Overview
Power Automate enables seamless automation of AI Builder models, allowing businesses to process documents, images, and predictions automatically. By integrating AI Builder with Power Automate, you can create intelligent workflows that analyze data, extract information, and trigger actions without manual intervention.
1️⃣ What is AI Builder in Power Automate?
AI Builder is a low-code AI platform integrated into Microsoft Power Platform.
It enables automation using prebuilt and custom AI models for tasks like:
- Text Recognition & Sentiment Analysis
- Form Processing (Invoices, Receipts, Documents, etc.)
- Object Detection in Images
- Category Classification
- Prediction Modeling
🔹 Use Case Example: Automatically process incoming invoices, extract key details, and store them in SharePoint or Dataverse.
2️⃣ Common AI Builder Actions in Power Automate
| AI Model | Purpose | Example Use Case |
|---|---|---|
| Form Processing | Extracts data from structured forms | Read invoice amounts and save in Dataverse |
| Object Detection | Identifies objects in images | Detect products in warehouse photos |
| Prediction | Analyzes historical data to make predictions | Predict customer churn |
| Text Recognition (OCR) | Extracts text from images & PDFs | Read text from scanned documents |
| Category Classification | Sorts text into predefined categories | Categorize support tickets automatically |
| Sentiment Analysis | Detects positive, negative, or neutral sentiment | Analyze customer feedback in emails |
3️⃣ How to Automate AI Builder Actions in Power Automate
Step 1: Select a Trigger
Start by choosing a trigger to initiate the automation:
Examples:
- When an email is received with an attachment
- When a file is added to SharePoint
- When a Power Apps button is clicked
Example: Trigger when a new invoice is uploaded to a SharePoint folder.
Step 2: Add an AI Builder Action
Once the trigger is set, add an AI Builder action:
1️⃣ Click “+ New Step”
2️⃣ Search for AI Builder
3️⃣ Choose an AI model based on the use case:
- Process and extract data from a document
- Detect objects in an image
- Analyze text sentiment
Example: Use “Extract information from forms” to process invoices.
Step 3: Process the AI Builder Output
After AI Builder extracts data, use Power Automate actions to process the results.
🔹 Examples:
- Store extracted invoice details in Dataverse or SharePoint
- Send an approval request if the invoice amount exceeds a limit
- Add sentiment analysis results to a Teams notification
Example: Store extracted invoice amount, date, and vendor name in a SharePoint list.
Step 4: Handle Errors and Improve Accuracy
To improve AI model accuracy and handle failures:
Use conditions to validate extracted data.
If data is missing, send an approval request for manual verification.
Train custom AI models to improve recognition accuracy.
Example: If AI Builder fails to extract invoice details, send a manual review task in Power Automate.
Step 5: Automate the Next Steps
After processing the AI Builder results, automate the next actions:
Examples:
- Send a Teams notification for approvals
- Store extracted text in Dataverse or Excel
- Trigger Power Apps for manual validation
Example: After extracting invoice data, trigger an approval flow for financial review.
4️⃣ AI Builder and Power Automate Use Cases
| Use Case | Trigger | AI Model Used | Next Steps |
|---|---|---|---|
| Invoice Processing | New invoice uploaded | Form Processing | Save data in Dataverse & trigger approvals |
| Customer Sentiment Analysis | Email received | Sentiment Analysis | Alert support team for negative feedback |
| Product Defect Detection | New image uploaded | Object Detection | Create a defect report in SharePoint |
| Contract Data Extraction | New contract scanned | OCR Text Recognition | Store extracted terms in SharePoint |
| Expense Report Automation | Receipt uploaded to OneDrive | Form Processing | Auto-fill an expense report in Excel |