Narrative Drivers
What themes the AI uses when describing brands in your space.
Last updated: 2026-05-05
Narrative Drivers
Narrative Drivers shows which prompt categories each brand owns in AI answers. It's the categorical companion to your headline metrics: AI Score tells you how strong you are overall; Narrative Drivers tells you where that strength comes from — and where it's missing.
What it shows
A brand × category grid, locked to Unaided (organic) prompts so self-mention bias from Aided and Comparison prompts can't inflate it. Your brand sits on the first row, competitors follow, ranked by their average score.
Each cell is a coverage-adjusted AI Score for that brand in that category:
cell = AI Score when mentioned × coverage in that category
Coverage is the share of that category's scans where the brand appeared at least once. This is the identical formula used by the headline AI Score card and the Top Brands widget, so the numbers reconcile across the dashboard.
Hover any cell to see the arithmetic: the score when mentioned, the covered scan count, and the resulting adjusted value.
Reading the cells
| Cell | Meaning |
|---|---|
| Green (≥ 70) | Strong, consistent presence in this category. |
| Amber (40–69) | Present but inconsistent — either mid-ranked when named, or named in only some scans. |
| Red (< 70, incl. 0.0) | Weak. A red 0.0 means the category was scanned and the brand was never mentioned — a genuine zero, not missing data. |
| Grey "—" | Not tested — no scans exist in that category for the current filters. |
The distinction matters: a 0.0 is an actionable gap, a — just means you need prompts in that category.
Why misses count as zero
Averaging only the answers where a brand appeared makes every brand look excellent — the score then measures "how good are we when we win" rather than "how visible are we". Because a scan that never mentions the brand is a valid, informative result, it is counted as a zero. This is the same coverage-inclusive principle applied across AI Score, Coverage, Mention Rate and Top Brands.
Expanding a brand row shows the underlying per-prompt scores. Those are raw scores from scans where the brand was mentioned, so they will read higher than the coverage-adjusted cell above them — they're for diagnosis, not for reporting.
How to use it
- Find your category gaps. A red
0.0next to a competitor's green cell is a content/citation gap in that category, not a product gap — publish and get cited for that category's questions. - Read the insight line. The summary above the grid names the category where you lead furthest ahead of the competitor average and the one where you trail furthest behind, using the same coverage-adjusted numbers. It only appears when both categories have at least 3 scans.
- Prioritise reach before rank. If a cell is low because coverage is low (hover to check), fix visibility first; if coverage is high but the score when mentioned is low, fix how you're positioned in the answer.
Differences from Brand Perception
Brand Perception (a separate page) runs a deeper analysis on a per-brand basis with sentiment scoring, gap analysis, and recommendations. Narrative Drivers on the dashboard is the bird's-eye summary across all brands; use it to spot what to investigate, then drill into Brand Perception.
Related articles
- Brand Perception overview
- Co-occurrence analysis
- Top Brands widget
- Gap analysis