All you need to know about AI Builder & Microsoft Power Platform

Companies around the world have incorporated the advanced benefits of AI into everyday operations. From gauging how emerging trends in fashion will affect production to ensuring that perishable products in the dairy industry meet regulatory requirements around the clock, the ease and speed with which businesses make critical decisions is now underpinned through AI tools.

Many users are familiar with text-based tools such as GPT-4 and others that feature prominently in popular media. However, AI tools come in limitless forms offering a wide range of functionalities. Microsoft’s AI Builder, a key part of the Power Platform, offers a no-code/low-code environment which enables companies to build custom AI models that automate processes and extract insights without the hassle of training employees in advanced coding skills.

These models enhance accuracy by eliminating errors through manual efforts which benefit companies in numerous ways. Document processing, gauging customer sentiment with near-human intuition and providing tools to forecast future performance are just some examples. In this blog we will be exploring AI Builder’s capabilities and how it connects to and enhances those of the Power Platform.

AI Builder creates custom AI models, such as form processing and object detection, using their own data from sources like Dataverse, OneDrive, SharePoint, and Azure. It integrates seamlessly with Power Apps, Power Automate, and Microsoft 365, allowing businesses to build streamlined workflows with AI-driven insights at their core.
In Power Apps, users can embed AI models into their applications without writing any code. A company can use a form processor model in Power Apps to automatically extract and display data from submitted forms within the app, which ensures accurate data entry within minutes. Power Automate incorporates AI models into automated workflows, such as processing email attachments or triggering actions based on extracted information, which once again removes tedious manual efforts. When this data is linked to Microsoft 365, executives can draw insights, from text to interactive, with just a touch of their fingertips.

AI Builder also integrates with over 500 connectors, including Dynamics 365 and Azure, which allows for multi-platform data flows to guide further automation and predictive analytics.

With AI Builder, companies can make use of an extensive list of capabilities as outlined below.

• Document processing:

As a signature feature of this technology, document processing allows businesses to automatically collect data from forms, invoices, PDFs, and other structured or unstructured documents. This reduces manual data entry errors and speeds up workflows. In invoice processing, a custom AI tool can be trained to identify key details across different invoice formats. During the training phase, these identifiers will be tagged and the tool adopts machine learning technology to accurately identify the desired data from future documents. This information can be fed into Power Apps and Power Automate for the processing of payments and further analysis.

• Sentiment analysis:

Sentiment analysis is the process of analyzing text to determine whether it expresses positive, negative or neutral emotions. This is of increasing importance in the hospitality industry where customer reviews are critical in building reputation and loyalty. AI Builder includes a pre-built sentiment analysis model that can automatically analyze text to determine the overall sentiment being expressed. This streamlines customer feedback analysis allowing businesses to understand opinions and tailor responses accordingly. The model can feed information to Power Automate workflows to route negative feedback to customer support, while insights gained from positive feedback can be used for marketing purposes.

• Category classification:

This model helps businesses categorize text into predefined categories. For example, incoming customer emails can be automatically sorted into categories such as “billing inquiries,” “technical support,” or “general feedback,” allowing teams to prioritize tasks and respond more effectively. This classification can be directly implemented in Power Apps. A global chemical company used this feature to identify requests sent in emails and convert them automatically to form submissions through Microsoft 365 where they could be processed more efficiently.

• Entity extraction:

AI Builder’s entity extraction model identifies and extracts key data points like names, dates, locations, and more from large documents. This is particularly useful for processing complex legal documents and contracts which are a key feature in the food and beverage (F&B) industry where multinational companies have to navigate regulatory requirements across multiple countries. By connecting to data sources like SharePoint or Dataverse, businesses can automate the identification of necessary information, once again reducing the time spent on manual efforts.

• Prediction:

The predictive analytics model in AI Builder uses historical data to forecast future outcomes. Companies can predict sales trends, customer churn rates, or inventory needs which are critical in the fashion industry which experiences the emergence, rise and decline of trends in a span of a few weeks. This model can be integrated into Power Automate flows, where predictions trigger specific actions, such as ordering more stock or launching targeted marketing campaigns.

• Object detection:

Object detection models identify objects in images and provide precise coordinates, helping businesses analyze and manage visual data effectively. In the food and beverage industry, object detection can play a crucial role in reducing food waste by automating the monitoring of inventory in real-time. Manufacturers can use the technology to detect packaging defects that could lead to waste. By continuously monitoring inventory, companies can quickly respond to such issues to limit faulty production that could lead to discarded items.

• Image description:

AI Builder’s image description model automatically generates descriptive tags for images, making it easier to manage and search for content. This is useful for media companies or any business managing large volumes of visual content. Similarly, Microsoft’s Azure VI Video Indexer derives insights from stored videos to allow for informed decisions on how advertisements or product placements can be incorporated into promotional material.

With the AI tools at hand being employed across industries, the following are just a few ways in which they have brought about significant change.

• Quality control

Ardent Mills, a leading flour-milling company in North America, faced challenges with manual data entry and limited real-time reporting, affecting efficiency and quality control. To address these issues, the company developed the “Cross Check” app using PowerApps, integrating SharePoint Online to streamline data entry and enhance data accuracy.

Building on these advancements, Ardent Mills is piloting the use of Lobe with AI Builder to train models in their baking lab. This pilot focuses on bread baking and grain grading, using AI to detect and flag products that require further evaluation, ensuring quality control and improving operational processes even further.

• Sentiment analysis

The versatility of this tool shows its influence in the field of education services being revolutionized by AI tools. Microsoft Teams provides in-built features that let students react to taught material through simple gestures such as selecting an emoji or leaving a quick comment. The AI tool processes and identifies the responses so educators can make informed decisions on how to adapt their material for future lessons. Companies can similarly seamlessly integrate this tool into feedback and accessibility surveys on their websites to improve navigation and ease of access to their services.

• Object detection

Electrolock, a leading insulation products provider, improved inventory management by leveraging AI Builder’s object detection capabilities. This digital solution eliminated the cumbersome tasks of managing inventory through manual spreadsheets and email. By integrating AI Builder’s object detection within a PowerApps canvas app, they developed a mobile-friendly platform that allowed customers to track and pick inventory directly from the manufacturing floor. This digital system, connected to Microsoft Dataverse and their ERP, enabled accurate recording of inventory usage, quick purchase order creation, and precise revenue tracking. The solution improved inventory processing with the acquisition of accurate and timely data. Object detection ensured that goods were correctly identified and tracked, providing key insights into usage and replenishment needs.

• Outcome Prediction

Toyota Lanka, a member of the Toyota Group, faced delays in strategic decision-making due to manual data extraction and fragmented reporting from multiple data sources. To address this, Fortude developed a business intelligence platform built on Microsoft Power BI and modern data warehousing technologies, integrating predictive analytics and machine learning.These models optimized inventory management, sales forecasting, and operational decisions by analyzing complex, multilayered data in real-time. Fortude’s solution also included advanced data visualization techniques, allowing Toyota Lanka to monitor key performance metrics through interactive dashboards with 360° visualization.