已不再支援此瀏覽器。
請升級至 Microsoft Edge,以利用最新功能、安全性更新和技術支援。
You want to test how well your autonomous agent responds to different inputs before deployment. What should you do?
Use the Test chat feature in Copilot Studio to analyze the conversation flow.
Deploy the Copilot immediately and gather feedback from real users.
Export the chatbot as a Power Automate flow and then run simulations.
How can you ensure that your agent runs an action based only on user input when needed?
Use condition-based triggers to determine when an action should run.
Set every action to run automatically whenever a user interacts with the agent.
Configure the action to always run in the background, regardless of user interactions.
You want your agent to provide users with the most accurate and up-to-date company information. Which approach optimizes the results?
Configure your agent knowledge sources for SharePoint and Dataverse or upload internal documents to pull your company data to train the agent.
Manually enter all possible responses into predefined topics to ensure accuracy.
Turn on the Allow the AI to use its own general knowledge toggle without configuring knowledge sources.
You want to make your agent available on multiple platforms, such as Microsoft Teams, a website, and a mobile app. What is the best way to deploy it?
Use the Copilot Studio publishing options.
Manually create a separate agent instance for each channel to optimize responses for each platform.
Deploy the agent in one channel first and then rely on users to copy responses to other platforms.
You must answer all questions before checking your work.
此頁面對您有幫助嗎?
需要本主題的協助嗎?
想要嘗試使用 Ask Learn 來釐清或引導您完成本主題嗎?