Microsoft has introduced a new real-time streaming analytics experience for Dynamics 365 Contact Center. The new experience, also called the “Supervisor Experience”, is designed for supervisors who need to see what is happening in the contact center in near real-time.
In this post, we will look at what real-time streaming analytics is, how it differs from the existing real-time analytics reports, the different views available to supervisors, and some of the actions we can take directly from the new supervisor experience.
What Is Real-Time Streaming Analytics?
Dynamics 365 Contact Center already has analytics reports. These provide useful operational information such as queue performance, representative activity, active conversations, and service levels. The new real-time streaming analytics experience goes further. Instead of periodically refreshing an entire report, the new experience uses an event-driven architecture that continuously delivers changes to operational metrics. This means supervisors can watch values such as queue backlog, representative availability, active conversations, and longest wait time change as the contact center changes.
Where traditional reporting would periodically refresh views of the contact center, real-time streaming analytics continuously updates the operational view of the contact center:
Real-Time Streaming Analytics Versus the Existing Real-Time Reports
It is worth calling out that this is not simply a new visual layout for the existing Power BI-based Omnichannel real-time analytics dashboard. The existing real-time analytics experience includes reports such as Summary, Voice, Bot, Agent, Ongoing Conversation, Backlog Conversation, Intent Group, User Group, and Proactive Engagement.
The newer real-time streaming experience introduces its own supervisor interface with these main areas:
- Wallboard
- Assisted Service
- Queues
- Representatives
- Conversations
- Self Service
The important architectural difference is the streaming approach. Traditional dashboards refresh on scheduled intervals and may re-render the full report, whereas the new streaming experience receives live operational changes.
Live Metrics and Aggregated Metrics
One of the most important concepts in the new analytics experience is the difference between live metrics and aggregated metrics. Live metrics show the current state of the contact center and continuously update.
Examples include:
- Active conversations
- Queue backlog
- Online representatives
- Longest wait time
These metrics are identified in the interface using a Live indicator:
The UI shows a “Last Refreshed” time, which is constantly update in seconds.
Aggregated metrics are calculated over a selected reporting period. Examples include:
- Service level
- Abandon rate
- Average speed to answer
Note the available Period filters to set the aggregated time period:
Note the current period, which applies to aggregated metrics, not live metrics:
You can also filter by channels, business units, and queues, and save your filters by bookmarks so you can easily reapply them when needed.
If we select the last 15 minutes as a filter, an aggregated metric such as abandon rate will be calculated using conversations that fall within that period. Live metrics represent the current operational state and instead respond to applicable filters such as business unit, channel, or queue.
Accessing Real-Time Streaming Analytics
For the latest from Learn, go to this link: https://learn.microsoft.com/en-us/dynamics365/contact-center/use/realtime-streaming
The new experience is currently accessed through the Dynamics 365 Contact Center supervisor portal rather than through the familiar embedded analytics page in Copilot Service workspace. There are regional URLs for both production and First Release (FRE) environments.
For example, the current production URL for the United States is:
https://portal.us.contactcenterai.powerplatform.com/experience/supervisor
And if you have a first release environment, the FRE URL is:
https://portal.us.fre.contactcenterai.powerplatform.com/experience/supervisor
Other regions have their own endpoints, including Europe, Canada, Australia, United Kingdom, Japan, Germany, France, India, Brazil, Switzerland, UAE, and Asia.
Before using the experience, check the prerequisites for your environment.
These include:
- Omnichannel for Contact Center configured and in use
- The required supervisor security role
- At least one customer interaction channel, such as voice, chat, or asynchronous messaging
- Interaction data being generated in the environment
FRE:
PROD:
The Wallboard
The first view we will look at is the Wallboard. The Wallboard is designed to give supervisors a high-level operational view of the contact center. It brings together workload, queue health, representative availability, and service performance on one screen.
Depending on the available data and filters, we can monitor items such as:
- Active conversations
- Online representatives
- Representative presence
- Queue backlog
- Longest wait time
- Service level
- Abandon rate
- Average speed to answer
This is the view that would work particularly well on a large monitor in a contact center operations area. The experience also includes trend indicators. These help us see whether a metric is improving or getting worse over time. For example, if abandon rate begins increasing, the supervisor can identify that condition quickly rather than waiting for a historical report later in the day.
The wallboard can also be shown in dark mode, and full screen:
Thresholds and Visual Indicators
Another useful capability is the use of operational thresholds. Metrics can appear in one of three states:
- Normal
- At risk
- Critical
An At risk metric can be visually highlighted, while Critical can use a stronger warning indicator.
