How the AI readiness score works

RocketCite scores every page on your site across 43 criteria that directly affect whether AI systems can find, understand, and cite your content. The result is a single percentage — your AI Readiness Score — that tells you where you stand and what to improve.

This page explains the scoring model: what gets measured, how weights work, what the score tiers mean, and how the system adapts to different types of websites.

The Five Categories

Every criterion belongs to one of five categories. Each category targets a different aspect of AI visibility.

Content Quality (56.9% of default weight)

The largest category, and for good reason: AI systems are text-processing machines at their core. Content Quality measures how well your writing is structured, formatted, and composed for AI consumption.

This includes heading hierarchy, answer-first paragraph structure, readability level, factual density, tables, lists, FAQ sections, image alt text, freshness signals, internal linking, semantic completeness, and more. If you write clearly, structure logically, and include supporting data, this category takes care of itself.

18 criteria, 68.0 default points

Structured Data (17.6%)

Schema markup helps AI systems understand what your content is — not just what it says. Article schema tells systems “this is a blog post by this author, published on this date.” Product schema says “this is a product with this price and availability.” FAQ schema maps your questions to their answers in a machine-readable format.

Without structured data, AI systems have to guess your content type from context alone. With it, they can index and cite you precisely.

7 criteria, 21.0 default points

Technical Foundation (16.7%)

Before AI crawlers can evaluate your content, they need to reach it. Technical Foundation checks the infrastructure: Can bots access your pages? Is your content server-rendered or hidden behind JavaScript? Do you have a sitemap? Are your caching headers set correctly? Is there a valid llms.txt file?

Most WordPress sites handle the basics well out of the box. This category catches the gaps that silently block AI visibility.

9 criteria, 20.0 default points

E-E-A-T / Entity Signals (15.9%)

Experience, Expertise, Authoritativeness, and Trustworthiness. These are the signals that tell AI systems who is behind this content and why should it be trusted. Author bylines, Person schema, organization identity, sameAs links to Wikipedia or Wikidata, and source attribution all contribute here.

For sites where credibility matters — and it almost always does — this category separates trusted sources from anonymous content.

6 criteria, 19.0 default points

AI System Configuration (3.3%)

Site-wide controls for how AI crawlers interact with your content. This covers bot separation (allowing search/retrieval bots while blocking training crawlers), Google-Extended configuration, and Chinese bot management (Bytespider).

Small in weight, but critical in function: getting this wrong can block AI visibility entirely or expose your content to unwanted training use.

3 criteria, 4.0 default points

Scoring Tiers

Your AI Readiness Score is a percentage from 0 to 100. Here is what each tier means in practice:

A score of 60% or above means your content is in a healthy position for AI visibility. Most sites start between 30-50% before any optimization, so there is nearly always room to improve.

How Weights Work

Each criterion has a weight expressed as its maximum score (maxscore). A criterion with a maxscore of 7.0 can contribute up to 7 points to your raw total. One with a max_score of 1.5 contributes at most 1.5 points.

The final percentage is calculated as:

Score % = (sum of earned points / sum of max possible points) * 100

With default weights, the total max possible is 119.5 raw points. If your content earns 83.7 raw points, your score is 83.7 / 119.5 * 100 = 70%.

Weights reflect relative importance. Answer-First Paragraph has a maxscore of 10.0 because AI systems heavily prioritize content that answers queries directly and immediately. Caching Headers has a maxscore of 0.25 because while useful, caching alone has minimal impact on AI citation likelihood.

How individual criteria are scored

Each criterion has an evaluator that computes a raw score between 0 and its internal maximum. The score is then scaled to the effective weight:

scaled_score = (raw_score / internal_max) * effective_weight

If you earn 5 out of 10 internally on a criterion with an effective weight of 7.0, your scaled score is (5/10) * 7.0 = 3.5. This means changing a weight changes both the ceiling and the proportional contribution of that criterion.

The Weight Resolution Chain

RocketCite resolves weights through a layered system. Each layer can override the previous one:

1. Default configuration — The built-in weights shipped with the plugin (data/scoring-weights.php). These are the starting point for every site.
2. Business-type preset (Pro) — Pre-configured weight adjustments for specific business models: E-Commerce, Publisher/News, SaaS/Developer Docs, Local Business, and Portfolio/Creative. Each preset adjusts a handful of weights to match the priorities of that business type.
3. User overrides (Pro) — Manual weight adjustments made through the visual Scoring Weights Editor in the plugin settings. These are stored as differences from the defaults, so only your changes are saved.
4. Developer filters — WordPress filter hooks (rocketcite_scoring_weights and per-criterion filters) that allow theme or plugin developers to adjust weights programmatically. Filters have the highest priority — code always wins over UI settings and presets.
5. Validation — Every weight configuration passes through the WeightValidator before use. Invalid configurations (negative weights, zero totals, non-numeric values) are rejected and the system falls back to defaults.

The result is cached per request, so the resolution chain runs at most once per page load.

Business-Type Presets

Different types of websites have different priorities. An e-commerce store cares more about product schema and page speed than heading structure. A news publisher needs strong E-E-A-T signals and content quality. A portfolio site emphasizes visual content and personal branding.

RocketCite Pro includes six presets that adjust weights to match these priorities:

Presets are a starting point. You can select a preset and then make additional adjustments on top of it.

Why Transparency Matters

Most SEO tools treat their scoring algorithms as black boxes. You get a number, maybe a color, and a list of suggestions — but no insight into how the score is actually calculated or why certain factors matter more than others.

RocketCite takes a different approach. The entire scoring model is open:

• Every criterion, weight, and category is documented — you can see exactly what is measured and how much it counts.
• Weights are customizable — developers can adjust weights via WordPress filters (free), and Pro users get a visual editor and business-type presets.
• The resolution chain is deterministic — you can trace exactly how your effective weights were calculated: defaults, then preset, then overrides, then filters, then validation.
• The scoring model is versioned — when weights change in a new plugin version, the changelog calls it out so you can review the impact on your scores.

This transparency is not just a feature. When you understand how the score works, you make better decisions about what to improve. You stop chasing arbitrary numbers and start making changes that genuinely improve your AI visibility.