Narratives

Narratives let you zoom in on a single topic and track how AI models discuss it over time. While Perception gives you a broad overview, Narratives drills into the stories that matter most.

What is a narrative?

A narrative is a specific topic you want to monitor across AI models. It could be a product launch, a public controversy, a competitive claim, or anything where you need to know what AI is telling people.

Each narrative has:

• Name - A short label for the topic
• Description - Context for the AI analysis
• Keywords - Terms that help identify relevant responses (1-10)
• Comparison brand (optional) - A competitor to track on the same topic

Creating a narrative

Click New Narrative from the Perception tab or the Narratives page.

1. 1Give it a clear, descriptive name
2. 2Write a description that provides context for the analysis
3. 3Add keywords that relate to the topic
4. 4Optionally add a comparison brand to benchmark against

Once created, Trakkr immediately runs the first analysis across all AI models.

The narrative detail view

Each narrative opens into a detailed view with:

Score

An overall score (0-100) representing how accurately and favorably AI models discuss this topic. Higher is better.

Narrative arc

A trend chart showing how the score changes over 30, 60, or 90 days. If you added a comparison brand, their trend line appears as a dashed overlay.

Provider breakdown

See how each AI model (ChatGPT, Claude, Gemini, Perplexity) handles this topic individually. Some models may be more accurate than others.

Corrective penetration

The percentage of AI models that have adopted your corrective messaging on this topic. This is how you measure whether your communications are working.

Corrective messages

Corrective messages are the statements you want AI models to adopt. For example, if AI models are spreading outdated information, you publish a correction and then track whether AI picks it up.

To add a corrective message:

1. 1Open the narrative detail view
2. 2Click Add Corrective
3. 3Enter the message name and key phrases to detect
4. 4Optionally add a URL and publication date

Trakkr checks each AI model's response for your key phrases and reports the penetration rate.

Narrative statuses

Critical narratives float to the top of your list so they get immediate attention.

Best practices

• One topic per narrative - Keep narratives focused on a single topic for clearer tracking
• Use specific keywords - Generic keywords produce noisy results
• Add corrective messages early - Start tracking penetration as soon as you publish a response
• Review weekly - Check the narrative arc to spot trends before they become problems
• Resolve when done - Archive narratives that are no longer relevant to keep your list clean