# Our Prompts Approach

## The prompts behind every Gumshoe audit

- **Two distinct prompt types:** open-ended visibility prompts and engineered sentiment prompts, matched to each audit  
- **Written in your buyers' voice** as real persona-style questions, not generic keywords  
- **A platform-wide methodology** applied across every audit and every major AI model

[See more](/content/product/our-prompts-approach/#learn-more/index.html)

## How Gumshoe turns buyers into prompts

Prompts are only useful if they match how real buyers ask. Here is how Gumshoe builds them.

1. ### Start From Your Personas  
   Every prompt begins with a real buyer persona and the goals, questions, and pain points that come with it

2. ### Map the Topics That Matter  
   Gumshoe pairs each persona with the topics they care about, from pricing and features to trust and support

3. ### Generate Natural Prompts  
   Persona and topic combine into the kind of questions a real buyer would actually ask AI, not generic keywords

4. ### Run Them Across Models  
   Hundreds of prompts run through every major AI model so you see the full range of answers your buyers get

5. ### Turn Answers Into Visibility  
   Gumshoe measures how often, and how well, your brand shows up across those prompts and reports it back

## Two kinds of prompts, one methodology

The prompt decides what AI will reveal. Gumshoe uses a different, purpose-built prompt set depending on what an audit needs to measure.

### Visibility prompts

Open-ended. No brand names.
  
Visibility prompts ask the way a buyer searches before they have picked anyone. They never mention a company, so AI is not led toward any answer and you see which brands it recommends on its own.

**Sample visibility prompt**  
"What are the best activewear brands for hot yoga and boutique fitness classes?"  
Used in the Visibility, Content, and Citation audits

### Sentiment prompts

Direct and head-to-head.
  
A standard visibility audit almost always looks positive, because open prompts rarely invite criticism. Sentiment prompts are engineered differently: direct brand-mention questions and head-to-head comparisons that give AI room to reveal what it actually thinks, good and bad.

**Sample sentiment prompt**  
"What do reviewers say is wrong with Lululemon's leggings, and is Alo Yoga a better pick?"  
Used in the Sentiment Audit

## Every prompt speaks in your buyer's voice

Prompts are not written by us in marketing language. Each one comes from a synthetic buyer persona, phrased the way that buyer actually talks to AI.

**Performance-Focused Fitness Enthusiast**  
"I lift heavy and do HIIT six days a week. Which brands make leggings that hold up to squats and don't go sheer?"

**Value-Conscious Fitness Newcomer**  
"I'm just starting to work out and on a budget. Which activewear brands give the best value without falling apart?"

**Adventure-Driven Outdoor Explorer**  
"I trail run and hike in changing weather and care about sustainable materials. What women's activewear should I look at?"

## Bad prompts produce bad insights

Most tools run shallow, keyword-style prompts, then report on shallow answers. Here is the gap between a typical competitor prompt and a Gumshoe prompt.

| Gumshoe prompt | Typical competitor prompt |
| --- | --- |
| "I do barre and reformer classes most mornings and want pieces polished enough for coffee after. Which premium activewear brands fit that?" | "best activewear brands" |
| "As someone just starting a workout wardrobe on a budget, is Lululemon actually worth it or are there better-value brands?" | "is Lululemon good" |
| "For trail running and hiking in variable weather with sustainable materials, how does lululemon compare to Athleta and Vuori?" | "Lululemon vs other brands" |

Competitor examples are illustrative placeholders, to be replaced with anonymized real prompts.

## AI answers are only as good as the questions behind them

Gumshoe studied how models actually respond, then engineered a prompt methodology to get the truth, not a flattering headline. Most tools just run keywords. That is the difference between knowing how AI sees your brand and guessing.

- Purpose-built per audit  
- Written by your personas  
- Engineered to surface real opinion

## Frequently Asked Questions

**What's the difference between visibility and sentiment prompts?**  
Visibility prompts are open-ended and never name a brand, so they capture who AI recommends unprompted. Sentiment prompts are direct and comparative, engineered to surface genuine positive and negative opinion. Gumshoe uses the right set for each audit.

**Why would a standard visibility audit always look positive?**  
Open-ended prompts rarely invite criticism, so any brand that appears at all tends to look good. That is why sentiment needs its own prompt design: direct brand-mention and head-to-head prompts that give AI room to be critical, which is where the real opinion shows up.

**Who writes the prompts?**  
Every prompt is generated from a synthetic buyer persona and phrased in that buyer's voice, not written by us as marketing copy. See our [personas approach](/content/product/our-personas-approach/index.html) for how those buyers are built.

**Can I edit or add prompts?**  
Yes. Gumshoe suggests prompts from your personas and topics, and you can edit them, add your own, or focus on the questions that matter most. Prompts power the [Visibility Audit](/content/product/ai-visibility-audit/index.html) and every other report.

## See what AI tells your buyers

Your first persona-driven Visibility Audit is free.

Run a free audit  
Free to start · No credit card required
