- A page-by-page scoring template that grades your content on 10 AI search signals
- Score each page from 0-100 so you know exactly which content to fix first
- Includes the scoring rubric, a prioritization matrix, and a sample filled-out template
- Works for any CMS and any industry, just plug in your URLs
Most Content Audits Miss What AI Search Actually Cares About
Traditional content audits check things like word count, keyword density, and backlink count. That’s fine for Google’s classic algorithm. But AI search engines evaluate content differently.
ChatGPT, Perplexity, and Gemini care about things like entity coverage, factual density, source credibility, and structural clarity. A page can rank #1 on Google and still never get cited by an AI engine. I’ve seen it happen dozens of times.
This template scores each page on the 10 signals that actually matter for AI search visibility. You’ll end up with a ranked list of pages, sorted by improvement potential. Fix the worst ones first.
That 527% growth figure comes from BrightEdge’s 2025 year-in-review report. AI referral traffic is growing faster than any other channel. Your content either works for AI search or it doesn’t.
The 10 AI Search Signals
I’ve tested these signals across hundreds of pages that do and don’t get cited by AI engines. These are the patterns that separate cited content from ignored content.
Signal-by-Signal Scoring Rubric
Signal 1: Direct Answer Density (15 points)
AI engines love content that directly answers questions. Not content that dances around the topic for 500 words before getting to the point.
| Score | Criteria |
|---|---|
| 15 pts | Page leads with clear answers. First 200 words contain the core answer. Uses definition-style sentences (“X is…”) that AI can extract. |
| 10 pts | Answers are present but buried. Reader has to scroll past introductions, disclaimers, or filler to find them. |
| 5 pts | Partial answers. Page covers the topic but never gives a clear, extractable answer. |
| 0 pts | No direct answers. Content is all opinion or narrative with nothing an AI engine could quote. |
How to check: Read the first 200 words of your page. Could someone pull a standalone answer from them? If yes, you’re in good shape.
Signal 2: Entity Coverage (15 points)
Does the page mention the key entities (people, companies, technologies, concepts) that AI engines associate with this topic?
| Score | Criteria |
|---|---|
| 15 pts | Covers 8+ relevant entities with context. Mentions specific companies, tools, people, and concepts by name. |
| 10 pts | Covers 4-7 relevant entities. Some key players or concepts are missing. |
| 5 pts | Covers 1-3 entities. Most references are generic (“many companies” instead of naming them). |
| 0 pts | No entity references. Content is entirely generic with no specific names or examples. |
For background on why entity coverage matters, read our entity architecture guide. AI engines build understanding through entity relationships, not keywords.
Signal 3: Factual Specificity (12 points)
Does the page contain specific, verifiable facts? Numbers, dates, statistics, research citations?
| Score | Criteria |
|---|---|
| 12 pts | 5+ specific stats, dates, or data points with sources. Content reads like a research summary. |
| 8 pts | 3-4 specific facts with some sourcing. Mix of specific and general claims. |
| 4 pts | 1-2 specific facts. Most claims are vague (“many studies show…” without citing them). |
| 0 pts | No specific facts. All claims are general or unsubstantiated. |
Animalz’s research on content quality found that factual density is one of the strongest signals for content that gets referenced and shared. AI engines follow the same pattern.
Signal 4: Source Citations (10 points)
Does the page cite its sources? AI engines treat well-sourced content as more trustworthy.
| Score | Criteria |
|---|---|
| 10 pts | 5+ external citations to authoritative sources. Links to research, official documentation, or primary data. |
| 7 pts | 3-4 external citations. Mix of strong and weak sources. |
| 3 pts | 1-2 citations, or citations to non-authoritative sources. |
| 0 pts | No external citations at all. |
Signal 5: Structural Clarity (10 points)
Is the content structured in a way that AI engines can easily parse?
| Score | Criteria |
|---|---|
| 10 pts | Clear H2/H3 hierarchy. Short paragraphs (2-3 sentences). Lists and tables for data. Logical flow from general to specific. |
| 7 pts | Has headers and some structure, but sections are long and headers are vague. |
| 3 pts | Minimal structure. Wall-of-text with few headers. |
| 0 pts | No structure. Single block of text with no headers, lists, or visual breaks. |
Signal 6: Schema Markup (10 points)
| Score | Criteria |
|---|---|
| 10 pts | Has Article schema, FAQ schema, author markup, and relevant entity schema (Product, HowTo, etc.). |
| 7 pts | Has Article schema and one additional type. |
| 3 pts | Has basic Article or WebPage schema only. |
| 0 pts | No schema markup. |
Check your markup using Google’s Rich Results Test or our schema markup generator.
Signal 7: Freshness Signals (8 points)
| Score | Criteria |
|---|---|
| 8 pts | Updated within last 90 days. References current-year data. Has visible “last updated” date. |
| 5 pts | Updated within 6 months. Some references are current. |
| 2 pts | Updated within 12 months. Data feels dated. |
| 0 pts | No update date visible, or content is clearly outdated. |
Signal 8: Author Authority (8 points)
| Score | Criteria |
|---|---|
| 8 pts | Named author with bio, credentials, linked LinkedIn profile, and published work in the topic area. |
| 5 pts | Named author with basic bio. Some credentials mentioned. |
| 2 pts | Named author but no bio or credentials. |
| 0 pts | No author attribution. Content appears anonymous. |
Signal 9: FAQ/Question Coverage (7 points)
| Score | Criteria |
|---|---|
| 7 pts | Includes FAQ section with 3+ questions matching real user queries. Has FAQ schema. |
| 4 pts | Addresses common questions within the content but no dedicated FAQ section. |
| 2 pts | Touches on 1-2 questions tangentially. |
| 0 pts | Doesn’t address any questions users commonly ask about the topic. |
Signal 10: Technical Accessibility (5 points)
| Score | Criteria |
|---|---|
| 5 pts | Fast loading, mobile-friendly, no render-blocking issues. AI crawlers can access all content without JavaScript. |
| 3 pts | Minor technical issues. Content is mostly accessible. |
| 1 pt | Significant issues. Content requires JavaScript to render or is behind login walls. |
| 0 pts | Content blocked from crawlers via robots.txt or requires authentication. |
For technical accessibility checks specific to AI crawlers, see our technical AEO checklist.
