When a customer service team grows, they often face a big problem. Their shared support inbox fills up with noise. Real customer complaints need immediate attention. However, these important emails often get lost under vendor newsletters, out-of-office replies, and automated alerts. If you do not have an intelligent filter, every single email creates a new case in Dynamics 365. As a result, salespersons waste hours reading and deleting irrelevant records instead of actually helping customers.
Another issue happens after a case is resolved. When a salesperson figures out how to fix a complex problem, those troubleshooting steps stay locked inside the closed case record. The next salesperson who gets the exact same question has to start completely from scratch. This leads to repeated work and lost knowledge.
To solve these challenges, Microsoft introduced a new set of Out-of-the-Box AI Agents inside Dynamics 365 Customer Service. These AI agents work together to create a fully automated pipeline. They can read incoming emails, create cases, follow up with customers, and even write Knowledge Base articles.
Because this pipeline is so comprehensive, we are splitting our guide into two parts. In this first blog, we will explain the first three steps of the pipeline: AI Email Classification, the ARC Rule, and the core capabilities of the Case Management Agent.
The first step in our automated pipeline is AI Email Classification. Traditionally, routing emails required teams to maintain massive, complex lists of keywords. Not anymore. This feature uses Generative AI to read and understand the actual intent behind every incoming email.
You do not need to set up any keyword rules. Instead, you simply define categories and write a short description in plain English explaining what each category means.
In our setup, we created an “Out of Scope” category to catch and ignore junk emails or out-of-office replies. We also created a “Hardware and Warranty Support” category for real customer issues.
As highlighted in the configuration below, telling the AI what to do is incredibly easy. For our hardware category, we simply wrote a plain-English description. The AI reads this instruction and accurately classifies the emails the exact moment they arrive in the inbox.
After the AI categorizes the email, the system needs to take action. This is where the Automatic Record Creation (ARC) Rule comes in. The ARC Rule acts as a bridge. It decides whether or not to create a case in Dynamics 365.
In our configuration, we set a rule to only create a case if the email is tagged as “Hardware and Warranty Support“. If the AI tags an email as “Out of Scope” junk, the ARC Rule simply ignores it. This keeps your queue clean so your team can focus on real issues.
Once a real case is created, the Case Management Agent takes over. This AI agent has three powerful capabilities. But first, you must complete a quick setup. You need to configure Connection References (to connect Dataverse and Copilot Studio), enable the background Power Automate Flows, and publish the AI agent.
The AI agent also needs its own identity to send emails. You do this by creating an Application User in Microsoft Entra ID and giving it a Customer Service Representative security role. Now, every email the AI agent sends will look professional and official.
The first capability is Case Creation and Update. As highlighted in the configuration below, we set up specific rules for both actions.
First, for case creation, we instructed the AI to automatically predict and fill three important fields: Case Title, Description, and Priority. For instance, if an email subject contains the word “URGENT“, the AI automatically sets the Priority to High so the case arrives in the queue ready for immediate action.
Second, we enabled a Case Update rule for ongoing conversations. If the customer replies to an existing case, the AI reads their new email, understands the new details, and updates the case summary automatically. When a salesperson opens the case, they instantly see the latest information without having to read through a long email chain.
The second capability is Case Resolution. Normally, when a salesperson fixes a problem, they have to manually close the ticket and type a closing email. With Case Resolution, the AI can detect when a customer confirms their problem is fixed. The AI will then resolve the case on its own. It also drafts a personalized closing email based on the case history, rather than sending a generic automated message.
The third capability is Case Follow-up and Closure. In our workflow, if the AI asks the customer a clarifying question but the customer forgets to reply, the case does not just sit in the queue forever.
We set this up using Full automation. This means the AI sends follow-up emails and resolves cases completely on its own, without waiting for a salesperson to approve the action. We also created a rule that specifically targets cases where the Priority is “High“.
Furthermore, we set the trigger to activate only when the Status Reason is “In Progress“. To ensure the follow-up email is highly specific rather than a generic reminder, we instructed the AI to use the Case Title and Case Number for context.
