Skip to content

Latest commit

 

History

History
112 lines (55 loc) · 3.91 KB

Exercise04.md

File metadata and controls

112 lines (55 loc) · 3.91 KB
Raw

Exercise 4 - Detect positive, negative, or neutral sentiment

In this exercise rather than use a pre-built model. We will be creating our own model.

  • Using the same in-private browser instance, navigate to the Power Automate site if you are not already there.

  • In the top right of the screen, ensure the Environment is set to Dev One.

  • Choose Create from the navigation.

  • Then choose Automated cloud flow

  • In the following prompt select When a file is created for the OneDrive for Business trigger. Then click Create.

  • Click the When a file is created action.

  • There should be one Trigger on the page that needs to be configured.

  • Select the Trigger and select Root

  • Next click on the + sign and choose Add an action

  • In the slideout search for sentiment and choose Analyze positive or negative sentiment in text. This will add another step to the flow.

  • The properties side panel will open. In the Language section click on the dropdown and choose English

  • In the Text field we need to pass in the value provided from OneDrive. Specifically the content. Click inside the text field and it will be highlighted with a blue bounding box. Then click on the blue lightning bolt.

  • In the popup search for content in the text box and then choose File content

  • Next click on the + sign and choose Add an action

  • Search for condition and select Condition from the options provided

  • The resulting condition

  • Let's set the Condition parameters. Click in the Choose a value field and you will notice the blue box. Click on the lightning bolt. Then scroll down and select Overall text sentiment. In the second text box put in the word positive.

    sentencesentiment

  • In the True path add another Action

  • Search for notification and choose Send me an email notification

  • You will be prompted to create a new connection. Click Create New.

  • Click in the Subject text box and click on the lightning bolt and select Sentence sentiment

    subject

  • Do exactly the same steps for the Body content

  • Then repeat the same steps on the False path.

  • Click on Save and wait a few moments until you get a confirmation that the flow has been saved

  • Then click on Test and choose the radio button for Manually and finally click the Test button. Similar to the previous exercise you need to kickoff the test, switch the OneDrive and upload one of the sample review files from the downloaded sample data: AIBuilderLabFiles\Reviews.

    Note: It may take several minutes for the flow to start after you upload the review to OneDrive. Switch back to the browser tab with your flow to see it start.

  • After the flow completes it should resemble the following

  • You can also visit the Outlook web client to see the notification and text that was sent.

Summary

In this exercise, you created a flow that leveraged a pre-created AI Builder model to parse and analyze sentiment data out of uploaded review data all without writing any code.