> ## Documentation Index
> Fetch the complete documentation index at: https://www.pierview.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# Perception
> Understand how AI models describe your brand and how you compare to competitors
The **Perception** dashboard breaks down sentiment beyond a single score. It shows *which LLMs* are positive or negative about your brand, *which categories* drive that signal, and *what specific language* AI models use.
Sentiment is the quality layer on top of visibility. If you're visible but perception is negative, you're being mentioned for the wrong reasons.
***
## Sentiment score
Your overall sentiment score is a 0–100 value calculated from extracted perceptions in AI responses:
$$
\text{Sentiment Score} = \frac{\text{Positive perceptions}}{\text{Positive} + \text{Negative perceptions}} \times 100
$$
**Score ranges:**
| Range | Meaning |
| ------ | ------------------------------------------------------------ |
| 70–100 | Positive: AI consistently recommends or praises your brand |
| 50–69 | Mixed: present in the conversation, but not clearly endorsed |
| 0–49 | Negative: AI surfaces drawbacks, objections, or warnings |
***
## Sentiment by LLM
Different AI models often have different tones toward the same brand. The per-LLM breakdown shows:
* **Positive %**: share of positive perceptions for that model
* **Positive count / Negative count**: raw mention volumes
Use this to identify model-specific narratives. A brand can be recommended by Claude but flagged as expensive by ChatGPT. The fix differs per model because the sources each model relies on differ.
***
## Perception categories
Pierview classifies extracted perceptions into 8 categories:
| Category | What it captures |
| ------------------------ | ------------------------------------------------------ |
| **Pricing & Value** | Cost, pricing tiers, value for money |
| **Performance** | Speed, reliability, accuracy |
| **Usability & Apps** | Ease of use, onboarding, mobile/desktop experience |
| **Products & Features** | Specific capabilities, integrations, feature gaps |
| **Support & Service** | Customer support quality, documentation, response time |
| **Brand & Trust** | Reputation, credibility, brand recognition |
| **Delivery & Logistics** | Shipping, setup time, fulfillment (e-commerce/SaaS) |
| **Other Themes** | Anything that doesn't fit the above |
Each category shows a **tone** (positive, negative, neutral, or mixed) based on the distribution of mentions within it. Categories with more data are weighted more heavily.
### How to use category data
A negative score in **Pricing & Value** means AI responses frequently surface cost objections. The fix is content that directly addresses value: comparisons, ROI calculators, pricing FAQs.
A negative score in **Products & Features** often points to a specific missing feature being surfaced repeatedly. Check the extracted facts to see which feature.
***
## Brand strengths and weaknesses
Pierview extracts the specific phrases and attributes being mentioned, then surfaces the top:
* **Strengths**: most frequently cited positive attributes (e.g., "fast onboarding", "best for enterprise")
* **Weaknesses**: most frequently cited negative attributes (e.g., "limited integrations", "steep learning curve")
Each item shows:
* How often it appears across responses
* Which LLM(s) surface it
* The prompt context it appeared in
This is the most actionable output from the Perception dashboard. If a weakness appears across many prompts and LLMs, it's a signal worth addressing in content.
***
## Competitor sentiment comparison
The competitor panel shows sentiment scores and perception counts for each tracked competitor, side by side with your brand.
For each competitor you see:
* **Sentiment score** (0–100)
* **Positive fact count**
* **Negative fact count**
This reveals whether you're losing on perception even when you appear alongside competitors. If a competitor consistently scores 20 points higher, inspect their extracted strengths. That's the narrative gap to close.
***
## Filtering
The Perception dashboard supports:
* **Time range**: filter to last 7, 14, or 30 days to spot trend changes after a content update or launch
* **LLM filter**: isolate a single model to understand model-specific behavior
***
## Related
How the sentiment score is defined and what drives it.
Which sources back your brand mentions.
Which domains shape how AI talks about you.