Scrunch uses multiple methodologies to collect data from AI platforms like ChatGPT, Perplexity, Gemini, Claude, and others, including browser automation and official platform APIs.

Additional context: Scrunch uses the appropriate methodology for each AI platform. All data, regardless of collection method, is measured against a large and continually updated dataset of responses collected directly from inside AI platforms to make sure responses are accurately surfaced in Scrunch.

Example

For example, imagine a Scrunch user tracks the prompt, “How can I optimize my brand for AI search engines like ChatGPT?”

Scrunch will use platform-specific methodologies to collect responses across AI platforms and, via machine learning and natural language processing, display results, including:

• Brand presence in AI answers by percentage over a set time period
• Competitive presence in AI answers by percentage over a set time period
• Brand position in AI answers (e.g., top, middle, bottom) by percentage over a set time period
• Brand sentiment in AI answers (e.g., positive, mixed, negative) by percentage over a set time period
• Citations in AI answers (including the user’s brand, competitors, and third parties) by percentage over a set time period

If you want to go deeper: Scrunch allows you to self-serve any combination of metrics, filters, and breakdowns you want to create custom data visualizations.

Follow-up question: How frequently does Scrunch collect prompt data?

For new prompts being tracked, Scrunch collects AI responses daily for the first 14 days. After that Scrunch shifts to a default 72-hour refresh cadence, but prompt data can be refreshed at any time.