Gumshoe Reports are Diagnostic Tests
Gumshoe reports are not based on historical traffic or past user searches, because AI models don’t share that data. Unlike traditional SEO tools, we don’t have access to what users have asked AI tools in the past, and neither does anyone else.
Instead, Gumshoe is a real-time diagnostic. Each report is a snapshot that shows how AI models respond right now to specific persona-driven prompts.
These prompts simulate the types of questions users might ask. The report then shows:
- Which brands each model mentions
- How often they are mentioned
- What sources the model used
- What context or reasoning it provides
So What Does That Mean?
You’re not seeing data from the past week, month, or year.
- AI models don’t share usage logs, so we can’t report on what real users asked or clicked.
You are seeing how your brand ranks in AI responses right now.
- Gumshoe tests specific questions across personas and topics to simulate realistic user behavior.
Mentions reflect current results from that test run.
- These are not tied to a specific historical time window, such as “last 30 days.” They show how many times your brand was cited during that particular run.
Why It Works
AI models like ChatGPT, Claude, Gemini, and Perplexity are trained on a combination of public web content and inferred user behavior, but they don’t expose that training data or activity.
Gumshoe provides a method for testing how models behave based on their training, in a consistent and repeatable manner.
To learn how to schedule your reports to run on a regular timetable, click here.