Brand Perception

When someone asks an AI model about your category, it doesn’t just recommend brands, it also describes them. AI might say things like “great for design teams” or “strong on security, but weaker on integrations.” For many users, that description becomes their first impression, and they rarely see the sources behind it. Peec gives you two connected views for this:

• Brand Perception breaks AI’s view of your brand into attributes, the qualities models associate with brands in your space.
• Brand Insights zooms into a single brand and shows how it performs across models, topics, geographies, and competitors.

​

Brand Perception

Peec breaks AI’s view of your brand into attributes, the qualities models associate with brands in your industry (for example reliability, performance, value for money, or ease of implementation). Attributes aren’t a fixed checklist. Peec extracts them from real AI answers and clusters similar ones together, so the set reflects how your space is actually described. For every attribute you can see how strongly it’s associated with your brand and how that compares with your tracked competitors. Peec asks AI two different questions, and the page keeps them apart because they answer different things:

• Asked about your brand: “What is this brand known for?” The answer gives each attribute an association score from 0 to 100, where 100 means the attribute always comes up first when AI describes you.
• Asked about your industry: “Which brands are known for this attribute?” The answer gives every brand a market prominence score from 0 to 100, where 100 means the brand is always named first, and 0 when it is never named.

Both run 0 to 100, but they are not the same number and are not comparable. A brand can be described as “fast” constantly and still lose “fast” to every competitor in the market. That difference is usually the most useful thing on this page. Scores are averaged across the AI models you track. Use the model filter at the top of the page to see how perception shifts on a single model.

​

Summary

The headline states the finding in a sentence, for example “AI describes Gymshark as Fitness, but it competes best on Influencer”. Above it, a timestamp shows when the last run completed; hover it for the estimated next run. Below that:

• Industry: The industry your competitors are drawn from. Keep it specific, because it decides which brands you are measured against.
• Most associated: The attribute AI reaches for most when asked about your brand. This is the top bar in the chart below it.
• Best vs competitors: The attribute where you come closest to the brand leading it, with your position next to it. Position alone isn’t the ranking criterion, because attributes attract different numbers of brands, so a #3 of 12 and a #3 of 60 aren’t the same achievement. Hover for how many attributes you place top three on.
• Biggest gap: The attribute AI leans on harder than your market position earns: high when AI is asked about you, low when it is asked about your industry. Shown as two positions, for example #1 → #9. unplaced means you never surface for it in the market at all, which is the widest version of this gap.
• Strongest competitor: The competitor with the highest average score across all attributes. The brand AI describes most favorably overall in your space.

​

How AI describes your brand

Association scores: the attributes AI reaches for when it is asked about your brand, strongest first. If the messaging you’re investing in doesn’t show up here, AI hasn’t picked it up yet, and the chats AI cites are where to start looking. These scores answer the “asked about your brand” question, so they don’t line up with the breakdown below. An attribute can top this chart and still sit near the bottom of the market.

​

Brand shape

A radar chart of the same market prominence scores as the breakdown below, one axis per attribute, so you can see the shape of your position rather than read it row by row. Each axis is scaled to the brand leading that attribute, so the leader always reaches the edge. The chart shows three brands by default, and you can add more from the brand selector. You can also choose which attributes appear on the axes.

​

Brand comparison by attribute

A heat map of market prominence for every brand and attribute, from the “asked about your industry” question. Each cell is a brand’s 0 to 100 score, so you can scan across brands and down attributes to see who owns what. Answers a brand never surfaced in count against it, so a brand that shows up rarely scores lower than one named in every answer.

​

Attributes and sources

Every attribute with your position for it, and the sources feeding that position. Ranking is where you place among every brand scored for the attribute, by market prominence. Answers you never surfaced in count against you, the same way the breakdown scores them, so this agrees with the heat map above. Brands tied on score share the better position. A dash means you never surfaced for that attribute at all. Expand any attribute to see the actual sources AI retrieved when describing brands on that quality, with links through to the full source detail. This is where a position turns into a concrete list of pages to influence. For each source you get:

• Occurrences: How many times AI cited this source when associating your brand with the attribute, across all Brand Perception checks. Unlike retrievals, this isn’t limited by the dashboard date range or model filter, so it’s the truest measure of which sources shape the attribute. Sources are sorted by occurrences by default.
• Retrievals: The number of distinct chats in your tracked prompts where the source was retrieved, over the selected date range.
• Citation rate: The average number of citations per chat the source was retrieved in.
• URL type and Domain type: How the source page and its domain are classified.

​

Best practices

• Start with the biggest gap. It names the attribute AI credits you with more than the market does, which is usually more actionable than confirming a strength you already expected.
• Then read what you’re not known for. An attribute where you place well in the market but AI rarely mentions when asked about you is demand you have already earned and aren’t collecting.
• Check individual models. A brand can be perceived very differently in ChatGPT than in Gemini, so aggregated results can hide important differences. Use the model filter for the engines that matter most in your industry.
• Follow attributes to their sources. Use Attributes and sources to turn a low score into the specific pages worth influencing.

​

Brand Insights

Brand Insights is a deep-dive hub for analyzing how a single brand performs across topics, models, geographies, and competitors. Pick a brand and see where it dominates and how it compares with others across every dimension you track. The header carries a full filter bar (date range, model, topic, tag, and country, plus reset). The brand filter sets which brand you’re analyzing, and can be pointed at a competitor as well as your own brand. At the top, a summary line shows your rank (“You’re #N in AI Visibility”), who leads, and when the data last updated.

​

Main KPIs

A summary of the selected brand’s performance over your chosen time period:

• Visibility: Percentage of AI responses where this brand appears.
• Share of Voice: This brand’s share of all brand mentions across your tracked prompts.
• Sentiment: How positively AI platforms describe this brand, on a 0 to 100 scale.
• Position: Average ranking when this brand appears in AI responses (lower is better).
• Strongest / Weakest model: The AI model where the brand performs best or worst, based on combined visibility, sentiment, and position.

Change indicators show how each metric moved compared with the previous period of the same length.

​

Brand insights graph

The graph shows the brand’s performance over time. Toggle between Visibility, Sentiment, Position, and Share of Voice to see trends for each. It also overlays your retrieved percentage, so you can see how often your own domains are pulled into AI answers alongside the selected metric. Choose which of your domains to display, and compare brand authority against domain authority over time. Vertical markers on the chart flag events like new prompts being created or an AI model change.

​

Performance matrix

A configurable matrix that crosses any two dimensions against each other. You control both axes, choosing from:

• Models (ChatGPT, Claude, Perplexity, etc.)
• Brands (competitors)
• Topics
• Tags
• Geographies

Pick which metric to view (Visibility, Sentiment, Position, or Share of Voice) with the selector in the top-right. Cells are heat-mapped by score, each cell’s tooltip shows all four metrics, and you can reorder columns by dragging. For each axis you can pick exactly which items to show (up to 10), search them, or apply a top-N preset. For example:

• Which topics does the brand own on ChatGPT but not Perplexity?
• Where is a competitor outperforming my brand, and in which specific markets?

​

Top rankings

Shows how brands rank against each other for a dimension you choose (by AI model, topic, tag, or geography). Each column is a rank position, from #1 through the top brands, and each cell holds the brand sitting at that rank, with your own brand highlighted. Use the metric control to choose whether ranks are ordered by Visibility, Share of Voice, Sentiment, or Position.