AI-powered automation

Important:

This feature is currently part of an audit process and is not to be considered part of the FedRAMP Authorization until the review is finalized. See here the full list of features currently under review.

This chapter lists guidelines to effectively use Autopilot^TM for test automation in Studio Desktop. Autopilot for testers can support test automation engineers in the following actions:

Converting manual test cases into automations.
Converting text into code.
Generating synthetic test data.

Note: To use AI-powered automation, install Studio Desktop version 2024.10.1 or higher.

Converting manual test cases into automation

After you link Studio to Test Manager, the from Studio Desktop offers you a view of all test cases within your Test Manager project. The panel displays both automated and manual tests. Autopilot allows you to convert the manual tests into automated tests.

Visit Transforming manual tests into coded test cases to learn how to convert manual tests into coded test cases.

1. Prepare a consistent object repository

Autopilot uses UI Automation capabilities to convert manual test steps into automated ones. In order to reference the UI elements that you want to test, Autopilot needs object repository references to each of them. It's crucial to maintain a consistent naming convention, ensuring that you use identical control names within your manual test steps as those within your object repository.

For instance, consider the following manual test step: "Click on 'Submit loan application'". In this case, Autopilot will generate a Click activity, and attempt to locate an object repository element named Submit loan application. Consistency for control names within your manual steps facilitates a smooth automation conversion process.

2. Use common activity names

A typical manual test step might have the following structure: "Type 'john.doe@gmail.com' into the 'Email' field". This step includes the following information used for automating it:

Action: Type, which is recognized as the Type Into activity.
Data: 'john.doe@gmail.com'.
Target: 'Email', which is recognized as the corresponding object repository element.

Check the table for the verbs you should use in your manual steps. These verbs are then converted in the appropriate activity/API within Studio Desktop.

Manual test step	UiPath activity	Uipath API
"'Click' on 'My button'"	Click	Click
"'Type' 'hello world' into 'My textfield'"	Type Into	TypeInto
"'Get Text' from 'My label'"	Get Text	GetText
"'Check' 'My checkbox'"	Check	Check
"'Select item' '3 years' from 'Loan term'"	Select Item	SelectItem

3. Automating form filling

If a form that required data input is displayed on your screen, you can use the 'Fill Form' keyword to tell Autopilot to automate it.

Note: This functionality is available only when generating a coded test case.

Consider the following manual test step as an example: "Fill the form on the screen named 'myFormScreen' with the following values: Email: 'john@doe.com', Loan Amount: '10000', Loan Term: '3'". Autopilot will execute the 'Fill form' command within your coded automation, populating the values across all identified user controls within your specified form.

Converting text into code

You can leverage the Generate Code action within a coded automation from the context menu, to convert any prompt that you input into code. Autopilot^TM can understand numerous instructions, and can help you:

Generate any C# code.
Refactor existing code.
Generate a UiPath® automation.

Visit Generating code to check the various methods you can use for generating code.

Prompt examples

"Refactor the method below and make it more performant."
"Create a method to compare 2 strings for similarity with the input parameters 'string1' and 'string2'."
"Click on 'Products'".
"Run Job 'myDemoJob' on Orchestrator."

Generating synthetic test data

Test data management can consume up to 50% of your testing efforts. Autopilot^TM can help you save this time, by auto-generating synthetic test data for your test cases.

1. Create arguments for your data fields

When you generate your synthetic test data, Autopilot considers the existing arguments within your workflow, as well as the additional instructions provided in the prompt. With this information it will try to generate test data that leads to a high coverage within your test case.

To produce a robust set of test data, we recommend to use arguments wherever feasible, instead of relying solely on local variables, or specific string values. By using arguments, you allow Autopilot to generate test data that aligns more closely with the dynamic requirements of your workflow.

2. Provide instructions on your preferred data combination method

By default, Autopilot generates approximately ten data records, aiming to achieve high code or activity coverage within your test case. However, if your test strategy requires a specific algorithm, such as 'pairwise' or 'all combinations', include this instruction in your prompt.

For instance, consider the following example prompt: "Generate test data for every argument and combine the data fields pairwise."

3. Use the prompt instruction to customize your data set

You can also instruct Autopilot to perform various other tasks, such as generate a specific number of data records, or include additional data fields.

Consider the following example prompts: