Evidence layer
Probative surfaces brought into scope by this page
This page does more than point to governance files. It is also anchored to surfaces that make observation, traceability, fidelity, and audit more reconstructible. Their order below makes the minimal evidence chain explicit.
- 01Evidence artifactbridge-vocabulary.json
- 02Evidence artifactconcept-registry.json
- 03Evidence artifactserp-ownership.json
- 04Evidence artifactfamily-proof-requirements.json
bridge-vocabulary.json
/bridge-vocabulary.json
Published surface that contributes to making an evidence chain more reconstructible.
- Makes provable
- Part of the observation, trace, audit, or fidelity chain.
- Does not prove
- Neither total proof, obedience guarantee, nor implicit certification.
- Use when
- When a page needs to make its evidence regime explicit.
concept-registry.json
/concept-registry.json
Published surface that contributes to making an evidence chain more reconstructible.
- Makes provable
- Part of the observation, trace, audit, or fidelity chain.
- Does not prove
- Neither total proof, obedience guarantee, nor implicit certification.
- Use when
- When a page needs to make its evidence regime explicit.
serp-ownership.json
/serp-ownership.json
Published surface that contributes to making an evidence chain more reconstructible.
- Makes provable
- Part of the observation, trace, audit, or fidelity chain.
- Does not prove
- Neither total proof, obedience guarantee, nor implicit certification.
- Use when
- When a page needs to make its evidence regime explicit.
family-proof-requirements.json
/family-proof-requirements.json
Published surface that contributes to making an evidence chain more reconstructible.
- Makes provable
- Part of the observation, trace, audit, or fidelity chain.
- Does not prove
- Neither total proof, obedience guarantee, nor implicit certification.
- Use when
- When a page needs to make its evidence regime explicit.
AI brand monitoring
AI brand monitoring covers methods that observe what generative systems say about a brand over time. It may track presence, citations, associated competitors, categories, sentiment, recommendability, sources and wording changes.
Observation becomes interpretable only when the protocol is explicit.
Working definition
AI brand monitoring is the repeated, dated and comparable collection of generated outputs about a brand under a defined set of prompts, intents, models, languages, regions, browsing modes and evidence-retention rules.
The key word is comparable. A dashboard that changes prompts, models or samples without versioning those changes can create apparent drift caused by the protocol itself.
What monitoring can observe
| Measure | Use | Limit |
|---|---|---|
| Presence | Whether the brand appears | Not how it is understood |
| Citation | Visible sources | A cited source is not always governing |
| Share of voice | Frequency in a prompt corpus | Strongly dependent on that corpus |
| Sentiment | Tone | Not fidelity |
| Category | Framing | Requires a comparison canon |
| Competitors | Semantic neighbourhood | Does not prove a real market relation |
| Recommendability | When the brand is proposed | Does not guarantee legitimacy or stability |
| Stability | Change over time | Requires controlled conditions |
Baseline and prompt families
Defensible monitoring begins with an AI perception baseline. Prompts should represent real intents: definition, comparison, recommendation, verification, problem, purchase, risk, alternatives and use cases.
Brand-named prompts test direct knowledge but not generic emergence. Generic prompts test recommendability but may not expose identity details. Both are required.
What monitoring does not govern
Monitoring does not select the legitimate source for a claim, correct the canon, disambiguate an entity, prove that a modification caused a change, or decide whether qualified criticism should remain.
AI brand monitoring versus representation governance separates four operations: observation, diagnosis, correction and governance of evidence, limits and re-observation.
Minimum collection contract
Each observation should preserve the exact prompt and intent family; product or model and available version; date, language, known region and browsing state; full answer and citations; protocol identity and changes; dated comparison canon; and an outcome status such as incident, pattern, possible drift or inconclusive.
Decision rule
Monitoring is useful when it triggers proportionate investigation. Minor wording variation is not equivalent to repeated entity confusion near a purchase decision. The goal is not identical answers. It is detection of materially relevant gaps.
AI brand monitoring is the observability layer. Governance begins when the organization decides what the observations allow it to conclude and how corrections can be verified without invented causality.