A Dynamics 365 Customer Service admin has just configured the Case Management Agent to auto‑populate case fields from incoming emails and added field descriptions to help the AI make better predictions. The configuration looks right, but how will the admin ensure that the AI is able to predict case field values accurately? The new case enrichment simulation in Dynamics 365 Customer Service—now generally available—answers exactly that question.

Admins can measure AI‑powered field prediction accuracy against real organizational data before enabling case enrichment in production: test against historical records, review the results, refine field descriptions, and enable the feature when accuracy meets your bar. 

Empower admins to validate AI quality 

Case Management Agent simulations include: 

• Flexible test data — run predictions against historical case records or sample email and chat data uploaded via Excel 

• Field‑level match reports — see prediction accuracy across every field 

• Case‑level comparison view — review actual versus predicted values side by side 

• Excel export — share results offline with stakeholders 

• Re‑run capability — iterate after refining lookup field descriptions or prediction rules 

• Shadow runs — Validate predictions by activating case enrichment with live production data without writing to records. 

Run your first simulation today 

Don’t wait until production to find out how well the AI performs. Open the Copilot Service admin center, navigate to Case settings > Case Management Agent > Case creation and update > Simulation, and run your first test against historical data. Within minutes you will know exactly where to tune your field descriptions before turning on case enrichment. 

For step‑by‑step guidance, read the documentation: 

• Set up Case Management Agent | Microsoft Learn 

The post Run case enrichment simulation to assess AI prediction quality in Dynamics 365 Customer Service  appeared first on Microsoft Dynamics 365 Blog.

Customers expect quick and smooth help when they reach out to customer support—whether it’s about a delayed order, a product question, or a service issue. Good case management means keeping track of all these requests and associated customer conversations to make sure nothing falls through the cracks. 

Managing these cases is not always easy. Fortunately, the case create and case update capabilities of the Case Management Agent (CMA) make life easier for support teams and hence, customers. 

From manual to automated 

A case is like a digital record that holds all the details about the support request. Traditionally, whenever a customer contacts customer service—by chat, email, or voice—a service rep creates a case and manually tracks the issue from start to finish. Now, case create and case update capabilities make this process smarter and more automated than ever. 

Key features and capabilities 

We released the first version of CMA in public preview in April 2025, and now it’s generally available. We’ve taken customer feedback into account and added capabilities to make creating and updating cases more holistic than ever: 

1. Automatically create a case from chat when the customer service representative accepts the chat request and automatically update it at the end of the conversation. These capabilities are now supported for voice calls and social channels as well. 

2. When the conversation ends, the conversation summary is posted as a note on the case timeline.   

3. For any incoming customer email, case will be created through automatic record creation (ARC). The case will be kept updated with details from all incoming customer emails. 

4. The ability to configure rules based on which case or related entity fields will be predicted. This means you can configure a rule that updates a certain set of fields based on the product in discussion.  

5. Lookup fields are now supported for predictions. You can add field descriptions for capturing your business context which gets leveraged by AI for prediction. For example, for case category prediction, you can describe what each case category means in your business. 

6. We have made changes to the administrator settings layout and in the service rep experience when AI updates are made to the case they’re viewing. This will ensure a more intuitive experience. 

7. You can leverage multiple queue support in automatic record creation (ARC) for creating cases from emails.  

Example scenario 

Let’s say a customer sends an email to customer support about their device not working properly. As soon as this email is received, Case Management Agent creates a case with the relevant details from the customer’s email. The AI agent answers the email to request further details to solve the issue. The customer sends an email answering AI agent’s questions.  

In a pre-CMA world, the customer service rep would have to input these details in the case manually. Not anymore! CMA updates the configured case or related entity fields automatically using this new email from the customer. Everything is kept in one place, so the next agent who helps the customer has the full story—no confusion, no delays. 

Learn more 

Watch a quick video introduction. 

To learn more, read the documentation:  

• Set up Case Management Agent to create and update cases | Microsoft Learn 
• Use Case Management Agent to create and update cases | Microsoft Learn

The post Effortlessly create and update cases with Case Management Agent appeared first on Microsoft Dynamics 365 Blog.