GPT-4o powered AI summaries: Boosting efficiency in Model-Driven Apps

If you want to empower users with instant insights and accelerate decision-making in your model-driven apps? You can leverage the power of GPT-4 to easily add AI-generated summaries of records directly within your app!

This fantastic feature provides a concise overview of the latest actions and activities, enabling quicker responses and faster actions.

Thanks to Copilot in Dataverse, creating these custom summaries is a breeze. You have full control over the prompt and the data used, tailoring the summary to your specific processes and the fields that matter most.

Here's how you can add this powerful capability to your model-driven app (assuming you have an existing app in development):

  • In make.powerapps.com, navigate to Tables and select the desired table.

  • Under Customisations, click Row Summary. This opens the Copilot prompt editor.

  • In the left-hand section, define the prompt's task, action, data, restrictions, and desired output. For example: "Summarize the latest activity on this case, including the last three activities, their status, and any upcoming deadlines." (More examples below!)

  • To reference specific columns, type "/" , select the table, and then choose the column. This allows the prompt to pull precise data for the summary.

  • Click Test to preview the generated summary based on your prompt.

  • Use the model-selection dropdown (GPT-4o mini, GPT-4o, Premium o1) to choose the best model for your needs. GPT-4o mini is a great starting point for quick summaries.

  • Once satisfied, click Apply to main forms and then Close.

  • Navigate back to the table view and select Forms to see the forms displaying the summary.

  • Use the Row summary option in the ribbon to hide or edit the summary.

  • Refresh and publish your model-driven app to deploy the new feature.

  • Open the app, select a record, and witness the magic of AI-generated summaries!

Example prompts:

  • "Summarize the key details of this account, including industry, annual revenue, and primary contact."

  • "List the open tasks associated with this opportunity, ordered by due date."

  • "Identify any potential risks or issues related to this project, based on the latest notes and activities."

  • My example: “Summarize Application record in a way that is easy to understand for a case officer. Make sure to include all important information, including main ideas and important details, while keeping the order of the content logical. Remove any repetitive elements to make the summary as concise as possible without losing the original text’s integrity. If the original text is too short to condense, present it as the summary.

    Don't include the title in the summary. Create the summary in two paragraphs. For the first paragraph, use the instructions under the header Data. For the second paragraph, use the data under the header Activity. Leave a blank line between the two paragraphs. At the start of the response add "Here is a quick summary of the record  “Application.ApplicationID". Don't include the titles Data & Activity in the paragraphs.

    Data:

    Show the below values as bullet points using markdown text and bold the values. Make the application Identifier a navigation link to the contact record. The link must not include query string parameters in it.

    Application.ApplicationID

    Application.Builder

    Application.ApplicationID (Property).Postcode

    Application.StartDate

    Application.Description

    Application.ApplicantID (Applicant).FirstName

    Application.ApplicantID (Applicant).LastName

    Application.CaseOfficerID (Case Officer).Email

    Activity:

    Generate a summary paragraph on actions using the following activity and task table fields identifying top 2 key information and top 2 important actions to take.

    Application.Regarding (Activity).Subject

    Application.Regarding (Task).Subject“

It's that simple! Give it a try and supercharge your model-driven apps with the power of AI-generated summaries. Your users will thank you!

Next
Next

Gamifying Field Service Azure Maps: My Saudi Arabia Bootcamp Experience!