Mavera Documentation
  • Introduction to Mavera
    • About Mavera
    • Our mission and vision
  • Our Technology
    • Overview of Mavera's AI Ecosystem
    • What Makes us Different from ChatGPT
  • Our Frameworks
    • Ellie: The Orchestrator
    • Emma: The Adversarial System
    • Gremlins: Data Harvesters
    • Sprites: Data Annotators
    • Personas: Targeted AI Swarms
    • Heracles: Individual Customer Modeling
  • Key Concepts
    • What is a Persona?
    • Core Technology: Mavera's AI Personas
    • AI Personas vs. Traditional Personas
    • Parallels with Traditional Personas
    • How Our AI Personas Work
    • Determining AI Persona Sample Size
    • Understanding AI Swarms
    • Hard to Reach Audiences
    • The Role of Data Scraping and Annotation
    • Synthetic Data Generation
    • The Emotional Intelligence of Our AI in Marketing
  • Privacy and Ethics
    • Data Handling and Privacy Policies
    • Ethical AI Development and Usage
  • FAQs and Support
    • Frequently Asked Questions
    • Contact Support
    • Troubleshooting Guide
  • The AI Revolution in Marketing: Why You Need It
  • Benefits of Mavera's AI Personas
  • ⚒️Use Cases
    • Our Offerings Overview
    • Qualitative Customer Research and Insights
      • Qualitative Research: Example Output
    • Individual Customer Profiling and Segmentation
      • Customer Profiling: Example Output
    • Competitor Analysis and Market Research
      • Competitor Analysis: Example Output
    • Content Analysis and Sentiment Tracking
      • Content Analysis: Example Output
    • Keyword Research and Topic Discovery
      • Keyword Research: Example Output
    • Creative Ideation and Testing
      • Creative Ideation: Example Output
    • Predictive Analytics and Trend Forecasting
      • Predictive Analytics: Example Output
    • Personalized Content Creation and Targeting
      • Personalized Content: Example Output
    • Brand Perception and Reputation Management
      • Brand Perception: Example Output
    • Customer Journey Mapping and Optimization
      • Customer Journey: Example Output
    • Enhancing Existing Market Research
      • Enhancing Market Research: Example Output
    • Influencer Identification and Analysis
      • Influencer Identification: Example Output
    • Customer Churn Prediction and Prevention
      • Customer Churn Prediction: Example Output
    • Pricing Optimization and Elasticity Analysis
      • Pricing Optimization: Example Output
    • Product Feature Prioritization
      • Product Feature Prioritization: Example Output
    • Marketing Mix Modeling and Optimization
      • Marketing Mix Modeling: Example Output
    • Ad Creative Testing and Optimization
      • Ad Creative Testing: Example Output
  • Case Study: AI Persona vs. Deloitte Study
  • AI Search Engine Optimization
  • Handling 'Practical' Jobs: Mavera's Advanced Approach
  • Quality Assurance in AI Outputs: Volume-Driven
  • The State of AI in Marketing
  • Mavera's Unique Advantage
  • ROI of AI in Marketing
  • The 'Destination': Future of AI in Marketing
  • Getting Started with Mavera
  • Fast Food Questions
Powered by GitBook
On this page
  • Data Inputs
  • Process Visualization
  • Key Points:
  1. Key Concepts

How Our AI Personas Work

Data Inputs

👤 Human Queries 💬 AI Interactions 💾 Company Data 💾 Real-World Data

Process Visualization

User Input (Query, request, or interaction)
                  ↓
Parent AI (Understands query intent) → Write Proxy (Refactors query into tasks)
                  ↓
Child AI 1     Child AI 2     Child AI 3
(Executes       (Executes      (Executes
specific task)  specific task) specific task)
                  ↓
        Reward Models (Check accuracy, utility, etc.)
                  ↓
        Final Proxy (Synthesizes responses)
                  ↓
        User Output (Persona response or action)

Key Points:

  • User input initiates the AI Persona process

  • Parent AI understands intent and directs queries appropriately

  • Proxy AIs refactor queries and synthesize responses

  • Child AIs specialize in specific tasks for optimal performance

  • Reward models ensure high-quality, accurate outputs

  • Final output is delivered back to the user

  • System can process both human and AI inputs

  • Continuous integration of company and real-world data enhances accuracy

PreviousParallels with Traditional PersonasNextDetermining AI Persona Sample Size

Last updated 10 months ago