Monitoring dashboards add three analytical layers on top of individual scans.
Share of voice (SOV)
Mention rate measures you against the questions. Share of voice measures you against your competitors: of all brand mentions across a scan's answers, what fraction were you?
SOV reframes the goal. In a category where AI answers mention 5 brands per answer, a 40% mention rate can still mean you're the 8th most-mentioned name. The SOV chart shows your slice next to your top competitors' slices, scan over scan — so you can see whether you are gaining share, and from whom.
Sentiment
Every brand mention is classified in context as positive, neutral or negative. This distinguishes "Acme is a popular choice, though support reviews are mixed" from a clean recommendation. On the dashboard, sentiment appears summarized per scan and per prompt, alongside competitors' sentiment.
Sentiment matters most when it diverges: a high mention rate with sliding sentiment means AI answers are increasingly mentioning you with caveats — usually traceable to review sites and comparison content in the citation list.
Trend history
Each weekly scan appends one data point. Trend charts track mention rate, grade, SOV and sentiment over time, with scan-to-scan deltas marked throughout the dashboard.
Reading trends honestly:
- Judge by weeks, not by single scans. Model outputs vary run to run; one dip is noise, three consecutive dips are signal.
- Annotate changes mentally. When you ship fixes from the recommendations board, watch subsequent scans for movement — that link between action and trend is the whole point of monitoring.
- A stable prompt set is the precondition. Trends compare like with like because prompts stay constant; large prompt edits create step changes (details).
Where to find them
Everything lives on your monitoring dashboard: KPI cards up top (with deltas and explanatory tooltips), the trend and SOV charts below, sentiment per prompt in the prompt table drill-down.