Intelligent Brand Index (IBI)
The Intelligent Brand Index (IBI) is a comprehensive scoring system designed by Advanced Analytica to evaluate how effectively a brand's website communicates with AI systems and large language models (LLMs),search engines, and accessibility tools. It combines structured data quality, discoverability, AI-readiness, and accessibility into a single score out of 100.
The IBI system contains the accessibility scoring component of the IBI, powered by Lighthouse CLI and axe-core.
Purpose of IBI
As AI-generated answers, voice assistants, and search overviews increasingly depend on machine-readable content, brands must ensure they are discoverable and interpretable by machines. IBI provides a benchmark for:
- AI discoverability via structured data (e.g., JSON-LD + Schema.org)
- Governance and crawlability (robots.txt, sitemaps, AI guidance files)
- AI-readiness (semantic markup, content freshness, media accessibility)
- Accessibility based on WCAG 2.1 AA conformance
IBI Scoring Framework
The full IBI score is composed of four main categories:
| Category | Max Points |
|---|---|
| A. Structured-Data Quality | 50 |
| B. Discoverability & Governance | 30 |
| C. AI-Friendliness & Media | 20 |
| D. Accessibility (WCAG 2.1 AA) | 15 |
| Total Raw Points | 115 |
The final score is normalised to a 0–100 scale.
Expanded IBI Scoring Categories**
The IBI evaluates a website’s machine-readability and digital governance using a multi-dimensional score. It ensures your site is visually appealing, machine-readable and LLM-interpretable.
A. Structured Data Quality
Definition:
Measures the presence, accuracy, and coverage of semantic metadata embedded on the site using formats like JSON-LD, Microdata, or RDFa, following Schema.org vocabularies.
Why It Matters: AI systems (including Google, ChatGPT, and voice assistants) increasingly depend on structured data to understand entities, products, events, and questions and asnwers (QAs). Well-formed data increases search visibility and enables features like rich snippets and knowledge graph inclusion.
Scoring Components:
| Sub-Metric | Description | Pts |
|---|---|---|
| A-1. Schema.org Type Coverage | Presence of types like Organization, Brand, Product, Event, FAQ, etc. |
15 |
| A-2. Field Completeness | Degree to which required/recommended fields are populated (e.g., name, url) |
10 |
| A-3. Nested Entities & Relationships | Use of nested schema (e.g., Product has Brand, Offer, Review) |
10 |
| A-4. JSON-LD Syntax Validity | Checks for valid, parseable JSON-LD syntax | 5 |
| A-5. Duplicate / Conflicting Markup | Penalizes repeated or contradictory structured data entries | 5 |
| A-6. Use of Custom Extensions | Bonus for custom vocab extensions where applicable | 5 |
B. Discoverability & Governance
Definition: Audits whether a site is technically configured for bots and agents to discover, crawl, and interpret the content — while also supporting newer AI agents and systems.
Why It Matters: Good governance helps ensure that crawlers (from search engines to AI scrapers) can properly index and access your content — improving visibility in SERPs, chat and voice search.
Scoring Components:
| Sub-Metric | Description | Pts |
|---|---|---|
B-1. Valid robots.txt file |
Should exist, be reachable, and contain appropriate disallow/allow directives | 5 |
| B-2. Sitemap Presence | sitemap.xml is present and submitted |
5 |
B-3. AI Guidance (ai.txt) |
Support for ai.txt or llms.txt to guide AI agents (e.g., OpenAI, Anthropic) |
5 |
| B-4. Canonical Tag Usage | Correct canonical tag on pages to prevent duplicate content confusion | 5 |
| B-5. Page Indexability | Page should not be noindex unless intended |
5 |
| B-6. Crawl Budget Optimization | Avoid excessive redirects or dead links | 5 |
C. AI-Friendliness & Media Semantics
Definition: Evaluates how machine-readable your content is, especially rich media like images, video, and interactive widgets.
Why It Matters: AI systems like GPT-4, Gemini, and Perplexity attempt to interpret content across modalities. Media needs metadata, captions, and context to be useful to machine reasoning.
Scoring Components:
| Sub-Metric | Description | Pts |
|---|---|---|
| C-1. Descriptive Alt Text | Contextually accurate alt attributes on visual content |
5 |
| C-2. Video Metadata & Captions | Titles, transcripts, and subtitles present (especially for hero content) | 5 |
| C-3. Headless & Semantically Tagged Content | Uses <article>, <section>, <main>, etc. to define intent |
5 |
| C-4. Freshness & Update Signals | lastModified, updated timestamps, or freshness indicators for content |
5 |
D. Accessibility
[Already well defined in your original message — retained with slight expansion below]
Definition: Assesses the ability of people with impairments (e.g., vision impairment, limited mobility) to navigate and consume the content, per WCAG 2.1 AA.
Why It Matters: Accessibility is not just ethical and often legally required — it improves semantic clarity and boosts SEO/AI-readability for all users and agents.
Scoring Components:
| ID | Metric | Description | Max Points |
|---|---|---|---|
| D-1 | Alt text on images | All meaningful images have non-empty alt attributes | 3 |
| D-2 | Semantic heading order | Headings follow logical h1 > h2 > h3 structure |
2 |
| D-3 | Colour contrast (4.5:1 or better) | Meets WCAG contrast standards for legibility | 3 |
| D-4 | Keyboard navigation | All interactive elements are accessible without a mouse | 2 |
| D-5 | ARIA landmarks and skip links | Uses roles like main, navigation, complementary |
2 |
| D-6 | Critical accessibility violations < 5 | Lighthouse/axe-core reports fewer than 5 critical issues | 3 |
Final Score Computation
- Each raw category is summed (max 115 pts).
- Then normalised to a score out of 100:
ibi_score = round((raw_score / 115) * 100, 1)
How It Works
1. Run Lighthouse
python ibi_accessibility_scoring.py https://example.com output_dir/
This uses Lighthouse CLI (Node.js required) to generate a JSON report and extract accessibility data.
2. Extract Metrics
The script parses Lighthouse's audits and evaluates the presence and quality of WCAG-relevant elements.
3. Score Calculation
Each sub-metric is weighted according to the IBI framework and produces a total score out of 15.
Dependencies
- Node.js with Lighthouse CLI (
npm i -g lighthouse) - Python 3.8+
- Chrome installed and available to CLI
Output Example
Accessibility Sub-Metrics:
D1_alt_text: true
D2_headings_order: true
D3_contrast: false
D4_keyboard_nav: true
D5_landmarks: true
D6_critical_violations: true
Total Accessibility Score: 13 / 15
Notes
- This module evaluates accessibility only. For the full IBI score (structured data, governance, and AI-friendliness), integrate this into the broader pipeline.
- Only public URLs are supported unless Lighthouse is run with custom Chrome sessions.