AIVM-1.0 measurement framework
Five measurement modes separate spontaneous recall, current evidence, competitive comparison, reputation and purchase intent so incompatible experiments are not collapsed into one story.
How to interpret this document
This content describes technical and methodological behavior that is implemented or explicitly planned in the product. When a control depends on configuration, a provider, a secret, a contract or legal approval, that dependency must remain visible.
Why multiple methodologies exist
A brand can have strong spontaneous recall and still underperform when current web evidence is required. It can also have positive general perception but weak recommendation in a purchase context. AIVM-1.0 separates these questions instead of hiding them inside one monitoring mode.
Natural Visibility
Runs without web search and instructs the model not to force brand mentions. It is designed to observe spontaneous recall and knowledge already available to the model.
Grounded Visibility
Requires current web search and asks the model to support the answer with verifiable sources. It is appropriate when recency, external evidence and citation analysis matter.
Competitive Visibility
Frames the problem as a market comparison. Known competitors may be considered when relevant without forcing the monitored brand into the answer.
Reputation Visibility
Focuses on trust and public perception. The prompt design asks for balanced positive, negative and neutral evidence and reduces the risk of treating isolated allegations as established fact.
Purchase Intent Visibility
Simulates support for a real buying or contracting decision, considering fit, value, price when available, trust, availability, limitations and trade-offs.
Series signature
Each run persists a signature containing methodology, country, locale, selected models, prompt_keys and repetitions. Historical comparisons should also preserve compatible score, statistics and taxonomy versions.
Interpretation rule
A difference between two runs should only be interpreted as market movement when the experiments are sufficiently comparable. Changing methodology, model set or prompt set can change the result by experimental design rather than real brand movement.