The Scoring Template: How to Use It
Set up a spreadsheet with these columns:
| Column | Description |
|---|---|
| URL | The page URL |
| Target Query | The main question this page should answer |
| S1-S10 | Individual signal scores |
| Total | Sum of all 10 signals (max 100) |
| Priority | High/Medium/Low based on total score + traffic potential |
| Next Action | The single most impactful improvement for this page |
Sample Filled-Out Template
Here’s what a real audit looks like for a SaaS company’s blog:
| URL | S1 | S2 | S3 | S4 | S5 | S6 | S7 | S8 | S9 | S10 | Total |
|---|---|---|---|---|---|---|---|---|---|---|---|
| /what-is-crm | 15 | 10 | 8 | 7 | 10 | 7 | 5 | 8 | 7 | 5 | 82 |
| /crm-comparison | 10 | 15 | 12 | 10 | 7 | 3 | 2 | 5 | 4 | 5 | 73 |
| /sales-tips | 5 | 5 | 4 | 0 | 3 | 0 | 0 | 2 | 0 | 3 | 22 |
| /product-update | 0 | 5 | 8 | 3 | 7 | 3 | 8 | 5 | 0 | 5 | 44 |
See the pattern? The /what-is-crm page scores high because it directly answers a common question with specific details. The /sales-tips page scores low because it’s generic opinion content with no sources or structure.
Prioritization: Fix These Pages First
Don’t just fix the lowest-scoring pages first. Prioritize based on a combination of score and opportunity.
High Traffic + Low Score = Fix Immediately
Pages already getting organic traffic but scoring below 50 on AI signals. These have proven demand and just need AI-specific improvements.
High Intent + Medium Score = Quick Wins
Pages targeting high-intent queries (comparisons, “how to,” “best X”) scoring 50-70. Small improvements can push these into citation territory.
Low Traffic + Low Score = Rewrite or Remove
Pages with no traffic and low AI scores. These need a complete rewrite or should be consolidated with stronger pages.
High Score + Low Traffic = Promote
Pages that score well on AI signals but aren’t getting traffic. The content is good; it just needs distribution and authority building.
Common Patterns I See in Audits
After running this template across many sites, here are the recurring issues:
Signal 1 (Direct Answers) is almost always the weakest. Most content buries the answer. Blog posts start with 300 words of setup before getting to the point. AI engines need the answer up front.
Signal 6 (Schema) is often zero. Tons of sites have no schema markup at all, or just basic Article schema. Adding FAQ schema and proper author markup can bump a page by 10+ points.
Signal 4 (Source Citations) varies wildly. B2B content tends to cite sources. B2C content often doesn’t. AI engines notice the difference.
Fun fact I’ve noticed: pages that score above 70 on this template get cited by Perplexity about 3x more often than pages scoring below 40. That’s not a formal study (I wish I had the sample size for that), just a consistent pattern from the audits I’ve run.
Automating the Scoring
Scoring pages manually takes about 5 minutes each. For a site with 100 pages, that’s over 8 hours. Here’s how to speed things up:
- Use screaming frog for technical signals. Signals 5, 6, 7, and 10 can be partially automated with a Screaming Frog crawl. Export the data and fill those columns automatically.
- Use your CMS for freshness. Pull “last modified” dates from your CMS to score Signal 7.
- Batch-check schema. Google’s Rich Results API lets you check schema programmatically. Or use our schema markup guide to audit manually.
- Manual review for signals 1-3. These require actually reading the content. No shortcut here. But you can prioritize by starting with your highest-traffic pages.
Connecting This to Your AI Search Strategy
This template is one piece of a bigger puzzle. Here’s how it fits with other tools:
- Use the AI search readiness scorecard for a site-wide assessment
- Use the AI search improvement checklist to fix the specific issues this template identifies
- Use the free citation checker to verify whether improvements are leading to actual citations
The flow is: Score your content → Fix the gaps → Check for citations → Repeat.
For the full picture on AI search improvement, start with our pillar guide.
Want a Professional Content Audit for AI Search?
We score your top pages, identify quick wins, and build a prioritized action plan. Get your free AI visibility audit to start.
Frequently Asked Questions
About 5 minutes per page for manual scoring. A 50-page site takes roughly 4 hours. You can speed this up by automating the technical signals (schema, freshness, structure) and focusing manual review on content quality signals.
Aim for 70+ on pages targeting queries that AI engines commonly answer. Pages scoring 70+ get cited significantly more often than those below 40. Not every page needs a perfect score though. Focus on your highest-value content first.
Yes. The scoring criteria are CMS-agnostic. Whether you use WordPress, Webflow, Shopify, or a custom CMS, the 10 signals apply equally. The only difference is how you implement fixes (schema plugins vs manual code, etc.).
Re-score your top 20 pages every quarter. Re-score any page you’ve updated within 2 weeks of publishing changes to see if your improvements moved the needle. Add new pages to the template as you publish them.