StressGen

Documentation / Getting Started with AI Stress Testing

Getting Started with AI Stress Testing

Step-by-step guide to creating your first AI-powered stress scenario — account setup, scenario generation, and understanding results.

Table of Contents

Welcome to StressGen

StressGen generates stress testing scenarios rooted in six layers of context: regulatory guidance, live market data, market intelligence, statistical validation, historical precedent, and likelihood intelligence. Describe a hypothetical economic event in plain English, and the platform produces a complete, audit-ready set of risk factor shocks.

Every number in your scenario traces back to a source — a regulatory document, a market data feed, or a statistical model. This is not generic AI output. It is context-rooted stress testing designed for regulatory submission.

This guide walks you through creating your account, building your first scenario, and understanding the results.

Create Your Account

  1. 1Go to the Login page and sign up with your work email.
  2. 2Your organization administrator will assign you a role (see User Roles below).
  3. 3Once approved, log in and you'll land on the Dashboard.

Your First Scenario

Creating a scenario takes just a few steps:

  1. 1Navigate to Scenarios and click "New Scenario."
  2. 2Describe the economic event you want to model. For example: "The Federal Reserve raises interest rates by 100 basis points amid persistent inflation, triggering a broad equity sell-off."
  3. 3Choose which data sources to include (economic indicators, news, market data, etc.).
  4. 4Select the regulatory framework — DFAST, EBA, or PRA.
  5. 5Click "Generate Scenario" and wait for the AI to produce results (typically 1-3 minutes).
Tip
The more specific your description, the better the results. Include details about the trigger event, expected severity, and any particular sectors or regions you want emphasized.

Understanding Results

After generation completes, your scenario page shows:

  • Summary cards — Total number of shocks, confidence score, and a breakdown by category.
  • Shock table — Every affected risk factor with its category, current value, and projected change.
  • Visualizations — Charts showing shock distribution by asset class, severity, and direction.
  • AI narrative — A written explanation of why each shock was applied, grounded in the scenario description and supporting data.

User Roles

Your organization administrator assigns one of three roles:

  • Viewer — Read-only access to scenarios and results.
  • Analyst — Create, edit, and generate scenarios. Run stress tests and analyze results.
  • Admin — Full access including user management, scenario approval, and organization settings.

Feedback & Bug Reports

StressGen is under active development and your feedback shapes the product. If something feels off, confusing, or missing, we want to hear about it.

When reporting an issue, include the steps to reproduce it, what you expected to happen, what actually happened (including any error messages), and your browser and operating system. Send reports to support@stressgen.ai.

Tip
Screenshots or screen recordings are extremely helpful for reproducing visual issues.

What's Next

Now that you have the basics, explore the other sections of this documentation:

  • Dashboard — Your real-time market overview and activity hub
  • Scenarios — Deep dive into scenario creation, editing, and comparison
  • Portfolios — Track portfolio sensitivities with daily snapshots
  • Stress Tests — Compute P&L impact of scenarios on your portfolios
  • Risk Factors & Shocks — Learn about the different types of risk factors and how shocks are calculated
  • Data Sources — Configure which real-world data feeds into your scenarios
  • AI Assistant — Get answers and insights through conversational interaction