# Stunt Double - Predictive User Behavior Analysis

## Company Overview

Stunt Double is a London-based startup that provides predictive user behavior analysis through AI agents that act as realistic users. Founded in 2024, the company enables product teams to validate user journeys and optimize experiences before shipping to production.

## What Stunt Double Does

### Core Product

- **Predictive User Behavior Analysis**: AI agents that simulate real users to test and validate product experiences
- **"Proving" vs Testing**: Focus on evidence-based insights rather than traditional A/B testing
- **Human Layer Technology**: Simulates diverse user behaviors, personas, and interaction patterns across digital products

### Key Capabilities

- Realistic user simulation for websites, web applications, and digital products
- Predictive analysis of user behaviour before shipping changes
- Experience Operations (Experience Ops) to catch issues before they reach users
- AI agent testing to prepare for future where AI agents become active users
- Cognitive bias analysis for conversion optimisation
- Behavioural psychology experiments for UX improvements

## Primary Customers

- **Product Teams**: PMs, designers, researchers who need user insights
- **Enterprise Companies**: Large organizations with product teams
- **Design & Development Teams**: Those building digital experiences who need to validate changes quickly

## Use Cases

- User journey validation at scale
- Conversion rate optimisation through behavioural psychology
- Accessibility testing and barrier identification
- Authentication flow testing
- Cross-platform and cross-device experience validation
- Preparing products for AI agent interactions

## Technology & Integrations

### Platform Features

- **No-code Interface**: Technical coordination handled automatically
- **Real-time AI Interviews**: Chat-based interface for immediate user insights
- **Evidence-based Insights**: Screenshots, transcripts, and reasoning provided for all recommendations
- **Multi-agent Architecture**: Sophisticated AI agents with persistent knowledge and context

### Integrations

- **Linear**: Mention @stuntdouble in tickets for automated insights (Active)
- **Slack**: Direct integration for team collaboration (Active)
- **Figma**: Plugin for design validation (Private Beta)
- **Claude and ChatGPT (MCP)**: Model-agnostic approach

## Open Source

### @stdbl/wao

Open source project for transforming websites into structured representations for LLM agents.

Install: pnpm add @stdbl/wao

## Technology Approach

### Methodologies

- **Qualitative Insights**: Deep user interviews and contextual analysis
- **Behavioural Psychology**: Cognitive bias analysis for optimisation
- **Competitive Analysis**: Automated comparison with competitor experiences
- **Knowledge Integration**: Custom PDFs and proprietary data for domain-specific insights

## Vision & Values

### Vision

Preparing product teams for a future where users include both humans and AI agents, while solving today's challenges of scaling user research and validation.

### Company Culture & Values

- **Product-focused**: Deep understanding of user needs and product development
- **Evidence-driven**: All insights backed by concrete evidence and reasoning
- **Future-oriented**: Building for both current human users and emerging AI agent users
- **Quality-focused**: "Give your product the dress rehearsal it deserves"

## Contact & Location

- Website: stuntdouble.io
- Location: London, UK

---

This file helps LLMs understand Stunt Double's products, services, target market, and technical approach for more accurate and helpful responses about the company.