Data Collection

GeoSnap gathers data by interacting directly with AI engines through their user interfaces. Every question is sent as if posed by a real user, and the complete response is collected and stored for analysis.

This approach ensures that the data reflects the genuine user experience, rather than a filtered or simplified version accessible only via API.

Response Analysis

Each AI response is analyzed through different layers:

• Brand detection: identifying all brands mentioned in the response

• Position analysis: determining the relative position of each brand

• Sentiment extraction: analyzing the tone and perception associated with each brand

• Source identification: extracting all cited or referenced sources

• Intent mapping: associating the response with the type of intent of the question

Aggregation and Metrics

Data collected from hundreds of responses are aggregated into synthetic metrics. This aggregation considers the natural variability of AI responses: a single response is not significant, but the pattern emerging from hundreds of responses is statistically robust.

The main metrics (AI Rating, Sentiment, Competitors, Sources) are the result of this aggregation. Each metric is calculated using specific methodologies designed to produce reliable and comparable results over time.

Managing Variability

AI responses are inherently non-deterministic: the same question asked twice may yield different answers. GeoSnap manages this variability through volume (hundreds of questions), diversification (different formulations for the same intent), and statistical aggregation.

This approach produces stable and reliable metrics, despite the variable nature of underlying AI responses.

Data Update

Data is updated periodically to reflect the evolution of AI responses over time. LLMs are continuously updated, web sources change, and your brand perception evolves. Periodic monitoring allows tracking these changes and measuring the impact of your optimization actions.

Privacy and Security

GeoSnap does not collect personal data from AI platform users. The analysis concerns only the responses generated by AI engines in response to generic questions about your market category. The data collected is handled with the highest security standards and is accessible only to the project owner.