# AI readiness score

> How GetIntel's technical AI readiness score is calculated for each engine, what the 80 / 15 / 5 weighting means, and what raises it.

Source: https://getintel.ai/docs/academy/metrics/ai-readiness-score/

The **AI readiness score** (0 to 100) measures whether an AI engine can reach, read and use your key pages. GetIntel scores each of ChatGPT, Google AI, Perplexity, Claude and Copilot, and the overall score is the average of the five.

## How it's calculated

For each engine:

| Part | Weight | What it measures |
| --- | --- | --- |
| **Usable key pages** | 80 | The share of your key pages this engine can reach and read, weighted by page importance. |
| **Discover & Fresh** | 15 | Sitemap coverage, honest last-modified dates, internal links, freshness and IndexNow. |
| **Page health** | 5 | Unique titles, no broken internal links, no near-duplicates, Core Web Vitals. |

**Worked example.** Perplexity can use 9 of your 10 key pages (90%), discovery checks pass at 80% and health at 60%: 80 × 0.9 + 15 × 0.8 + 5 × 0.6 = **87**.

## Why engines score differently

Checks apply only where they matter. JavaScript-dependent content, for example, only hurts engines that don't run JavaScript: ChatGPT, Perplexity and Claude. A site that renders its content in the browser can score well on Google AI and poorly on ChatGPT.

## Where to find it

**Technical → Overview**: the overall score and its history, a score per engine, and every check. Fixes appear under **Actions → Technical**.

## What raises it

Start with **Reach**: a bot blocked by robots.txt or a firewall makes every page unusable for that engine, which costs most of the 80 points. See [AI crawlers show nothing or a bot is blocked](/docs/academy/troubleshooting/bot-blocked/).