{"id":37347,"date":"2024-02-09T15:56:24","date_gmt":"2024-02-09T10:26:24","guid":{"rendered":"https:\/\/www.inogic.com\/blog\/?p=37347"},"modified":"2024-02-09T15:56:24","modified_gmt":"2024-02-09T10:26:24","slug":"incorporating-start-and-wait-for-an-approval-of-text-action-for-the-ai-generated-response","status":"publish","type":"post","link":"https:\/\/www.inogic.com\/blog\/2024\/02\/incorporating-start-and-wait-for-an-approval-of-text-action-for-the-ai-generated-response\/","title":{"rendered":"Incorporating Start and wait for an approval of text action for the AI generated response"},"content":{"rendered":"

In the previous blog<\/a> we have already explored the AI prompt integration with Power Automate using the \u201cCreate text with GPT using a prompt\u201d\u00a0action, At that time while designing the Power Automate we didn\u2019t add any Approval process where humans could review the AI-generated response.<\/p>\n

Human review through the approval process is important and recommended which ensures the quality of output generated by the GPT model. You can review this MS doc<\/a> to understand why Human review is important in any automation. Sometimes AI model responses can be misleading and need human intervention to adjust and approve those responses to improve the quality.<\/p>\n

In this blog let us focus on incorporating the Approvals along with the \u201cCreate text with GPT using a prompt\u201d\u00a0action and observe the advantage of combining AI with approval where a human can review the AI-generated response and can either approve or reject it (with maybe some modification or maybe with no modification in AI response) which then ultimately enhances the overall quality of AI-generated model outcome.<\/p>\n

In Power Automate while exploring the actions around the approvals connector, recently I came across the new action introduced specifically for validating AI-generated text responses.<\/p>\n

In the previous blog <\/a>we have already created a custom AI prompt named \u201cQA Team Members Quarterly Progress Review\u201d, now let us integrate it within Power Automate.<\/p>\n

Step1:<\/strong><\/p>\n

For simplicity let us select the \u201cManually Trigger a flow\u201d trigger and pass 2 inputs as text boxes.<\/p>\n

Then add an action from the AI Builder connector named \u201cCreate text with GPT using a prompt\u201d as shown below, wherein you can dynamically pass those 2 declared inputs in the previous step:<\/p>\n

\"AI<\/p>\n

NOTE: <\/strong>As you can observe in the above screenshot, the Flow checker shows the warning that appeared and in detail provides a way to fix it, we can skip fixing it but it is always recommended to incorporate the Approval after the \u201cCreate text with GPT using a prompt\u201d.<\/p>\n

\"AI<\/p>\n

Step2:<\/strong><\/p>\n

After #step1, now add the Approvals action, under it select \u201cStart and wait for an approval of text\u201d specifically as this is newly introduced for handling the AI-generated text response.<\/p>\n

\"AI<\/p>\n

You can find more details about the \u201cStart and wait for an approval of text\u201d approval action in this MS doc<\/a>.<\/p>\n

Fill in the details of the required parameters below:<\/p>\n