AI Builder: Prebuilt models

What is AI Builder and why use prebuilt models

AI Builder is a feature of the Microsoft Power Platform designed to enable the creation and integration of AI models into apps, flows, and business processes. Through a low-code approach and an intuitive graphical interface, users can select prebuilt models or create custom ones.

Prebuilt models are particularly useful because they require no training phase: they are immediately available and optimized for common scenarios such as text recognition, document processing, and linguistic analysis. This allows for quick integration in Power Apps Maker Portal or Power Automate flows.

Main advantages

• Immediate implementation with no training needed.
• Native integration with Power Apps, Power Automate, and Dataverse.
• Optimized models for common business scenarios.
• Reliable results using Microsoft AI technologies.

List of available prebuilt models

AI Builder preconfigured models are divided into two main categories: visual models and linguistic models. The table below summarizes the main types and their use cases.

Category	Model	Description
Computer Vision	Business Card Reader	Extracts structured data from business card images such as name, email, and company.
Computer Vision	Identity Document Reader	Recognizes and reads identity documents such as passports or ID cards.
Document Processing	Invoice Processing	Analyzes invoices to automatically extract amounts, dates, and relevant fields.
Document Processing	Receipt Processing	Processes photographed receipts to extract payment data and vendor details.
Computer Vision	Text Recognition (OCR)	Recognizes text from images or PDFs, useful for archiving and digitalizing documents.
Natural Language Processing (NLP)	Category Classification	Automatically classifies text into predefined categories, useful for tickets or emails.
Natural Language Processing (NLP)	Entity Extraction	Identifies entities like names, places, dates, or amounts within text.
Natural Language Processing (NLP)	Key Phrase Extraction	Extracts key phrases from text for content analysis and automatic summaries.
Natural Language Processing (NLP)	Language Detection	Automatically detects the language of a text.
Natural Language Processing (NLP)	Text Translation	Translates text into different languages using Microsoft Azure linguistic services.
Natural Language Processing (NLP)	Sentiment Analysis	Determines the tone and sentiment of a text (positive, negative, neutral).

Integration of prebuilt models with Power Platform

AI Builder prebuilt models can be easily integrated into different components of the Microsoft platform:

• Power Automate: use AI Builder actions in flows to automate processes based on images or text.
• Power Apps: insert AI Builder controls in Canvas or Model-driven apps to analyze data in real time.
• Dataverse: store and manage model results in custom tables.
• Power BI: visualize data processed by AI models in interactive reports and dashboards.

This interoperability is made possible by the Power Platform and its services integrated with Microsoft Azure, ensuring security, scalability, and compliance.

How prebuilt models work

The following diagram shows the typical integration flow of a prebuilt AI Builder model within a Power Automate flow or Power Apps app.

When to use prebuilt models

Prebuilt models are ideal when you want to quickly implement AI features without managing datasets or training processes. Example applications include:

• Automating invoice or receipt management.
• Analyzing customer feedback sentiment.
• Automatically detecting the language of incoming emails.
• Digitizing business cards and paper documents.

When more specific or customized behavior is needed, it is preferable to use AI Builder custom models, which allow training with your own data.