Most sites that hire AI SEO agencies for the first time are paying professional rates to fix things they could have caught and fixed themselves in a weekend.
I’m not saying that to talk you out of hiring — I run AI SEO services myself. I’m saying it because the sites that get the most value from professional AI SEO work are the ones that already understand their baseline. They know what’s wrong. They know what the agency is fixing. They can evaluate whether it’s actually getting done.
This checklist covers 27 signals across five categories. Work through it before you spend a dollar on AI SEO services. What you find will either tell you that the DIY path is faster than you thought, or give you a precise brief for a professional engagement.
How to use this checklist
Run through each item and mark it as:
- ✅ Done — signal is in place and working
- ⚠️ Partial — in place on some pages but not all
- ❌ Missing — not implemented
At the end, count your ❌ and ⚠️ marks. If you have more than 8 in the schema and entity categories (items 1–14), fix those before anything else — they’re the fastest-moving signals in AI citation patterns.
Category 1 — Schema (10 signals)
These are binary: either the schema is present and validating, or it isn’t. Check each page type, not just your homepage.
1. Article schema on every blog post
Validate in Google’s Rich Results Test. Must include: headline, datePublished, dateModified, author (linked to your Person node via @id), publisher, and mainEntityOfPage.
The most common failure: dateModified is missing or matches datePublished and never updates. AI models weight recency — stale dateModified is a citation liability.
2. FAQPage schema on cluster articles Every cluster spoke article — not just your pillar — should have FAQPage schema. Questions must match actual user search queries (pull from People Also Ask for your target keyword, not generic questions you invented).
3. Person schema on your About page
Required fields: name, url, sameAs (array of external profile URLs), knowsAbout, jobTitle. The sameAs property is the entity recognition anchor — it connects your on-site identity to your external profiles in AI model knowledge graphs.
4. BreadcrumbList on every page Not just blog posts — every page on the site. Simple implementation. High impact on AI navigation signals.
5. HowTo schema on process articles
Any article with a step-by-step structure should have HowTo schema. Each step needs a name and text minimum. This is underimplemented on 90% of sites I audit.
6. Speakable schema on definition blocks Mark your direct-answer blocks and FAQ answers with Speakable schema. Specifically designed for Google AI Overviews and voice retrieval. Still underused — early mover advantage applies.
7. WebSite schema on homepage
With name, url, description, and publisher linking to your Person node. This creates the root entity connection that ties every page’s schema to the site-level identity.
8. ImageObject schema on featured images
Each featured image should have @id, url, and contentUrl. Links to the Article’s image property. Improves image attribution in AI-generated answers.
9. Schema validated in Rich Results Test Not just implemented — validated. Errors in schema are worse than missing schema because they signal inaccurate structured data. Every page type should have a screenshot-on-file validation.
10. No conflicting schema types FAQPage schema on a page with no visible FAQs, LocalBusiness schema on a non-local page — these create trust signals in the wrong direction. Audit every page for schema-content alignment.
Category 2 — Entity signals (4 signals)
11. Author name consistent across all platforms Your name on your site must exactly match your LinkedIn profile, your YouTube channel, your Amazon Author page, your Twitter/X handle display name, and any other platform where you appear. AI models build entity confidence from consistent cross-platform signals. Inconsistencies create ambiguity.
12. External profiles verified and linked
LinkedIn, YouTube, and any authoritative external profiles should be linked from your About page AND listed in your Person schema sameAs property. The link and the schema entry must match.
13. About page is information-dense Not a marketing page — an information page. It should answer: who are you, what’s your expertise, what sites or work can verify this, and how can someone contact you. Thin, promotional About pages perform worse on entity recognition than informational ones.
14. Google Knowledge Panel exists (if applicable) If you have one, verify the information is accurate and your website is linked. If you don’t have one yet, building it isn’t the priority — the schema and external links above will eventually trigger it as a byproduct.
Category 3 — E-E-A-T signals (7 signals)
15. Named author byline above H1 Not in the footer. Not at the bottom of the article. Above the H1 — visible before the reader encounters the headline. This is the signal that tells quality raters (human and algorithmic) who is responsible for this content.
16. Author page linked from byline The byline should link to an About or Author page that itself links to external verification. The chain: article → author page → LinkedIn/YouTube/published work. Each link in the chain builds attribution confidence.
