SpyderBot Documentation
FAQ
Is Brand Insights measuring search rankings?
No. Brand Insights measures patterns in collected AI responses. Ranking metrics apply only when eligible responses contain ordered or explicit ranking evidence.
Is AI Share of Voice the same as Google Share of Search?
No. AI Share of Voice describes relative brand presence within the declared Brand Insights response population and competitor set.
Does a mention mean the brand was recommended?
No. A mention shows that the brand appeared. A recommendation requires an eligible endorsement, shortlist, or suggested choice.
Does a citation mean a user visited the website?
No. Citation and referral data are different. Website visits require AI Traffic or another compatible analytics source.
Why can multiple brands have high Mention Coverage?
One response can mention several brands. Mention Coverage is calculated independently for each brand and does not need to sum to 100%. AI Share of Voice is the comparative distribution that uses a shared denominator.
What does complete prompt-system coverage mean?
Every eligible combination of prompts and AI systems contains at least one valid response. It does not mean equal response depth in every cell.
Is execution count enough for a report?
No. Review Prompt Set coverage, prompt-system coverage, response depth, valid-response rate, prompt-family coverage, AI-system scope, competitor scope, and metric-level sample sizes.
Does a high Reliability state mean the AI response is true?
No. Reliability describes evidence fitness and consistency within the report scope.
Why is Founder Negative Context split into Organic and Risk-aided?
Prompts that explicitly ask for controversy or risk naturally produce more negative material. Separating them prevents risk-seeking prompts from inflating the organic brand-context metric.
Can Brand Insights measure conversions?
Not from prompt responses. Conversions require compatible first-party event tracking through AI Traffic or another connected system.
Can Brand Insights explain why an AI model made an internal decision?
No. It can show observable patterns, response evidence, citations, associations, and possible contributing factors.
How often should an analysis be repeated?
Repeat when establishing a baseline, evaluating a significant change, monitoring an important period, or measuring compatible trends. Use Prompt Observatory for continuous prompt-level monitoring.