Discovery Scoring Engine

The Discovery Scoring Engine scans your pages for AI readability, identifies exactly what is blocking each one, routes specific fixes to the right team member, generates the manifest files AI crawlers need, and tracks your progress across scans over time.

Overview

Most brands know their AI visibility score but do not know which specific pages are holding them back — or what exactly needs to change on each one. The Discovery Scoring Engine bridges that gap. It works at the page level, not the brand level, and produces outputs that are immediately actionable for both content teams and developers.

Step 1: Set up your brand entity block

Before scanning, you define your brand entity — the information that will anchor your AI manifests and give AI engines a reliable source of truth for your brand.

• Brand name — The exact name as it should appear to AI engines.
• Website URL — Your primary domain.
• Brand mission — One sentence describing what your brand does and for whom.
• Core innovation — Your proprietary methods, technology, or approach that differentiates you.
• Key personnel — Founders, executives, or subject matter experts who give your brand human authority.

This information is written directly into the header of your llms.txt manifest, so AI systems can identify your brand entity before reading any page-level content.

Step 2: Add your target URLs

Paste the URLs you want to scan — one per line. You can scan up to 20 URLs per session. The module automatically classifies each URL by content intent:

• Educational — Blog posts, guides, research, whitepapers, case studies.
• Transactional — Pricing pages, demo requests, contact forms, free trial pages.
• Technical — Documentation, API references, support articles, changelogs.

You can override the classification for any URL before scanning. Intent classification determines how URLs are organized in your manifest files.

Step 3: Scan for AI visibility

Click Scan for AI Visibility and the module evaluates each URL across three dimensions:

Dimension	What it measures
Semantic Density	How much of the page is substantive, fact-based content versus navigation, ads, banners, and UI chrome that AI should ignore.
Structural Readiness	Whether the page uses a clear heading hierarchy and structured data signals that help AI understand the content organization.
Entity Authority	Whether the page establishes brand credibility — through expert authorship, proprietary terminology, external citations, or named personnel.

Each dimension is scored, and the combined result produces a visibility status for the page:

• AI-Visible — The page is structurally ready for AI crawlers and should be included in both manifest files.
• Partially Visible — The page has issues that reduce AI readability. It is listed in llms.txt with a Fix Required flag, but excluded from the deep knowledge base.
• Not Visible — The page has significant structural or content problems and is excluded from manifest files until fixed.

Step 4: Review the Site Health Map

After scanning, your results are organized into a folder-level Site Health Map. Each folder shows a health bar indicating the proportion of AI-Visible pages within it, and a count of pages by status.

Executive Summary

At the top of the results, a headline shows your overall position: how many pages are AI-Visible, how many AI engines cannot currently read, and — if you have run previous scans — how much has improved since your last scan.

Quick Wins

Immediately below the summary, the module surfaces the three pages closest to AI-Visible status. These are your highest-ROI fixes — small changes that will immediately move pages from Partially Visible to fully visible.

Step 5: Act on Fix Cards

Every page that is not AI-Visible has a Fix Card — an inline panel that explains exactly what is wrong and what to do about it. Fix Cards are written for the person who will action them, not for a technical audience.

Each fix is tagged with who should own it:

• Content Writer Fix — The problem is too much navigation noise, too little brand-specific content, or missing authority signals. The fix card describes what the writer should add or restructure in plain language.
• Developer Fix — The problem is broken heading structure, missing schema markup, or a technical issue with how the page is built. The fix card describes the specific technical change required.

For pages with both types of issues, both cards appear — one for the writer, one for the developer.

Step 6: Generate manifests

Select the URLs you want to include in your manifests and click Generate Manifests. The module generates two files:

• llms.txt — Your brand entity block followed by a categorized list of URLs with directive hints. AI-Visible pages are listed cleanly. Partially Visible pages are listed with a [Fix Required] flag. Not Visible pages are excluded.
• llms-full.txt — A full knowledge base containing the cleaned, AI-readable text of your AI-Visible pages, bundled for AI systems that perform document retrieval before generating responses. Only AI-Visible pages are included.

Review before deploying

Before downloading, the module shows you an editable preview of llms.txt. Review the directive hints — the one-sentence descriptions of each page that AI systems use to understand what that page is for. Edit any that do not accurately reflect your content. What you approve here is what AI engines will read.

Step 7: Deploy

Both files are plain text. They require no server configuration, no build step, and no database. They go at your website root:

• yoursite.com/llms.txt
• yoursite.com/llms-full.txt

The module includes a Developer Handoff section with platform-specific deployment instructions for Vercel, WordPress, Webflow, and other hosting environments. Click Copy Deployment Brief to send your developer everything they need in a single clipboard paste — instructions, both files, and zero ambiguity.

Progress tracking

Every scan is saved to your scan history. The next time you run a scan for the same brand, the executive summary shows your delta — how many pages moved from Not Visible or Partially Visible to AI-Visible since your last scan. Your scan history is available in a collapsible panel for trend tracking and reporting.

AI Crawler Simulation

After generating your manifests, you can run an AI Crawler Simulation — a test that reads your llms.txt as an AI crawler would and returns a readability score, the brand entities it successfully identified, the directive hints it followed, and any content gaps it detected in your manifest. Use this to validate your manifest before deploying it to production.