17. First-person language on content pages Phrases like “in my experience,” “I’ve tested this,” “when I audited a client site last month” — language that demonstrates direct experience. AI models are trained to differentiate editorial content from generated filler. First-person specificity is the clearest signal.
18. Insider Tip callouts on every content page A formatted aside that shares a non-obvious insight from direct experience. Not a generic tip — a specific one, tied to a specific observation. One per article minimum.
19. Primary source citations Links to original research, government data, or peer-reviewed studies — not other blogs. Every factual claim should trace to a primary source. The citation is itself an expertise signal.
20. Original imagery where applicable Site-specific photos, screenshots, or diagrams — not stock photography. For informational content, annotated screenshots and diagrams count. For service content, photos of actual work.
21. Credentials stated inline Not just in the bio — in the body copy where the claim is made. “As someone who has built 30+ authority sites over 25 years, I’ve found…” anchors the credential to the specific claim it supports.
Category 4 — Content architecture (4 signals)
22. At least one complete pillar + cluster structure A pillar page (3,000–8,000 words, comprehensive) plus at least three cluster spokes (1,500–3,500 words each), all internally linked. Generative Engine Optimization: The Complete Guide is an example of a pillar page with its cluster properly built out.
23. All spokes link up to the pillar 3–4 times Not once. Three to four times, with anchor text that includes the primary keyword. This is the internal linking pattern that builds topical authority at the hub level.
24. Pillar links down to every spoke Every cluster spoke article should be linked from the pillar page. The pillar is the hub — it should point outward to every spoke in the cluster.
25. No orphaned articles Every article should receive at least one internal link from a related page. Orphaned content — articles with no internal links pointing to them — is invisible to AI retrieval systems regardless of how good the content is.
Category 5 — Content quality (2 signals)
26. Direct-answer block in first 200 words Every article and pillar page should answer its primary query directly within the first 200 words. Not a teaser. Not a promise to answer later. The actual answer, in 40–80 words, in declarative prose.
27. FAQ section with PAA-sourced questions At least four questions at the bottom of cluster articles, sourced from People Also Ask data for the target keyword. Answers should be 40–100 words — direct, complete, and attributable. These feed the FAQPage schema and the AI retrieval pipeline simultaneously.
Scoring and next steps
Count your marks:
| Score | Interpretation |
|---|---|
| 23–27 ✅ | Strong foundation — focus on topical depth and content volume |
| 17–22 ✅ | Solid — fix remaining gaps in schema and entity first |
| 10–16 ✅ | Moderate — significant opportunity; start with schema (items 1–10) |
| Under 10 ✅ | Starting point — prioritise the schema layer before anything else |
If you scored under 17: most of your gaps are in schema and entity signals — fixable yourself with a weekend of focused work.
If you scored 17–22: your gaps are likely in E-E-A-T and content architecture — these take longer but compound over time.
If you want an expert second opinion: the AI Search Audit ($49) delivers an outside diagnostic of all 27 signals plus citation testing across all four engines, delivered in 5 business days.
Frequently asked questions
What is a GEO readiness audit?
A GEO readiness audit checks whether your site has the schema, entity, E-E-A-T, and content architecture signals that AI search engines use to decide who to cite. It's the diagnostic step before any AI SEO investment — either DIY fixes or hiring a professional.
How long does a GEO audit take?
A thorough self-audit using this checklist takes 2–4 hours for a site with under 50 pages. A professional AI search audit with citation testing across all four engines typically takes 3–5 business days.
What are the most common GEO failures I'll find in an audit?
The most common gaps are: missing FAQPage schema on cluster articles, no Person schema on the About page, author byline below the H1 instead of above it, missing dateModified in Article schema, and no direct-answer paragraph in the first 200 words of cluster articles.
Do I need to fix all 27 signals before AI citations will improve?
No. Schema and entity signals (items 1–14 on this checklist) have the fastest impact — typically 2–6 weeks to register in citation patterns. E-E-A-T and content signals take longer but compound over time. Prioritise the schema fixes first, then work through the rest.
What tools do I need to run a GEO audit?
Google's Rich Results Test for schema validation. Google Search Console for crawl coverage and indexed pages. A spreadsheet for citation rate tracking. Optionally: Screaming Frog for crawl-level schema inventory. You don't need premium tools to complete this checklist.
Related: AI SEO Services: What They Cover and What They Cost · DIY vs Hiring a GEO Agency · Inside an AI Search Audit · Get the AI Search Audit ($49)