SpyderBot Documentation
Prompt Set
Prompt Set explains which questions produced the report and whether the report covers the intended brand, category, market, competitor, commercial, and leadership contexts.
What it answers
- Which prompts were included?
- Why were they selected?
- Which prompts are Stable or Adaptive?
- Which Intelligence groups use each prompt?
- Which AI systems were eligible?
- How many valid responses exist for each prompt-system pair?
How prompts are selected
SpyderBot begins with Prompt Universe candidates and may add qualified supplemental prompts from approved brand, category, product, competitor, market, or evidence context.
Selection considers Prompt Priority Score, required Intelligence-group coverage, intent and prompt-family coverage, semantic diversity, Stable trend anchors, evidence availability, competitor and market relevance, and redundancy, ambiguity, and bias risks.
Exact selection weights and internal scoring rules are proprietary. Priority Score helps select prompts; it does not automatically make one prompt count more heavily in analytical results.
Stable and Adaptive prompts
Stable prompts support compatible comparison over time. Adaptive prompts reflect current products, competitors, market questions, opportunities, risks, or evidence gaps. Both types belong to the same immutable Prompt Set snapshot for the report run.
Prompt families
- Unbranded discovery and visibility
- Brand context and organic sentiment
- Comparison, ranking, and recommendation
- Commerce and purchase intent
- Leadership and finance â organic
- Leadership risk-aided
- Entity, citation, and factual control
Prompt Set health
Review canonical prompts, Stable and Adaptive counts, AI systems, prompt-system cells, valid-cell coverage, valid-response depth, executions, and valid versus invalid responses.
Public documentation describes why a prompt was selected using reason labels, not proprietary scoring coefficients.