AI Citation Tracking

Instructions

Track and analyze how AI answer engines cite, reference, or surface brand content. Unlike traditional rank tracking, AI citation tracking requires prompt-based querying and response analysis.

Citation Tracking Methodology

1. Prompt Library Development

Build a library of tracking prompts that represent target queries:

• Informational prompts: “What is [topic]?”, “How does [topic] work?”
• Comparative prompts: “What is the best [product category]?”, “[Brand A] vs [Brand B]”
• Recommendation prompts: “What tool should I use for [task]?”
• Brand prompts: “Tell me about [brand name]”, “Is [brand] reliable?”

For each prompt:

• Record the exact prompt text
• Run across all target platforms (ChatGPT, Perplexity, Google AI Overviews, Copilot)
• Capture the full response including citations/links
• Tag whether brand was mentioned, cited with link, cited without link, or absent
• Note competitor mentions in the same response

2. Citation Classification

Classify each citation occurrence:

Type	Description	Value
Direct citation with link	Brand/content cited with clickable source link	Highest
Named citation	Brand mentioned by name as a source without link	High
Content paraphrase	Content clearly used but brand not named	Medium
Absent	Brand not mentioned; competitor cited instead	Gap — action needed
Incorrect attribution	Content attributed to wrong source	Negative — correction needed

3. Key Metrics

Track these metrics over time:

• Share of Answer (SoA): Percentage of target prompts where brand appears in the response
• Citation depth: Position within the response (first mentioned vs. buried)
• Link rate: Percentage of citations that include a clickable link back
• Platform coverage: Which AI platforms cite you vs. which don’t
• Competitive displacement: Prompts where you replaced a competitor (or vice versa)
• Sentiment: Positive, neutral, or negative framing when cited

4. Tracking Cadence

Frequency	Activity
Weekly	Run core prompt library (top 20 prompts) across all platforms
Monthly	Run full prompt library (50-100 prompts), compile trend report
Quarterly	Full competitive benchmark, refresh prompt library, strategy review

5. Competitive Benchmarking

For each target topic:

• Identify which brands/sources AI platforms currently cite
• Rank competitors by Share of Answer
• Analyze what cited content has that yours doesn’t (structure, authority, freshness)
• Track changes in competitive citation share over time

6. Attribution Analysis

When your content is cited, analyze the attribution path:

• Which specific page was cited?
• Was it the canonical URL or a syndicated copy?
• Did the AI system extract the correct information?
• Was the citation contextually accurate?
• Did the platform link to the most relevant page?

Automated Tracking Approaches

• API-based: Use ChatGPT API, Perplexity API to run prompt libraries programmatically
• Manual spot-checking: Supplement automation with browser-based verification
• Third-party tools: Evaluate tools like Profound, Peec AI, Otterly for tracking capabilities
• Custom dashboards: Build tracking dashboards that visualize SoA trends

Response to Citation Gaps

When brand is absent from AI responses:

1. Verify content exists for the topic
2. Check if content is structured for AI extraction
3. Confirm Schema markup is in place
4. Review llms.txt and crawler access
5. Assess E-E-A-T signals vs. cited competitors
6. Create or optimize content to fill the gap
7. Re-check after 2-4 weeks (AI index refresh cycle)

Inputs Required

• Brand name and primary domain
• Target topics and keywords
• Competitor list (3-5 primary competitors)
• Existing prompt library (if any)
• Access to AI platforms for querying
• Historical tracking data (if available)

Output Format

Monthly Citation Report

## AI Citation Tracking Report — [Month Year]

### Executive Summary
- Share of Answer: [X]% (Δ [+/-X]% from last month)
- Platform coverage: [X/4 platforms citing]
- Top gap: [topic where competitor is cited instead]

### Share of Answer by Platform
| Platform | SoA | Δ MoM | Top Cited Competitor |
|----------|-----|-------|---------------------|
| ChatGPT | X% | +X% | [Competitor] |
| Perplexity | X% | +X% | [Competitor] |
| AI Overviews | X% | +X% | [Competitor] |
| Copilot | X% | +X% | [Competitor] |

### Citation Quality
- Direct citations with link: [X]%
- Named citations without link: [X]%
- Content paraphrased (unattributed): [X]%

### Competitive Landscape
| Topic | Our SoA | Top Competitor | Their SoA |
|-------|---------|---------------|-----------|
| ... | ... | ... | ... |

### Action Items
1. [Gap to close] — Priority: [High/Med/Low]
2. [Content to optimize] — Priority: [High/Med/Low]

Anti-Patterns

• Tracking rankings instead of citations — AI answers don’t have “rank 1”; measure Share of Answer
• Querying once and extrapolating — AI responses vary; track over time for reliable trends
• Ignoring platform differences — Each AI engine cites differently; track each separately
• Manual-only tracking — Doesn’t scale; build automated prompt libraries early
• Vanity metrics — “Mentioned once” is not the same as “consistently cited as primary source”
• Stale prompt libraries — User query patterns change; refresh prompt libraries quarterly
• Ignoring negative citations — Track and address cases where brand is mentioned negatively