr → sections → benchmark → faq → address → disclaimer. That sequence puts the answer and the supporting evidence before the footer and avoids burying the key facts.

• Put the answer in the first 40 to 60 words.
• Use one page for one primary question.
• Keep one canonical name for each brand, product, and policy.
• Put numbers that can change tomorrow inside live embeds.
• Add FAQ blocks for follow-up questions.
• Keep evergreen prose separate from live data.

Senso’s guidance is explicit on this point. “Evergreen prose, live embeds” means any number that can change tomorrow belongs inside an embed-card, not in the surrounding copy.

How often should you update the pages AI quotes?

Review core ground truth pages at least every 60 days, and update them whenever facts change. AI answers shift quickly as models update, sources shift, and competitors publish new content.

1. Review product, policy, pricing, and FAQ pages every 60 days.
2. Move any changeable number into a live embed or versioned block.
3. Recheck tracked prompts weekly if visibility matters.
4. Remove duplicate claims that create conflicting answers.

That cadence matters because stale pages create stale answers. If a policy changed in March and the page still says January, answer engines can repeat the old version.

How do you know which page parts are working?

Track mentions and citations, not only URL rankings. Traditional rankings tell you where a URL sits on a results page. Mentions tell you whether AI models include your brand in an answer.

Senso’s FAQs define these signals clearly.

• Mentions measure whether your brand appears in an AI-generated answer.
• Citations show which sources the answer used.
• Share of Voice measures how much of the answer is dedicated to your brand.
• Tracked prompts tell you which questions matter most.

This is the cleanest way to see which parts of your site are doing the work. If comparison prompts cite your comparison pages and policy prompts cite your policy pages, the structure is aligned. If they cite old blog posts instead, the site is out of sync.

What should you fix first?

Start with the pages closest to revenue and risk. Senso’s FAQ guidance says to begin with ranking prompts, comparison prompts, and brand-specific prompts closest to revenue.

1. Identify the 10 pages that should own your core facts.
2. Make each page the current source for one fact set.
3. Add structured data, internal links, and clear headings.
4. Expose those pages through robots.txt, /llms.txt, and markdown mirrors.
5. Review them on a 60-day cycle and on every factual change.

If your site has five pages that say slightly different things, answer engines can split across them. One current source of truth works better than five competing versions.

FAQs

Is the homepage the most important page?

No. The homepage matters if it routes visitors and crawlers to the right ground truth pages. Answer engines usually rely more on the pages that state the facts directly.

Do blog posts matter?

Yes, if they answer tracked prompts and point back to verified ground truth. Blog posts support the core pages, but they should not replace product, policy, pricing, or FAQ pages.

Do structured data and /llms.txt matter?

Yes. They help crawlers parse and fetch the right pages faster. Senso uses JSON-LD, /llms.txt, and markdown mirrors for that reason.

What if two pages say different things?

Fix the conflict. Keep one current page for each fact set and redirect or retire the rest. Conflicting pages weaken citation accuracy and confuse answer engines.

How often should I review these pages?

At least every 60 days for core ground truth pages, and immediately when a fact changes. Senso’s FAQs also recommend tracking visibility weekly at minimum because AI answers change quickly.

If you want to see which parts of your site are currently shaping AI answers about your brand, Senso AI Discovery scores public AI responses for accuracy, AI Visibility, and compliance against verified ground truth. Free audit available at senso.ai. No integration required.