Methodology
Directional visibility, calculated from documented AI responses
Vislyt measures a defined sample of prompts, providers, models and dates. It does not claim access to internal algorithms or an exact universal ranking.
Mention rate
Prompt runs in which the tracked brand is detected divided by all completed prompt runs in the selected slice.
Visibility score
Position 1 receives 1.0, position 2 receives 0.8, position 3 receives 0.6, position 4 receives 0.4, position 5+ receives 0.2 and absence receives 0. The mean is multiplied by 100.
Share of Voice
The tracked brand’s weighted position score divided by the combined weighted score of all tracked brands.
Average position
The arithmetic mean of detected positions, excluding runs where the brand is absent.
Detection
Unicode normalization, case-insensitive word boundaries, configurable aliases and list-order parsing run first. Model assistance is reserved for uncertain cases.
How Vislyt measures ChatGPT visibility
Vislyt runs tracked prompts through OpenAI’s API with web search enabled. It stores the generated answer, citation annotations, model information, scan date and request configuration. These results are directional and may differ from personal ChatGPT sessions because the consumer product can use different models, settings, personalization and product configurations.
Limits
Outputs may change between runs. Personalization, provider updates and model behavior can affect results. Scores should be read as evidence-backed directional signals.
Start with evidence you can inspect.
Run a free visibility check first, or start a 14-day trial to build a repeatable tracking workflow.