Brand Perception overview

How AI models describe you across sentiment dimensions.

Last updated: 2026-05-05

Brand Perception overview

Brand Perception is a dedicated analysis layer that goes beyond visibility counts to answer how AI models describe your brand. Where the dashboard tells you "you appear in 42% of queries", Brand Perception tells you "when you do appear, you're framed as affordable and easy to use, but rarely as enterprise-ready or secure".

What it analyses

For each Brand Perception run, Pranas:

Pulls the most recent in-scope AI Scan responses for your brand and your registered competitors.
Sends them to an LLM that extracts attributes (descriptive traits like "fast", "secure", "expensive"), sentiment polarity (positive / neutral / negative per attribute), and driver themes (clusters of related attributes).
Computes a 0–1 sentiment score per attribute and an overall composite score.
Cross-references the result against your registered competitors so every metric is comparative, not absolute.
Persists the analysis to a dedicated table so the same numbers appear in the UI and in PDF exports.

Where to find it

The Brand Perception page lives under the main navigation. The dashboard also surfaces a summary card with the top drivers and the overall sentiment score; click through for the full diagnostics.

What you'll see

Sentiment Donut — overall positive / neutral / negative split for your brand.
Brand Alignment Score — how closely the AI's perception matches your stated positioning.
Sentiment Drivers Panel — the top positive and negative themes shaping your perception.
Feature Category Heatmap — your strength per feature category vs. each competitor.
Cross-Model Comparison — how perception differs across ChatGPT, Gemini, Perplexity, etc.
Model Divergence Alerts — flags when one model's perception of you diverges sharply from the others (often an early signal of stale training data).
Gap Analysis — concrete attributes you claim but the AI doesn't echo back.
Recommendations — generated remediation actions per gap.

Panel reference

Panel	What it shows	Best used to
Sentiment Donut	Positive / neutral / negative split for your brand	Get the headline read in one glance
Brand Alignment Score	How closely AI perception matches your stated positioning	Check whether your messaging has landed
Sentiment Drivers Panel	Top positive and negative themes	Decide which narrative to reinforce
Feature Category Heatmap	Strength per feature category versus each competitor	Spot competitive holes and dominance zones
Cross-Model Comparison	Perception differences across ChatGPT, Gemini, Perplexity and others	See which surface misreads you
Model Divergence Alerts	One model diverging sharply from the rest	Catch stale training data early
Gap Analysis	Attributes you claim but the AI doesn't echo	Build a content remediation list
Recommendations	Generated remediation actions per gap	Hand work to Content Studio

When to run it

After every meaningful content push (new positioning page, major launch) to measure pickup.
Quarterly as a baseline cadence even without changes.
Before a board update so the perception narrative is fresh.
When SOV moves but Visibility doesn't — the shift is usually in narrative, not raw count.

Brand Perception runs are independent of AI Scans and use a smaller credit footprint, but they require recent scan data to read from. If you haven't scanned in 30 days, run a scan first.

Browsing history

Each run is saved with a timestamp. Use the history popover to compare today's perception against last quarter's and see exactly which drivers shifted. The page is scope-isolated — when you load a historical run, every panel re-reads from that run's persisted data, so you never accidentally mix two runs.

Perception scoring methodology
Gap analysis
Feature category heatmap
Browsing perception analysis history