Quick Answer
Personas in Scrunch help you generate prompts and inject audience context into AI responses during data collection. Persona injection is now used across supported platforms, but the amount of persona context varies by platform, which means the same prompt can produce different results depending on where it is collected.
How Personas Work in Scrunch
Auto-Generated Customer Profiles
When you set up a brand, Scrunch automatically generates up to three personas by analyzing your website to understand who your company targets. These personas are:
Fully editable, you can refine descriptions, archive, or add new ones
Optional, not every prompt needs a persona
Strategic, best used when you need audience-specific insights
The Two Main Functions
1. Prompt Generation with AI
When you click "Generate Prompts with AI" (or during initial setup), persona descriptions help create realistic questions that each audience type would ask. The persona context helps your monitoring reflect actual customer behavior.
2. Persona Context Injection During Collection
Scrunch injects persona context into prompts at collection time so AI models respond as if they are answering that specific audience. This means the same question can return different answers for different personas.
How it works by platform:
ChatGPT, Claude, and Grok: Scrunch injects the persona name and description
Perplexity, Google AI Mode, Google AI Overviews, Gemini, Copilot, and Meta: Scrunch injects the persona name only
Some providers respond more strongly to persona context than others. Because of that, persona-based differences may be subtle on some platforms and more pronounced on others.
Persona injection can affect your results. In some cases, citation metrics may decrease after persona context is applied because providers handle persona context differently and may generate different responses or cite different sources.
Availability:
Persona injection is rolling out across platforms and applies to future collections. Historical data collected before persona injection changes remains unchanged.
3. Data Segmentation & Filtering
Filter your dashboard, prompts table, and reports by persona to analyze:
Which audience segments are searching for what
Performance differences across customer types
How AI responses differ for different user roles or experience levels
Customer journey stage analysis
Learn more about filtering and segmentation in Understanding Filters in Scrunch and Understanding the Explorer Tab.
Setup & Customization
Creating & Managing Personas
Initial Setup:
Scrunch auto-generates 3 personas during brand setup
Based on website analysis and target audience detection
You can edit, archive, or replace these immediately
Adding New Personas:
No limit to how many you can create
Write descriptions that help with AI prompt generation and persona context injection
Each persona can be assigned to multiple prompts
Assigning Personas to Prompts:
During manual prompt creation
During bulk CSV upload (include a persona column with the persona name)
Via "Generate Prompts with AI" feature
Prompts can be created without personas (generic monitoring)
Archiving personas:
If you archive a persona that is currently associated with a prompt, that prompt is paused. Review persona assignments before archiving so you do not unintentionally stop collection on active prompts.
For step-by-step instructions on adding and organizing prompts, see How to Add Prompts in Scrunch and Uploading Prompts via CSV.
Writing Effective Persona Descriptions
On platforms that support full persona injection, your persona descriptions directly influence how AI models respond. On platforms that use persona name only, the description still helps with prompt generation and internal organization, but it will have less direct impact on collected responses.
Write descriptions that:
Define the persona's role, responsibilities, and context (for example, "Small business owner managing a team of 5-10 employees")
Highlight their pain points and priorities (for example, "Limited budget, needs easy-to-use tools")
Specify their level of expertise or familiarity with your product category (for example, "First-time CRM buyer, not technically savvy")
Include demographic or firmographic details when relevant (for example, "Works in retail industry")
Advanced Tip: You can use XML formatting within persona descriptions to influence prompt generation precision and recall, though this is rarely needed for most use cases.
When to Use Personas
Use Personas When:
You need audience-specific insights
"Do enterprise buyers see us differently than small businesses?"
"Are we showing up for technical vs. non-technical searchers?"
"How do responses differ for different user roles or experience levels?"
You need customer journey segmentation
Early-stage awareness questions vs. evaluation-stage questions
Different personas at different buying stages
Role-based visibility (end users vs. decision makers)
You want to test personalized AI responses
Compare how AI responds to the same question from different audience types
Identify when AI oversimplifies or overcomplicates based on persona characteristics
Discover audience-specific objections or concerns in AI responses
Skip Personas When:
You're tracking broad brand visibility across all audiences
You're monitoring generic industry terms or competitor comparisons
You don't need to segment data by audience type
You're working with limited prompt budgets and want to maximize coverage
Strategic Use Cases
Better Reporting & User-Group Insights
Instead of seeing "40% brand presence across all prompts," you can segment:
Small Business Owner persona: 55% presence
Enterprise IT Director persona: 32% presence
Individual Consumer persona: 48% presence
This reveals which audiences are best-served by current AI visibility.
Journey Stage Analysis
Create personas representing different stages:
"Early-Stage Researcher" persona → Awareness-stage prompts
"Active Evaluator" persona → Comparison and advice prompts
"Ready to Buy" persona → Specific product or pricing prompts
Then analyze where you're strongest and where you need content improvements.
Personalized Response Testing
With persona injection, you can test how AI responds to different audience types asking the same question:
Compare technical vs. non-technical explanations of your product
Identify when AI oversimplifies or overcomplicates based on persona
Discover audience-specific objections or concerns that surface in responses
Optimize content for specific personas based on how AI adapts its answers
Keep platform differences in mind when reviewing results. A strong persona effect on ChatGPT, Claude, or Grok may look smaller on platforms where only the persona name is injected.
Frequently Asked Questions
Q: Do personas personalize AI responses during data collection?
A: Yes. Scrunch injects persona context during collection, but the exact behavior depends on the platform. ChatGPT, Claude, and Grok receive the persona name and description, while Perplexity, Google AI Mode, Google AI Overviews, Gemini, Copilot, and Meta receive the persona name only.
Q: Why do persona-based results look different across platforms?
A: Because providers handle persona context differently. Some platforms receive the full persona name and description, while others receive only the persona name, so the same prompt may produce different response patterns.
Q: Do I have to use personas?
A: No. Personas are optional and strategic. Many customers successfully monitor AI visibility using generic prompts without personas. Use them when you need audience segmentation or personalized response analysis.
Q: How many personas should I create?
A: Start with 3-5 that represent your core audience segments. More is not always better, focus on personas that will actually inform your strategy and reporting.
Q: Can I bulk upload prompts with persona assignments?
A: Yes, CSV uploads support persona assignment. Include a persona column in your CSV with the persona name. See Uploading Prompts via CSV for details.
Q: What happens if I archive a persona that is attached to prompts?
A: Any prompt associated with that archived persona is paused. If you want collection to continue, reassign those prompts before archiving the persona.
Q: Will persona changes affect my historical data?
A: No. Persona injection changes only apply to future collections. Historical responses collected before the change remain unchanged.
Q: Can persona injection affect citation metrics?
A: Yes. Because providers may change how they answer and what they cite when persona context is added, you may see changes, including possible drops, in citation metrics.
Q: How do I control which country my prompts are collected from?
A: Geographic targeting is set at the brand level (Primary Location) with optional prompt-level overrides. Personas do not control geographic targeting. See Understanding Geographies in Scrunch for details.