---
updatedAt: 2026-03-09T11:31:56.000Z
agentTools:
projectIndex: https://help.rainforestqa.com/llms.txt
---
# AI-Powered Test Creation
## Overview
Rainforest gives you three ways to create tests with AI, each one designed to reduce the time you spend writing and maintaining test steps manually.
* **AI Test Generation** takes a plain-English prompt and turns it into a full draft test. Give it a goal like "add a new user and assign them the admin role" and Rainforest spins up a virtual machine, navigates your app, and writes the steps for you. You review and refine the output in the Visual Editor before running it.
* **Self-Healing Snippets** are blocks of steps generated from a plain-English prompt that live inside your test as a snippet. If any step in that block fails during a run, Rainforest automatically tries to regenerate the steps and self-heal. If it succeeds, your test is updated automatically with no manual fix required.
* **AI Actions** are individual step types you can add anywhere in a test. AI Assertion answers a yes/no question based on what's currently visible on screen, useful for dynamic content that's hard to screenshot-match. AI Search takes a plain-English description of an element, finds it on screen, and performs whatever action you've selected, acting as a fallback when standard targeting isn't reliable enough.
## When to use AI generation
AI generation works best when:
* You're testing a workflow you can describe in plain text
* You want a starting point quickly and plan to review and refine the generated steps
* You're creating multiple tests in parallel and don't want to write each one from scratch
It works less well when:
* Your test requires complex conditional logic, since the generator writes steps for what it sees on screen and can't anticipate the counterfactual scenario
* Your workflow has app-specific behavior that isn't obvious from the UI, in which case add context to your site rules or prompt before generating
*If you have any questions, reach out to us at .*