| name | aso-optimizer |
| description | Optimize website content for AI search visibility and durable Google Search readiness. Use when creating, editing, reviewing, auditing, or publishing blog posts, landing pages, documentation, product pages, ecommerce pages, local-business pages, or other public web content where generative AI search features, Google AI Overviews/AI Mode, SEO, crawlability, snippets, helpful content, or content-quality strategy matter. |
ASO Optimizer
Overview
Use this skill as a practical AI Search Optimization workflow for public web content. Treat Google's guidance as a correction away from "AEO/GEO hacks" and toward strong SEO fundamentals: unique value, technical discoverability, satisfying human readers, and pages that browser agents can inspect.
Read references/google-search-ai-optimization.md when you need source-backed rationale, a fuller checklist, or citations from the Google guide.
Core Rule
Optimize for the search experience, not for a separate AI-only system. Preserve the user's brand voice and goals, but make every public page easier for people, Google Search systems, and browser-based agents to understand, cite, and act on.
Do not turn content into generic AI-search bait. If a suggested optimization does not make the page more useful, trustworthy, specific, crawlable, or satisfying for a real visitor, do not make it.
Workflow
-
Identify the page type and search intent.
- Classify the asset: blog post, landing page, product page, ecommerce category, documentation, local-business page, comparison, review, or support content.
- Name the audience, the problem they are trying to solve, and what a satisfying visit must deliver.
- Note whether the page needs freshness, first-hand experience, product/service details, local details, media, structured data, or conversion paths.
-
Run the non-commodity content test.
- Add or strengthen original information: first-hand experience, expert judgment, data, screenshots, examples, benchmarks, decisions, tradeoffs, failure modes, or direct product/service knowledge.
- Remove thin commodity sections that repeat common internet advice without adding a viewpoint.
- Prefer a smaller page with specific evidence over a larger page padded with generic FAQs or keyword variations.
- Make headings useful for human scanning; do not create mechanical "AI chunk" headings.
-
Improve usefulness and trust.
- Answer the primary intent early, then support it with evidence and detail.
- Make claims attributable: show who knows the topic, how they know it, and where the facts or product details come from.
- Add concrete examples and visuals where they help understanding.
- Clarify scope, limitations, pricing, availability, prerequisites, and next steps when relevant.
- Keep copy natural; rewrite hype, vague superlatives, and filler into specific claims.
-
Check technical discoverability.
- Ensure the important content is indexable, crawlable, and not blocked by
robots.txt, noindex, auth walls, client-only rendering failures, or broken canonicalization.
- Confirm the page is eligible for snippets when visibility in Google generative AI features matters.
- Use semantic HTML where practical for readability, accessibility, and agent/browser interpretation.
- For JavaScript-heavy pages, verify the content renders in a way Google can process and users can access.
- Reduce duplicate or near-duplicate URLs that waste crawl attention or confuse canonical intent.
-
Match page-specific SEO support.
- For articles and docs, check title, meta description, canonical URL, heading hierarchy, date semantics where relevant, internal links, author/source signals, images, and article-appropriate structured data.
- For landing pages, make the offer, audience, proof, differentiators, conversion action, and objections clear without hiding the main content behind effects or interstitials.
- For product/ecommerce pages, ensure product data, availability, variants, return/shipping details, reviews where genuine, product images/video, Merchant Center/feed alignment, and relevant structured data are coherent.
- For local-business pages, align name/address/phone, service area, hours, reviews/testimonials where genuine, business details, Google Business Profile assumptions, and local intent.
- For comparison/review pages, require real evaluation criteria, hands-on evidence where possible, tradeoffs, and an honest recommendation boundary.
-
Prepare for agentic experiences when relevant.
- Make primary actions visible in the DOM and accessibility tree.
- Use clear labels for forms, buttons, prices, variants, availability, contact options, and policies.
- Avoid UI that depends only on hover, visual-only cues, screenshots of text, or inaccessible custom controls.
- Verify that a browser agent could inspect the page, understand the main task, and complete the intended action without hidden context.
-
Produce an actionable output.
- For edits, directly improve the content or code when possible.
- For audits, return prioritized findings grouped as
Content value, Technical discovery, Page experience, Media/structured data, and Avoided AI-search hacks.
- Include exact file/page references when working in a repo.
- Distinguish must-fix issues from optional enhancements.
Anti-Patterns
Avoid these unless the user explicitly asks and there is a separate non-search reason:
- Creating
llms.txt, Markdown mirrors, or special AI-only files solely to appear in Google generative AI search.
- Breaking pages into artificial tiny chunks for AI systems.
- Rewriting natural copy into keyword permutations or exhaustive long-tail query variants.
- Creating many near-duplicate pages to capture fan-out queries.
- Pursuing fake mentions, synthetic citations, or low-quality backlink/PR schemes.
- Treating structured data as a magic AI-search ranking lever. Use it when it accurately represents visible page content and supports normal rich-result eligibility.
- Publishing AI-generated text without human value, review, experience, or compliance with Search Essentials and spam policies.
Verification
Before finalizing a content or page change, check:
- The page has a clear audience, primary intent, and satisfying answer.
- At least one concrete element would be hard for a generic model or competitor summary to recreate.
- The title, headings, intro, internal links, and calls to action match the real page purpose.
- Important content is visible to users, crawlable, indexable, and snippet-eligible when applicable.
- Images or videos are relevant, accessible, and supported by descriptive surrounding context.
- Structured data, if present, matches visible content and the page type.
- The recommendations did not add AI-search gimmicks that make the page worse for humans.
For high-stakes or current SEO claims, verify against current official Google Search documentation before citing or changing policy-sensitive guidance.