What is coming To Dynamics 365 Customer Service in 2026 Release Wave 1

Microsoft has just released the 2026 Release Wave 1 notes for Dynamics 365 business applications and the Power Platform, and of course I couldn’t wait to dive in! Over the past few days I’ve been digging through the details to get a better understanding of what’s coming as part of this wave. Just like with every release wave, I’ll be putting together a series of articles and videos to walk through the new functionality and highlight the features that stand out the most. In this article, we’ll take a look at what’s coming in Dynamics 365 Customer Service. Before I get started, I wanted to let you know that the legacy ‘opt-in’ for these features are no longer available. We’ll have to use the monthly release channel to get access to these features The Early Access concept is deprecated and has been replaced by the monthly release channel.

Control evaluation volume with sampling

The Quality Evaluation Agent in Dynamics 365 Customer Service and Contact Center was introduced to us back in October of last year and yes I still owe you an article on this guy! This is an AI agent that helps improve customer experiences by measuring the quality of service provided in conversations and cases. Assessing quality in customer service used to be a manual process, but with this new agent this becomes a lot easier. The agent scores completed cases and conversations autonomously to make sure the service provided complies with an organization standards which can be configured inside the agent. It provides evaluation summaries and recommendations to make future interactions better.

The Control evaluation volume with sampling feature gives supervisors more granular control over how many cases/conversations are being evaluated. According to the documentation sampling will be added to the evaluation plans and will have options like:

• % Based sampling to evaluate a defined number of interactions
• Absolute count sampling to cap the total number of interactions evaluated per schedule

 I am not sure what this will look like, but it sounds like this will add a more granular control on top of the filtering capabilities that we have today.

Define critical questions in evaluation criteria

Today we can configure criteria scoring, which allows us to configure the weight percentage of questions which represents the importance of the question in the overall evaluation.  With this new feature, Supervisors will be able to mark certain questions as critical and choose responses that will automatically fail the entire evaluation, no matter what the other scores are. This helps make sure that important requirements like compliance or required disclosures are always followed. Below is an image of what the criteria experience is today.

Validate knowledge and process adherence with evaluations

I feel like this was missing from the Quality Evaluation Agent, so I am very happy with this feature! This allows us to bring approved knowledge and business processes into the QE agent! Because think about it: how can the agent evaluate if the customer service rep followed the correct steps and processes if it doesn’t have access to what those steps and processes are? The way that this will work is that the QEA will compare the customer service rep’s responses against knowledge sources to determine whether the correct steps and processes were followed during the engagement with the customer.

Simulate case-resolution flows before production

This feature is related to the Case Management Agent, and it will allow us to validate the case resolution logic prior to putting it into production. We will be able to simulate a full end-to-end case resolution flow which can include the drafting of customer response emails without actually sending emails to the customers, intent handling, etc. We’ll be able to run those simulations from the admin center where it will track details like the instructions that were used, responses that were created, agents or tools that were used etc. We will be able to capture metrics from that data to see where the gaps are prior to using this in a production setting.

Use shadow mode for Case Management Agent predictions

The ‘shadow mode for Case Management Agent predictions‘ feature is similar to the feature I just described above. This allows organizations to test and validate the performance of the Case Management Agent using real cases but without modifying cases or sending customer communications. By enabling this feature you can review predicted outcomes like the customer intent that was identified, any case field updates suggested by the agent, any draft responses and recommended follow ups and/or resolutions the Agent suggests and the rationale behind those suggestions. Admins can review these and identify any issues prior to using the agent capabilities in production.

View customer sentiment indicators on a case

I don’t think this feature needs a lot of explanation as the title says it all! As you already suspected, this functionality gives customer service reps a quick view of customer sentiment, helping them spot unhappy customers early and (hopefully) respond faster. By bringing sentiment from email, chat, and voice into one simple indicator for cases, it makes it easier to prioritize cases and make better service decisions. Showing sentiment directly on case forms and case views also saves time because customer service reps don’t have to review every single interaction to understand how the customer feels. This is something that will need to be enabled by system administrators.

Generate AI-driven dynamic forecasts with intelligent method selection

There are also a couple of features to enhance the supervisor experience in Dynamics 365 Customer Service. The first feature is called ‘Generate AI-driven dynamic forecasts with intelligent method selection’ is related to the Workforce Management capabilities in the application. Keep in mind that Workforce Management must be installed to take advantage of this functionality.

This feature allows supervisors to use AI-based dynamic forecasting by looking at historical data. It then generates the forecast for the selected time period and explains the reasoning behind it, why it chose that method, and how confident it is in the newly created forecast. This feature needs to be enabled by an admin before supervisors can start using it.

Enable enhanced screen recording controls for admins

You might have missed the message in the M365 Admin center (I sure did!) on February 25th that informed us of a new feature in Dynamics 365 Customer Service which allows for full-screen recording of customer service rep activity during customer interactions. This feature will be generally available starting March 27, 2026. This feature records exactly what the CSR sees and does on their computer screen during a session, giving supervisors a clear, realistic view of what really happened. It can even capture system audio, and once the session ends, the recording is securely stored in Dataverse. Any local copies are automatically deleted to help protect sensitive data.
Supervisors can then review or download these recordings for coaching, quality assurance, or compliance. Each recording includes details like the case ID and agent name, making it easy to find. Recordings can be up to two hours long and are automatically created based on activity, so there’s little to no setup required. If needed, customer service reps can also start and stop recordings manually. Admins stay in control and can turn this feature off by removing the desktop capture app, disabling automatic triggers, or limiting access so only the right users can view or download recordings.

It is my understanding that the feature mentioned in the release wave builds on top of the functionality I just explained by providing admins more control over how screen recordings are captured and used. With security roles we can decide who can view or download recordings and we can also link recordings to the right cases, making it easier to keep things organized and investigate issues faster. According to the documentation, permissions are clearer too, so supervisors, auditors, and agents each have the right level of access.
Admins can manually attach recordings to cases or conversations when needed (I.E when a session doesn’t originate from a voice call) making sure everything is tied to the correct record. There are also improvements behind the scenes, like better error handling to clean up failed uploads if something unexpected happens, I.E when the video capture app closes unexpectedly.
Don’t worry, all of these updates are turned off by default, giving you time to review and decide what works best for your organization. When you’re ready, you can enable one or more of these features in settings and control who has access through security roles. I hope you enjoyed this article! Be sure to check in again soon or subscribe here to never miss another post!