This is useful because supervisors do not need to manually interpret every metric individually. Instead, they can immediately see where conditions are deteriorating. In dashboard areas where details are available, we can open the metric to inspect more information such as:
- Threshold range
- Metric definition
- Calculation
- Trend information
This moves the experience beyond simply displaying numbers, it gives us an operational signal that something needs attention.
The Queues View
The Queues view lets us move from the contact-center-wide picture into individual queues. We can filter queues using criteria such as:
- Period
- Channel
- Business Unit
- Queue
We can then inspect operational metrics for each queue. Note that queues are grouped by Business Unit:
Some of the available queue metrics include:
- Queue backlog
- Longest wait time
- Average speed to answer
- Abandon rate
- Service level
- Online representatives
- Representative presence
- Conversations answered versus offered
- Transfers in versus out
- Callback statistics
The queue backlog is particularly useful because it represents incoming conversations that remain open and are waiting to be assigned to a representative. This can include conversations that entered the queue directly, escalated from a virtual agent, or were transferred from another queue.
The Longest wait time metric shows how long the longest currently waiting conversation has been waiting. Together, backlog and longest wait time can give us an immediate indication that a queue is starting to struggle.
For example, imagine we see a Billing Queue with high backlog, high longest wait time, and low number of service representatives, a supervisor can investigate representative availability and potentially move capacity to the queue. We will see shortly how this can be done in these new dashboards.
The Representatives View
The Representatives view gives supervisors real-time visibility into individual representative status, availability, capacity, workload, and queue assignments.
At a high level, we can see counts for:
- Online representatives
- Offline representatives
- Representatives in active conversations
- Idle representatives
- Representatives in wrap-up
We can then inspect individual representatives in the grid.
For each representative, the experience can show information such as:
- Presence
- Duration in the current presence
- Available capacity
- Consumed capacity
- Answered versus offered assignments
- Active conversations
- Unanswered assignments
- Assigned queues
- Skills
This is where the new experience becomes operational rather than purely analytical. Supervisors can review the history of status changes of a representative presence:
And change the status of a service representative right in the UI:
Supervisors can use the information to understand whether the contact center has available capacity and where that capacity currently sits. Supervisors can move CSRs to new queues based on demand:
And also update skills and capacity profiles and units:
This creates a natural workflow:
The Conversations View
The Conversations view provides a centralized view of conversations in the contact center. Supervisors can search and filter conversations and then inspect individual interactions. This becomes useful when a high-level KPI tells us there is a problem but we need to understand which conversations are contributing to it.
For example, the Wallboard may tell us that the longest wait time has become critical. We can then move to the Queues view to identify the affected queue and finally use Conversations to inspect the actual conversations waiting in that queue.
The supervisor experience is therefore not only about monitoring aggregated numbers. It also allows us to move closer to the individual interactions behind those numbers.
Assisted Service
The Assisted Service view provides a more detailed analysis of human-assisted contact center operations. Where the Wallboard is useful for quickly answering the question:
What is happening right now?
Assisted Service helps us dig further into:
How is assisted service performing across the selected period?
This can include performance trends, service metrics, and conversation insights. It gives supervisors a way to move beyond the immediate operational state and start looking at short-term performance patterns.
Assisted Service Operations Summary:
Overflow Metrics:
Self Service
The Self Service area gives supervisors visibility into automated customer journeys. This can include areas such as:
- Containment
- Escalations
- Self-service engagement volume
- Supported languages
- Customer intents
- Operational health
This is increasingly important in a contact center where Copilot Studio agents and other self-service experiences may handle a significant percentage of interactions before a human representative becomes involved.
For example, we may find that overall conversation volume has increased but representative workload has not.
The Self Service view can help us understand whether more conversations are being contained automatically.
On the other hand, a sudden increase in escalations could explain why a human queue is becoming overloaded.
Configuring the Metrics
With the understanding that not all organizations use the same thresholds in certain metrics, we have the ability to configure report thresholds. To do this, click on Configure:
You will see a list of the available metrics to configure. Select a metric, for example, Service Level:
Here we can define thresholds based on our business:
And also define specific queue thresholds:
The ability to see how the metrics will look through preview values is nice as well:
Final Thoughts
Real-time streaming analytics is an interesting addition to Dynamics 365 Contact Center because it changes the supervisor experience from viewing periodically refreshed reports to monitoring an operational system as it changes. We can start with the Wallboard, identify queues that need attention, drill into representatives and conversations, take operational action, and then immediately watch the effect on the contact center.
The distinction between live metrics and aggregated metrics is especially important, as is understanding that this is a new streaming supervisor experience rather than simply another version of the existing Omnichannel real-time reports. If you are already using Dynamics 365 Contact Center, this is definitely a feature worth testing in a suitable environment while it is in preview.
