SEO and AI Search Resources for Audits, Tests and Decisions
Use this library when you need to diagnose a crawl problem, verify an AI-search claim, run a bounded experiment or choose the next publishing action. Each resource is organized around a job—not a vague promise—and the featured worksheets preserve evidence, limits and rollback decisions.
Start here: choose the problem you have now. Do not download everything or apply every checklist to every page.
Choose a resource by the problem you need to solve
| Your problem | Start with | Useful output |
|---|---|---|
| A page is not being crawled or indexed | Technical SEO pre-publish checklist | A verified status, canonical, directive and discovery record |
| Traffic dropped | Search traffic drop checklist | A dated diagnosis with confounders and next checks |
| An AI crawler appears in logs | Crawler identity verification guide | Claimed, verified and served fields kept separate |
| You want to test an AEO change | AEO experiment validator | A bounded hypothesis, guardrails and decision rule |
| You need better search snippets | SERP snippet preview | A readable title and description before release |
Crawler and indexing controls
Start with the AI crawler field guide to distinguish search, training, user-initiated and control-token roles. For OpenAI specifically, use ChatGPT-User and robots.txt: what you can—and cannot—block before changing a WAF or robots policy. The guide includes a downloadable decision ledger and four-layer test matrix.
When a request claims a known agent name, use the user-agent and IP verification workflow. A user-agent label is a claim, not identity proof. For a traditional indexing release, the technical pre-publish checklist covers status, canonical, robots, sitemap, structured data and public delivery.
Content evidence and citation quality
Use How to make content easier to cite in AI search to build passages with a clear subject, scope, evidence and limitation. Then read Citation Wars: can GEO optimization make content worse? to understand why a citation-seeking edit can fail even when the surface signal improves.
The citation-wars analysis adds a worked before-and-after repair, separates paper results from repository settings, and provides a downloadable quality gate. For broader claim review, use AI search ranking factors: what platforms document—and what they don’t. It classifies a statement as documented, observed or unknown before it becomes advice.
Experiments and measurement
Use the small SEO experiment method when you can isolate one material treatment and preserve a comparison. The AEO experiment validator turns that plan into a preflight check. Neither tool can manufacture causality; both are designed to expose missing denominators, shifting prompts and unrecorded confounders before you publish a conclusion.
For ChatGPT discovery and referrals, the ChatGPT SEO operating guide keeps crawl access, citations, clicks and business outcomes in separate ledgers. For reporting across platforms, AI visibility reports explains why mentions, citations, referrals and revenue should not be compressed into one proprietary score.
Page repair and publishing workflows
If an existing page needs stronger information gain, follow the page-level AI-search rewrite workflow. It asks for an answer contract, claim-source repair, a real example, a limitation and a rollback baseline. For repeatable drafting, auditable AI SEO prompts separates research, synthesis, writing and review instead of letting one prompt invent the whole chain.
Before choosing software, browse the tools library. Before releasing a repeatable process, browse the checklists library. Completed analyses and methods live in Research; they should state the sample, date and limits rather than presenting an isolated number as a benchmark.
How to use the library without creating busywork
- Name the decision. Write what you will allow, fix, test, publish or stop.
- Choose one primary resource. Add a second only when it supplies a different job, such as verification after diagnosis.
- Replace every sample row. Files marked
EXAMPLE-REMOVEare templates, not evidence about your property. - Attach dates and sources. Crawler names, platform controls and observed behavior change.
- Preserve the baseline. Record the current state and rollback route before a production edit.
- Report uncertainty. “Not observed” and “not measurable” are valid outcomes.
SearchEngineAnswer updates this hub when a resource materially improves a task, corrects an outdated control or adds a reusable artifact. We do not list a page merely because it targets a new keyword. See the editorial policy for sourcing and corrections, or Start Here for the publication’s core learning path.