Finally, we configured the exact timing. The AI will follow up a maximum of 3 times, waiting exactly 5 hours between each attempt. If the customer replies and confirms the issue is fixed, the AI automatically sets the case status to “Problem Solved”. If the customer never responds after all three attempts, the case is flagged for a salesperson to review. No customer is ever forgotten.
So far, we checked how the AI can classify an email, create a case, update details, and even handle follow-ups completely on its own. But what happens when cases need to be routed to different departments with different rules? And how do we permanently capture the knowledge when a case is closed?
In Part 2 of this series, we will cover the remaining pieces of the pipeline: the Customer Intent Agent and the Customer Knowledge Management Agent. We will also walk through full end-to-end demo scenarios to see exactly what this looks like from both the customer’s and the salesperson’s perspective.
Conclusion
Managing a high-volume support inbox does not have to be a manual, overwhelming task. By setting up AI Email Classification, ARC Rules, and the Case Management Agent, organizations can instantly eliminate inbox noise and fully automate the triage process. This zero-code setup ensures that every genuine customer email gets a lightning-fast response and that no ticket is ever abandoned in the queue. The AI handles the repetitive background work, freeing up your salespersons to focus entirely on solving complex problems.
FAQs
What is AI Email Classification in Dynamics 365 Customer Service?
AI Email Classification uses Generative AI to understand incoming emails and categorize them based on their intent, without requiring complex keyword rules.
How does ARC automate case creation in Dynamics 365 CRM?
The Automatic Record Creation (ARC) Rule uses email classifications to determine whether an incoming email should create a case in Dynamics 365 CRM.
What can the Case Management Agent do in Dynamics 365 Customer Service?
The Case Management Agent can create and update cases, resolve cases, send follow-up emails, and automate case closure.
How does AI automate case follow-ups in Dynamics 365 CRM?
The Case Management Agent can automatically send follow-up emails based on defined conditions, such as case priority and status, and resolve or flag cases based on customer responses.
Can Dynamics 365 Customer Service AI Agents automate email-to-case management?
Yes. AI Email Classification, ARC Rules, and the Case Management Agent can work together to classify incoming emails, create relevant cases, update case information, and automate follow-ups and resolution.
{ “@context”: “https://schema.org”, “@type”: “FAQPage”, “mainEntity”: [{ “@type”: “Question”, “name”: “What is AI Email Classification in Dynamics 365 Customer Service?”, “acceptedAnswer”: { “@type”: “Answer”, “text”: “AI Email Classification uses Generative AI to understand incoming emails and categorize them based on their intent, without requiring complex keyword rules.” } },{ “@type”: “Question”, “name”: “How does ARC automate case creation in Dynamics 365 CRM?”, “acceptedAnswer”: { “@type”: “Answer”, “text”: “The Automatic Record Creation (ARC) Rule uses email classifications to determine whether an incoming email should create a case in Dynamics 365 CRM.” } },{ “@type”: “Question”, “name”: “What can the Case Management Agent do in Dynamics 365 Customer Service?”, “acceptedAnswer”: { “@type”: “Answer”, “text”: “The Case Management Agent can create and update cases, resolve cases, send follow-up emails, and automate case closure.” } },{ “@type”: “Question”, “name”: “How does AI automate case follow-ups in Dynamics 365 CRM?”, “acceptedAnswer”: { “@type”: “Answer”, “text”: “The Case Management Agent can automatically send follow-up emails based on defined conditions, such as case priority and status, and resolve or flag cases based on customer responses.” } },{ “@type”: “Question”, “name”: “Can Dynamics 365 Customer Service AI Agents automate email-to-case management?”, “acceptedAnswer”: { “@type”: “Answer”, “text”: “Yes. AI Email Classification, ARC Rules, and the Case Management Agent can work together to classify incoming emails, create relevant cases, update case information, and automate follow-ups and resolution.” } }]
}
The post Enhancing Customer Service Automation with AI Agents in Dynamics 365 CRM – Part 1 first appeared on Microsoft Dynamics 365 CRM Tips and Tricks.
