À propos d'Autopilot for Testers

Autopilot for Testers is a collection of AI-powered digital systems, also known as agents, designed to boost the productivity of testers throughout the entire testing lifecycle. These capabilities are integrated into UiPath® Studio Desktop and UiPath® Test Manager, two key components of UiPath® Test Suite^TM.

Autopilot for Testers offers capabilities that can be grouped into the following categories:

• Agentic test design: Autopilot^TM in Test Manager supports you to evaluate requirements for quality aspects such as clarity, completeness, and consistency. Autopilot also helps you generate manual test cases for requirements (such as user stories) and SAP transactions.
• Agentic test automation: Autopilot^TM in Studio Desktop supports you to convert text such as manual test cases into coded and low-code automated UI and API test cases, generate test data for data-driven testing, refactor coded test automation, fix validation errors, generate expressions, and perform fuzzy verifications, among others.
• Agentic test management: Autopilot^TM in Test Manager allows you to get actionable insights into test results through the Test insights report.

The following sections what Autopilot^TM can help you with, when you create testing projects.

Quality-check requirements

Start by creating a requirement, such as Submit loan request for the UiBank Application. Describe the application flow and criteria needed for the loan application process. Then select Evaluate quality and evaluate the requirement using Autopilot^TM, to generate a list of suggestions that you can directly implement.

Provide supporting documents to Autopilot^TM, and additional guidance through a prompt that you choose from your library or that you type in yourself.

Après avoir déclenché l’évaluation, développez chaque suggestion pour mettre à jour son statut en fonction de la progression de votre implémentation. Vous pouvez choisir d’ajouter la suggestion à vos exigences, en indiquant son statut comme En cours ou Terminé. Une option permettant de supprimer les suggestions est également disponible.

Générez davantage de suggestions ou régénérez-les avec différents documents support ou des conseils supplémentaires à l’aide des options Plus de suggestions et Régénérer.

Vous pouvez également conserver les suggestions pour référence future en les exportant vers Word.

Visit Quality-check requirements and AI-powered evaluation: Best practices to understand how to efficiently evaluate your requirements using Autopilot.

Generate tests for requirements

Use Autopilot^TM to generate a list of potential test cases. You can generate test cases from the requirement details, followed by uploaded documents, and additional instructions.

Ouvrez une exigence et sélectionnez Générer des tests afin de générer des tests pour cette exigence.

Refine the generation process with documents and instructions for Autopilot^TM to use in generating the test cases. After you select Generate tests, review the generated test cases and create tests if satisfied, or refine them with more details if otherwise.

Visit Generate tests for requirements and AI-powered generation: Best practices, to check how to use the test generation feature at its full potential.

Generate tests for SAP transactions

With Autopilot^TM, you can generate test cases for SAP transactions from the Heatmap, and generate test cases for gaps discovered in the Change impact analysis, using uploaded documents and additional instructions. You can further refine the generation process with documents and instructions for Autopilot^TM to use in generating the test cases.

Visit Generating test cases for a specific transaction, as well as AI-powered generation: Best practices, to check how to successfully generate tests for SAP transactions.

Generate coded automations

Autopilot^TM assists you in generating coded automations, either from text and existent code, or from manual test cases created in Test Manager.

Generating coded automations from text and existent code

In a new or existent coded automation, you can use Autopilot^TM to generate code from natural language, from comments in your automation, or from existent lines of code. Use the Ctrl + Shift + G shortcut, or the Autopilot icon, to generate code.

Generating coded automations from manual tests

After you connect Studio Desktop to Test Manager, navigate to the Test Explorer and search for your manual tests. From here, you can choose one of the following scenarios to generate a coded test case for it:

• Create a coded test case, where the steps are shown as comments, using Create Coded Test Case.. Then prompt Autopilot^TM to generate code based on the comments.
• Directly generate a fully functional coded test case using Generate Coded Test Case with Autopilot.

Visit Creating a coded test case from a manual test case and Generating coded test cases using AI for more information about generating test cases.

Generate coded API automation

If you want to create a coded test case that automates a scenario involving APIs, Autopilot^TM can help with generating the code for this scenario. In the code editor, right-click and select Generate Code then offer Autopilot^TM the necessary instructions. For instance, you should offer Autopilot^TM the API that you want to use, and the API key that it should access.

