※ Please note that parts of this page are machine translated.
Overview
Autify AI Agent, our AI-powered assistant, streamlines the entire quality assurance process. Whether you’re writing new scenarios, troubleshooting errors, analyzing PRDs, or understanding coverage, the AI Agent is designed to boost your productivity and confidence.
Based on your experience with Autify or your specific use cases, we advise reading the following sections related to your needs to get the most benefit, rather than reading everything.
If you are:
heavily troubled by unstable locator definitions, read the “AI agent” section.
spending a lot of time checking scenario execution errors due to frequent site update, read the “AI fix” section.
testing Salesforce, read the “AI agent” section.
What can the AI Agent do?
Our agent offers four powerful capabilities:
Capability | What it does |
|---|---|
Scenario Creation | Instantly creates test scenarios from natural language instructions |
Step Auto-Fix | Detects and fixes errors in test steps, suggesting code or logic corrections automatically |
Test Case Generation | Parses your uploaded product requirements document (PRD) and generates relevant test cases |
Coverage Report Generation | Analyzes uploaded test cases to produce clear, actionable coverage reports |
Getting started
Onboarding checklist
Currently, AI features are only supported in remote workspaces.
Available features may differ based on your contract. Please reach out to your sales representative for details. </aside>
Log in and navigate to a remote workspace.
Find feature buttons from these locations:
AI agent for scenario creation: Hover menu bar during recording
AI Fix: Each execution result detail page in Result tab (the button will show only when a test hits error)
Generate Test Cases & Generate Coverage Report*: Scenarios tab (Labeled “Generate with AI”)*
Follow the step-by-step prompt or wizard for each task.
How to use each feature effectively
A. Scenario creation with AI Agent
When:
You will appreciate this feature when you…
Know the overall scenario and requirements, but not the details
The AI Agent can fill in the missing steps to complete your scenario.
Are not familiar with the page structure of the site under test
The AI Agent can analyze the site and guide you through the process.
Are unsure how to perform a specific step during recording
For example, if you don’t know the correct action to take—or whether it’s even possible—the AI Agent can propose the best approach.
Need to create more advanced code, such as custom locators
The AI Agent can generate and suggest code for complex scenarios.
How:
Start recording.
Click “AI agent” in the hover bar, and enter a description of the test you’d like to run (for example, “click the cheapest product on the product page, add it to the cart, and verify that the item is correctly added to the cart.”).
The AI Agent will automatically generate and record the scenario based on your input.
Tips:
Be explicit about workflows, data types, or boundary cases.
Besides a whole scenario, AI agent can be used to create a single step.
Consider what the AI creates to be about 80% complete. Reviewing and making manual adjustments based on the AI’s output is highly recommended.
B. Step auto-fix
When:
You will appreciate this feature when you are facing…
Frequent site changes
When the website under test changes outside your visibility or without prior notice, AI Fix can adapt scenarios to keep your tests running.
Complex locators
If your scenario relies on complicated or unstable locators, AI Fix can help simplify and adjust them for more reliable execution.
Unfamiliar site structure
When you’re testing a website whose structure or behaviour you’re not fully familiar with, AI Fix can guide and apply fixes for you.
How:
Open your test results.
Navigate to the step where the error occurred.
Click the “Fix with AI” button to launch the AI Agent.
The AI Agent will analyze the error, propose a fix, and apply it for you.
Tips:
The “Fix with AI” button may not appear if the AI cannot provide a suitable fix.
Review suggested changes before approval.
If the fix misses your expectation, provide direct feedback to improve future results.
C. Test case generation from PRD
When:
You will appreciate this feature when you…
Are unsure if your test cases are comprehensive.
The AI can help ensure your tests cover all necessary scenarios.
Want to create test cases optimized for Autify.
The AI can generate test cases that are best suited for the Autify platform.
Need a baseline set of test cases to refine later.
You can quickly generate an initial set of tests and then manually edit them for more detailed and specific scenarios.
How:
Navigate to Scenario window and click Generate with AI. From the drop-down menu, select Generate Test Cases.
In the AI Agent window, upload the specification document for the feature you want to create test cases for by clicking the file icon at the bottom.
You can also directly type the requirements into the text area in addition.
To guide the AI, you can provide specific instructions. For example: "Based on the attached specification, generate only the happy-path test cases for the user stories."
Click Send (paper airplane icon). AI Agent will generate the test cases and save as a CSV file in your downloads folder.
The time it takes can vary depending on the file size.
Open the downloaded CSV file to review the generated test cases.
Try to interact with AI to get better answer if the generated cases don't meet your requirements or if you need additional cases.
Each time you regenerate, a new CSV file will be downloaded.
Tips:
Ensure your PRD is clear and well-structured for best results.
Ask for coverage on specific modules or functionalities (“Test payment gateway section comprehensively”).
Maximizing your success
Workspace Level Custom Prompt Configuration:
You may want to create and run tests that are specific to your organization’s unique context. By providing this context to the AI in advance, the AI Agent can generate scenarios that are better optimized for your organization’s needs.
This ensures that the generated logic aligns more closely with your workflows, requirements, and testing goals.
Prompt Engineering:
Use specific, goal-oriented prompts for each feature. For complex scenarios, break requests into parts.
Experiment & Iterate:
Try variations, refine outputs, and compare automated versus manual results.
Review & Edit:
AI suggestions are editable— use the output as a baseline, and customize to match your team’s standards.
Responsible, secure, and ethical use
Do not upload sensitive production data.
For Test Case Generation & Coverage Report Generation: Review AI-generated outputs for relevance, accuracy, and bias.
Use the feedback system to report issues or suggest improvements.
Frequently Asked Questions
Can the agent handle large/complex PRDs?
Yes, but the best results come from well-organized and clearly written documents.
Do these AI features learn from my feedback?
No, for now Autify AI doesn’t learn from user input.
What file types are supported for upload?
Currently: PDF, CSV, and Markdown.