SpyderBot Documentation
What is SpyderBot
This page explains what SpyderBot measures, which product to use, and what its results can and cannot tell you.
SpyderBot in one sentence
SpyderBot helps organizations measure how selected AI systems represent their brands, investigate the prompts and responses behind those results, andâwhen website measurement is connectedâunderstand AI and bot activity, AI referrals, and business outcomes.
What SpyderBot helps you understand
People increasingly use AI systems to discover products, compare providers, evaluate companies, and decide what to do next. SpyderBot helps teams answer four connected questions:
- How is our brand represented in AI responses?
- Which questions and response patterns matter most?
- How are AI systems and bots interacting with our website?
- What evidence supports the result, and what should we investigate next?
SpyderBot provides different products for these questions because they rely on different evidence.
| Product | Primary question | Main evidence |
|---|---|---|
| Brand Insights | Where does the brand stand across a defined AI-response scope? | Prompts, valid AI responses, extracted observations, citations, and report provenance |
| Prompt Intelligence | What do AI systems say for important questions, and how does it change? | Prompt-level responses, citations, sources, observations, and compatible historical comparisons |
| AI Traffic | How are AI systems, bots, and AI-referred people interacting with the website? | Connected request, session, event, goal, commerce, and audit data |
| Reports | How do I open, compare, schedule, share, or export a saved result? | Saved report records and source-product metadata |
What SpyderBot does not claim
SpyderBot does not provide direct access to a model's private reasoning, beliefs, or training process. It does not guarantee that an optimization will change a future AI response. An observed relationship is not automatically a cause.
Every result must be interpreted within its declared scope, which may include prompts or business intents, AI systems and search surfaces, language and market, competitors, observation period, connected data sources, methodology or report version, and available evidence and permissions.
Observed, estimated, and interpreted data
- Observed: directly captured in an eligible response, request, event, session, source field, or audit result.
- Calculated: derived deterministically from declared eligible data.
- Estimated: modeled or extrapolated rather than measured from a connected first-party source.
- Classified or inferred: interpreted from evidence under a documented user-facing category or analytical finding.
- Unavailable or insufficient: the required data, permission, connection, or valid population is missing.
These labels are not interchangeable. An estimate is not first-party traffic. A citation is not a visit. A bot request is not proof that a page was used in an AI answer.
Open Choose Your Workflow to select the product that matches your question, or use Quick Start for the shortest path to a first useful result.