Refactor coded automations

Autopilot^TM helps you enhance coded automations through refactoring. Consider a situation where a coded test case contains a segment of code that could be more readable. To start the refactoring process, follow these steps: select the desired code, right-click the selection, then select Generate Code. Lastly, offer Autopilot instructions on how to refactor the selected code.

Generate low-code automations

Autopilot^TM assists you in generating low-code test cases, either from text or from manual tests created in Test Manager.

Generating low-code test cases from manual tests

After you connect Studio Desktop to Test Manager, navigate to the Test Explorer and search for your manual tests. Right-click a manual test and select Generate Test Case with Autopilot.

Generating low-code test cases from text

Open your blank low-code test case and select generate with Autopilot from the Designer panel. Enter the desired test steps, and then select Generate to trigger the test case generation.

For more information, visit Generating low-code test cases using AI.

Generate synthetic test data

You can choose Generate with Autopilot when adding test data to your test cases. Autopilot^TM initially generates potential arguments and variables, which you can further refine as required. Visit AI-generated test data to check how to generate synthetic test data using AI.

Generate test insights report

Obtenez des informations exploitables sur vos résultats de test en générant un rapport avec Autopilot, qui détaillera les raisons pour lesquelles vos cas de test échouent continuellement.

Pour générer le rapport, accédez à Exécution. Sélectionnez Générer des résultats ( Generate Insights), choisissez vos résultats de test, puis sélectionnez Générer les résultats ( Generate Insights). Accédez à votre rapport dans Exécution (Execution) sous l'onglet Insights (Insights).

Visit Generate test insights report and AI-powered insights: Best practices to better understand how to identify issues in your test executions.

Import manual test cases - Preview

Use Autopilot^TM to import manual test cases from Excel files. You can import from one file at a time. You can import tests from multiple sheets. The import process transfers all information to Test Manager, unless specified otherwise. For instance, test case properties such as Priority, Status, or Owner are imported into Test Manager as custom field values at the test case level.

To better identify the imported test cases, you can instruct Autopilot, in the Provide import instructions section, to place certain labels on the imported test cases.

To check how to efficiently import manual test cases, visit Importing manual test cases.

Search Test Manager project - Preview

With the help of the Autopilot search, you can search any test object within a project using natural language. If you are not sure what to search for, you can use one of the example search queries that Autopilot provides. Additionally, after returning the results, Autopilot allows you to perform actions on the resulted objects.

For information on using the search and exploring search queries Searching with Autopilot.

Licences

Consultez la page Licences Autopilot pour obtenir des informations sur la façon dont les activités Autopilot sont mesurées et attribuées.

Accès utilisateur

User access management with Autopilot for Testers

La politique de gouvernance de AI Trust Layer vous permet de gérer l'utilisation des fonctionnalités optimisées par l'IA au sein de votre organisation. Bien que tous les membres disposent d'un accès par défaut à ces fonctionnalités, vous pouvez utiliser cette stratégie afin de restreindre les accès, le cas échéant. La politique de gouvernance de AI Trust Layer vous permet de limiter l'accès d'un utilisateur à certaines fonctionnalités spécifiques optimisées par l'IA, ou bien à l'ensemble de ces fonctionnalités, que ce soit au niveau de l'utilisateur, du groupe ou du locataire. Par ailleurs, elle vous permet de choisir les produits d’IA auxquels les utilisateurs peuvent accéder. Vous pouvez créer, modifier et mettre en œuvre cette politique de gouvernance dans AutomationOps.

Si vous souhaitez déployer une stratégie de gouvernance AI Trust Layer et continuer à utiliser les capacités de test optimisées par l'IA, veillez à indiquer Oui dans le champ Activer/Désactiver les fonctionnalités de l'option Activer les fonctionnalités Test Manager.

Consultez les ressources suivantes pour découvrir comment créer, configurer et déployer une politique de gouvernance pour votre organisation.

• Créer une politique de gouvernance
• Paramètres des politiques AI Trust Layer
• Déployer les politiques de gouvernance