> ## Documentation Index > Fetch the complete documentation index at: https://docs.qwairy.co/llms.txt > Use this file to discover all available pages before exploring further. # What does AI actually say about my brand? > Read verbatim AI-generated responses about your brand, then surface the recurring themes, narrative patterns, and sentiment signals that define your reputation in generative AI. **TL;DR** — AI builds a narrative about your brand every time someone asks, and you need to read it verbatim before you can manage it. Go to **Monitor > Response Analysis** to read the exact paragraphs AI generates, then cross-reference with **Analyze > Perception** to see which themes dominate across all responses. Check the **Sentiment Analysis** overlay to distinguish between being mentioned frequently and being mentioned favorably. Pro tip: a high visibility score only means AI talks about you a lot — always read the actual response text to know whether that attention is an asset or a liability. ## The Question > **"What does AI actually say about my brand when someone asks about it?"** When a prospect asks ChatGPT, Claude, or Perplexity about your brand — whether to compare options, evaluate a purchase, or understand your positioning — they receive a generated paragraph, not a search result. That paragraph shapes perception before your website is ever visited. Knowing exactly what it says, and why, is the starting point for any AI brand strategy. This question covers the full content of those responses: the claims made, the attributes highlighted, the tone used, and the recurring narrative that emerges across models and prompts. **You might also be wondering:** * "Is what AI says about me accurate, or is it distorted by outdated training data?" * "Which attributes does AI emphasize most when describing my brand?" * "Do all AI providers describe my brand the same way, or do they diverge?" *** ## Where to Go in Qwairy Navigate to **Monitor > Response Analysis** — this is where you read the actual text AI generated. Focus on the **response body** for each answer: read the full paragraph, not just the score. Use the **Provider** filter to compare what ChatGPT says versus what Claude or Perplexity says for the same prompt. Cross-reference with **Analyze > Perception** to move from individual responses to aggregated signals. Use the **Theme extraction** panel to identify which attributes appear most frequently across all responses. Use the **Period** filter to see whether the narrative has shifted recently. Open **Analyze > Sentiment** to overlay emotional tone on the content you just read. A response can mention your brand frequently but in a cautious or comparative framing — Sentiment Analysis quantifies that distinction. Use **Workspace > Shared Views** to create a read-only perception snapshot for stakeholders. Lock the filters to a specific provider set and period before sharing. *** ## What to Look For ### Responses — Verbatim AI Output The Responses page lists every AI-generated answer collected during your monitoring runs. Each row shows the prompt, the provider, the date, the visibility score, and the full response text. | Element | What it tells you | | -------------------- | ---------------------------------------------------------------------------- | | **Response body** | The exact narrative AI constructs about your brand | | **Visibility score** | How prominently your brand appears in the answer (0–100) | | **Provider badge** | Which AI model generated this specific answer | | **Prompt context** | The question that triggered this response — useful for understanding framing | | **Cited sources** | The URLs AI referenced to construct its answer | Read several responses before drawing conclusions. Look for sentences that repeat across providers — those are the stable claims your brand has "earned" in training data or retrieval. Look for sentences that contradict each other across providers — those signal inconsistency worth addressing. ### Brand Perception — Aggregated Narrative Brand Perception processes all collected responses and extracts structured signals: dominant themes, attribute clusters, and a narrative summary that describes how AI collectively portrays your brand. | Element | What it tells you | | ----------------------- | ------------------------------------------------------------------------------ | | **Theme cloud** | Which topics and attributes appear most often in AI descriptions of your brand | | **Narrative summary** | A synthesized paragraph of what AI "believes" about your brand | | **Attribute strength** | How confidently AI associates each attribute with your brand | | **Comparison baseline** | How your narrative compares to competitors in the same space | > **Pro Tip**: Combine the verbatim **Responses** view with the **Brand Perception** theme summary. If a theme appears in Brand Perception but you cannot find it in actual responses, check the date range — it may reflect older data that has since shifted. ### Filters That Help | Filter | How to use it for this question | | --------------- | ---------------------------------------------------------------------------------------- | | **Provider** | Isolate a single AI model to understand its specific narrative about you | | **Period** | Compare what AI said last quarter versus this quarter to detect drift | | **Topic / Tag** | Narrow to a product line, market segment, or use case to get targeted narrative analysis | | **Prompt** | Read all responses generated by a specific monitoring prompt to see consistency | *** ## How to Interpret the Results ### Good result AI responses consistently describe your brand using the attributes you intend to own: your category leadership, key differentiators, or core value proposition. The narrative summary in Brand Perception closely matches your actual positioning. Sentiment is predominantly positive or neutral-informative. Across 3+ providers, the descriptions are similar in substance, even if phrased differently. ### Needs attention AI introduces attributes you do not claim, omits your primary differentiators, or describes your brand in a category you are trying to exit. Responses contain factual errors (wrong pricing, discontinued products, obsolete team details). Different providers give contradictory accounts — one positions you as enterprise-focused while another describes you as a startup tool. A high visibility score does not mean AI is saying good things. Your brand can appear in 80% of responses while being framed as a secondary option or a cautionary example. Always read the response text alongside the score — the score measures presence, not quality of portrayal. *** ## Example > **Scenario**: A luxury watch brand notices that AI responses consistently describe them as a "mid-range fashion brand" rather than a heritage watchmaker. Their recent repositioning toward haute horlogerie and limited-edition craftsmanship is not reflected in how ChatGPT, Claude, or Perplexity describe them. They want to understand the full narrative before deciding how to intervene. 1. Open **Monitor > Response Analysis**, filter by Provider = "All" and Period = "Last 90 days". Read 10–15 responses containing the brand name. Note the recurring descriptors: "fashion-forward", "accessible luxury", "popular among millennials" — none of which reflect their heritage positioning. 2. Open **Analyze > Perception** and review the Theme extraction panel. Confirm that "fashion accessory" and "affordable luxury" are the dominant clusters. Note that "Swiss craftsmanship", "in-house movement", and "limited edition" — their core positioning pillars — do not appear. 3. Open **Analyze > Sentiment** and confirm that while sentiment is positive, the positive signals cluster around "style" and "trend" rather than "exclusivity" or "horological tradition". 4. Export the Brand Perception summary and bring it to the brand strategy team as evidence that AI is reinforcing the old mass-market narrative rather than the new heritage positioning, informing a content push toward watchmaking authority publications and collector communities. *** ## Go Further Export AI-generated responses and brand perception data as CSV for stakeholder presentations Build a brand narrative tracking dashboard in Looker Studio using the answer-details data source Create a shared view of your brand perception data for leadership or investors to review ## Related Questions Compare what AI says versus what you intend to communicate — and measure the gap. Extract a structured SWOT picture from AI-generated brand descriptions. Detect where ChatGPT, Claude, and Perplexity diverge in how they portray you. Shift from what AI says to whether AI mentions you at all.