Advanced Power Platform: Customer Portals, Virtual Agents, and AI Models
Dynamics 365 features many out-of-the-box AI capabilities, bridging the gap between paper and digital. AI functionality within D365 is constantly being improved with the intention of streamlining processes.
Four Examples of AI Components Built into Business Central
- Sales invoices predicted to be overdue uses AI technology to determine the likelihood a customer will pay a transaction on time.
- The AI image analyzer determines inventory item characteristics based on photos. Items can be searched for and categorized by attribute.
- Azure AI Forecast can be enabled when building a model in cash flow forecasts.
- The inventory forecasting AI model is used to configure inventory forecasts with the date and interval thresholds.
Outside of out-of-the-box AI functionality, the Power Platform gives you the power to innovate your business by customizing your systems and processes with little to no coding.
- Analyze your data with Power BI.
- Solve business challenges with low-code, customized Power Apps.
- Automate and streamline processes with Power Automate.
- Assist customers and prospects with Power Virtual Agents.
In this blog, we’ll highlight some of the capabilities within each of these areas that businesses are using to be more productive.
Common Data Service AI Capabilities
Business systems can be accessed from within the common data service through virtual entities, which automate processing.
AI builder lab solutions in Power Apps offer useful solutions in Dynamics 365. Object detection product, which is included in AI builder lab solutions under entities, includes a table with records representing product inside the system. Users typically reference a table source when processing AI capabilities, making the object detection product feature a great place to start with AI.
Model capabilities to be set up and used in a Power App or Power Automate flow are found under the build section of the AI builder. Model capabilities include category classification, which is used to categorize text information, form processing, which can convert a handwritten document into a digital record, and object detection, which is used to recognize products, product type, and product quantity.
Object detection is a recently enhanced feature in Power Apps. Model domains have been added to object detection, making the AI algorithm much more precise and consistent. In addition to identifying objects, this tool is also able to recognize logos.
Repetition is used to improve the computer’s object recognition ability. Computers can be trained to better recognize items by inputting object images into object detection. At least 15 images are needed per object, and 30 images are recommended for the best results. Inputting images is a time-consuming process and will improve in the future.
Portal capability is one of the newer features in the Power Platform. Portals access data from the Common Data Service and can be tailored as needed. To generate a new portal, click on “new app” and select “portal”. Preset templates are available to simplify this process.
Power Virtual Agents
Power Virtual Agents are a derivative of the Azure bot framework. Bots eliminate repetitive and mundane pieces of work that a human would normally execute. When mundane work is automated, humans can focus on more important issues instead. For example, bots can be used to monitor open customer service chats, reducing the number of conversations customer service representatives must respond to.
Learn more about the Power Platform in the Power Platform section of the Stoneridge Software Blog.