AI Builder in Power Automate allows users to integrate AI-driven automation into workflows, helping with tasks like image recognition, text analysis, form processing, and predictive analytics. It enables businesses to automate manual processes, reduce errors, and improve decision-making using machine learning models.
Extract text from documents automatically
Analyze sentiment in customer feedback
Recognize objects in images and process forms
Detect and categorize emails intelligently
1️⃣ What is AI Builder?
AI Builder is Microsoft Power Platform’s AI toolset that enables users to add AI capabilities to Power Automate, Power Apps, and Dynamics 365.
It includes prebuilt AI models and custom AI models that can be trained on business data.
Common AI Models in AI Builder:
✔️ Form Processing – Extracts text from invoices, receipts, and forms
✔️ Text Recognition (OCR) – Reads text from images and scanned documents
✔️ Object Detection – Identifies objects in images
✔️ Sentiment Analysis – Determines the tone of text (positive, neutral, negative)
✔️ Entity Extraction – Identifies key data like names, dates, and locations
✔️ Prediction Models – Forecasts outcomes based on past data
2️⃣ How to Use AI Builder in Power Automate
Step 1: Enable AI Builder in Power Automate
1️⃣ Go to Power Automate → Navigate to AI Builder
2️⃣ Click “Explore” and select a model
3️⃣ Choose “Train a model” (for custom models) or use a prebuilt model
Step 2: Create an AI-Powered Flow
1️⃣ Open Power Automate → Click “Create” → Select “Cloud flow”
2️⃣ Choose a trigger (e.g., “When an email arrives”)
3️⃣ Add an AI Builder action (e.g., “Extract text from an image”)
4️⃣ Process the AI output (e.g., Save to SharePoint, SQL, or Dataverse)
5️⃣ Send a response (e.g., Send an approval email or update a record)
Example:
✔️ Trigger: Email with an attachment (invoice)
✔️ AI Model: Form Processing extracts invoice details
✔️ Action: Store extracted data in SharePoint or Excel
✔️ Outcome: Automates invoice processing without manual entry
3️⃣ Use Cases: AI Builder + Power Automate
Scenario 1: Automating Invoice Processing
Use Case: Extract data from invoices and save it automatically.
Flow Steps:
1️⃣ Trigger: Email receives an invoice PDF
2️⃣ AI Builder (Form Processing) extracts invoice number, date, and amount
3️⃣ Data is saved to SharePoint or a database
4️⃣ Power Automate notifies the finance team
Example:
- Before AI: Manually inputting invoice details
- With AI Builder: AI extracts & stores data automatically 🚀
Scenario 2: Sentiment Analysis for Customer Feedback
Use Case: Analyze customer emails or surveys to detect satisfaction levels.
Flow Steps:
1️⃣ Trigger: A customer submits a feedback form
2️⃣ AI Builder (Sentiment Analysis) detects positive, neutral, or negative tone
3️⃣ If negative, Power Automate alerts the support team
4️⃣ If positive, send a thank-you email
Example:
- Customer Feedback: “I am very unhappy with the service.”
- Bot Response: AI detects negative sentiment → Escalates to customer support
Scenario 3: Automating Resume Screening for HR
Use Case: AI Builder scans resumes and extracts key details (name, skills, experience).
Flow Steps:
1️⃣ Trigger: A candidate submits a resume
2️⃣ AI Builder (Entity Extraction) pulls name, skills, experience
3️⃣ Data is stored in SharePoint or Dataverse
4️⃣ HR team is notified for review
Example:
- Before AI: HR manually reads every resume
- With AI Builder: Resumes are automatically processed and categorized
Scenario 4: Object Detection for Quality Control in Manufacturing
Use Case: Detect defective products in a production line.
Flow Steps:
1️⃣ Trigger: Camera captures product images
2️⃣ AI Builder (Object Detection) identifies defects
3️⃣ If defective, Power Automate alerts the quality team
4️⃣ Data is logged for tracking
Example:
- Before AI: Manual product inspection
- With AI Builder: AI automatically detects issues & alerts teams
4️⃣ Best Practices for Using AI Builder in Power Automate
✔️ Use Prebuilt Models First → Avoid training new models unless necessary
✔️ Optimize Data Inputs → High-quality data improves accuracy
✔️ Enable Error Handling → Handle cases where AI predictions are incorrect
✔️ Combine with Dataverse → Store AI-extracted data efficiently
✔️ Monitor Model Performance → Retrain models with fresh data for better accuracy
Tip: Use AI Builder + Power Automate + Power BI for real-time analytics.