# CiteCue Methodology

## Overview

CiteCue analyzes why AI assistants recommend—or ignore—your business. The methodology combines website inspection, prompt generation, multi-model querying, gap analysis, and prioritized remediation.

## Phase 1: Website scan

CiteCue inspects your site for signals that influence AI recommendations:

- **Content quality** — Clarity, depth, and usefulness of product and educational content
- **Semantic coverage** — Whether you answer the full range of customer questions
- **Authority** — External citations, reviews, and domain trust
- **Trust signals** — Methodology, expert credentials, transparency, and original evidence
- **AI crawlability** — llms.txt, robots.txt, sitemap, structured data, static readable pages
- **Entity strength** — Consistent brand, product, and organization identity

These dimensions combine into an **AI Readiness Score**.

## Phase 2: Prompt generation

CiteCue generates hundreds of realistic customer prompts across categories:

- Recommendation queries ("What is the best X for Y?")
- Comparison queries ("X vs Y for small teams")
- Buying-stage queries ("Easiest X to set up")
- Problem-solving queries ("How do I solve Z with X?")

Prompts reflect how real buyers ask AI—not keyword-stuffed search queries.

## Phase 3: Multi-model analysis

Each prompt is sent to leading AI models. CiteCue tracks:

- Whether your brand is **mentioned** or **recommended**
- Which **competitors** appear instead
- Which **sources** AI cites as evidence
- **Position** and prominence in the answer
- **Sentiment** and confidence indicators
- **Frequency** across prompt variations

Supported platforms: ChatGPT, Claude, Gemini, Perplexity (Google AI Overviews planned).

## Phase 4: Gap explanation

For every missed recommendation, CiteCue traces the root cause:

- Missing comparison or FAQ content
- Weak trust or authority signals
- Incomplete semantic coverage for a topic cluster
- Technical barriers to AI access (no structured data, JS-only content, missing llms.txt)
- Competitor content that better matches the prompt intent

## Phase 5: Prioritized fixes

Recommendations are scored by:

- **Impact** — How many high-value prompts the fix addresses
- **Difficulty** — Implementation effort (content, technical, or both)
- **Evidence** — Why this fix should improve recommendation rates

Deliverables can include ready-to-use FAQs, comparison pages, buying guides, structured data markup, and AI-readable pages.

## Phase 6: Deployment

Fixes can be published directly to supported CMS platforms (WordPress, Shopify, Webflow—on roadmap) or exported as a developer implementation package.

## Free audit

The waitlist audit samples your website, runs a subset of priority prompts, and returns an initial visibility score with top gaps to investigate.

Join the waitlist: https://citecue.com/#free-audit

## Related resources

- [AI visibility guide](https://citecue.com/ai-visibility.md)
- [AEO guide](https://citecue.com/aeo-guide.md)
- [For AI assistants](https://citecue.com/for-ai.md)