Evidence first. Comparable remeasurement second.

Promptence is designed to separate observed AI visibility from assumptions. A measurement is tied to defined buyer demand, provider-specific observations, timestamps, and persisted evidence so later measurements can be compared on a governed basis.

1. Define buyer demand

Measurement starts with the company, its market context, and buyer-intent questions or demand clusters. The goal is to test questions that represent plausible customer demand, not prompts written only to force a brand mention.

2. Record provider-specific observations

Queries are measured through configured AI-provider integrations. Promptence persists the relevant run context and observations so results are attributable to a measurement rather than presented as an unsupported score.

3. Evaluate visibility and evidence

Measurements can include brand mentions, competitor presence, citations when available, and share-of-demand style metrics derived from persisted observations. The evidence status and measurement context matter alongside the headline metric.

4. Prioritize changes

Observed gaps can be converted into recommendations and change sets. Managed workflows keep review and approval as explicit gates before work is treated as implemented.

5. Remeasure comparable demand

After implementation, Promptence can measure the same governed demand again and compare persisted before-and-after evidence. A measured delta is reported as a delta; it is not automatically presented as proof that one change caused the entire outcome.

Limitations

Generative systems are probabilistic. Model versions, provider behavior, retrieval, geography, freshness, and other factors can change outputs. Promptence therefore focuses on repeatable evidence, provenance, confidence, and remeasurement rather than guaranteed placement in an AI answer.