Help Understanding your results
Prompts & answers
Open a prompt to read the AI answers, compare samples and inspect the mentions and cited pages behind its results.
Prompts & answers overview
Open AI visibility → Prompts & answers. The table shows results for each prompt, with its category and last-scan status. Search for a prompt or narrow the list with the category and model filters. Click a sortable column heading to change the order.
The five result columns show:
- Visibility — the share of measured answers that mention your brand, with the mentioned-answer count and total.
- Sentiment — the tone of the mentions of your brand.
- Competitors — tracked competitors named in the answers.
- Position — your average position among the brands mentioned; it is not a Google search ranking.
- Sources — distinct URLs cited in the answers.
While a scan is running, result cells show a loading indicator. A missing result or disabled model is different from a completed answer that did not mention you. Check the scan status before comparing values.
To add or organise prompts, see Prompts and categories.
Open a prompt and choose a scan
Click the prompt’s row. The detail page shows an overall result and a row for each AI model. Click a metric to open the relevant detail tab; use its question-mark control to read the explanation.
Use the scan selector above the results to switch from the latest results to a saved scan. Check the selected date when comparing answers. The latest view uses the most recent completed answers available for each prompt and model; their dates can differ when a model did not complete a newer scan.
Models that are off, unavailable or have an error display their status separately. Change enabled models in AI models.
Read the answers and samples
The detail page has three tabs:
- Visibility & Sentiments shows the answer with brand mentions and sentiment phrases highlighted.
- Competitors & Position shows the competitor mentions and brand ordering in the answer.
- Sources shows the cited URLs for each model.
When a model has several answer samples, use its sample selector to read them individually. Check the model, scan and sample before comparing two passages. One sample does not represent every answer in the overall result.
The sentiment value summarises tone around your brand. Individual coloured phrases are evidence and can differ from that summary. Review the actual text when a score needs explaining; the colour of a phrase about another brand is not a score for yours.
Follow a result to its evidence
Start with a row you want to investigate, open the answer, then inspect its competitor mentions or cited pages. A source link opens the cited page; it does not prove that every statement in the answer came from that page.
Use Sentiment and Sources to compare these details across prompts. Use AI trends to compare measurements over successive scans.