Real-time analytics for record routing

What it does

A live supervisor dashboard for work routed as records — cases and other entity records going through unified routing — rather than for conversations. It sits in Copilot Service workspace under Service and shows current queue backlog, in-flight work items and rep status without waiting for the nightly analytics refresh.

Key facts

Seven reports: Summary, Queue, Agent, Ongoing work items, Backlog work items, Intent group and User group. Summary is the landing page.
Intent group and User group are in preview. The other five are generally available.
Summary, Queue and Agent auto-refresh continuously, and the supervisor can pause and resume that refresh. Ongoing work items and Backlog work items do not auto-refresh — they need a manual refresh.
The Queue report only covers entity record queues. Conversation queues are out of scope here; those belong to the omnichannel real-time dashboard.
Access is via the CSR Manager role by default, and the underlying report entity is msdyn_dataanalyticsreport_ur_recordrouting_rt. Custom roles need read on that entity.
Row-level scoping honours Dataverse security, including business-unit-level permissions, so a supervisor in one business unit sees their own slice rather than the whole organisation.
Filters available are Modified on, Agent, Agent status, Agent skill, Queue, Work item status, Work item skill, Time zone, and optionally Business Unit.
It is customisable in the embedded Power BI editor along with the other supported report sets, and supports automatic page refresh at 5 or 10 second intervals when you go down the custom data model route.
Supervisors can save personalised bookmark views of their filter state.

When to use / skip

Use it wherever unified routing is actually distributing cases and you have supervisors whose job is intraday intervention. That is the whole point: seeing that a skill-based queue has quietly built a backlog at 11am, rather than reading about it tomorrow. If the client's supervisors sit on the floor and rebalance work during the day, this is the report they will live in.

Skip it if routing is not unified routing. Plenty of Customer Service implementations still assign cases through basic queues, workflows or manual pickup, and this dashboard has nothing to show for that — the metrics come from the routing engine, not from the case table. Also skip if supervision is genuinely retrospective: a team that reviews performance weekly gets more from the historical Agent report and will find a live dashboard a distraction.

Do not reach for it as a substitute for the omnichannel real-time dashboard on a project that has chat and voice. They are different report sets over different data, and a supervisor covering both will need both.

Configuration decisions

Whether record routing (unified routing for records) is in scope at all — this dashboard is meaningless without it, so the decision is made upstream in the routing design, not here.
Which supervisors get CSR Manager versus a custom role, and whether you grant read on the record-routing report entity to team leaders as well as managers.
Whether to scope by business unit. Getting the business unit hierarchy right in Dataverse is what makes the scoping work, and retrofitting business units later is painful.
Whether the Business Unit filter needs to be exposed, which pulls in data model customisation as a dependency.
Whether to adopt the preview Intent group and User group reports now or wait for GA — relevant if the client is regulated or has a no-preview policy.
Which skills and queue structure the Queue and Agent reports will group by. Ugly queue naming shows up directly on the supervisor's screen, so agree naming conventions before the reports go live.

Gotchas

The two work-item reports not auto-refreshing catches everyone. A supervisor leaves Backlog work items open all morning, sees a stale number and reports a data bug. Point it out in training, once, loudly.
Preview reports can change shape or naming between waves. If you build supervisor process around Intent group today, budget for rework.
Business-unit scoping means what the consultant sees in a system-admin context is not what the supervisor sees. Always sanity-check the dashboard signed in as a real supervisor account before declaring it done.
Agent status metrics reflect presence in the routing system. If the client has not adopted presence properly, or reps leave themselves Available while at lunch, the Agent report is confidently wrong and supervisors will lose trust in the whole analytics suite.
Continuous refresh on a dashboard left open all day is not free. On environments with heavy routing volume, expect questions about client performance on older supervisor hardware.
Real-time reports show the current picture only. There is no history here — if a supervisor wants to explain what happened at 3pm yesterday, they need the historical reports, and those are on a 24-hour cycle.

Consultant notes

Demo this to supervisors, not to the sponsor. The value is operational and only lands with someone who has actually had to rebalance a queue mid-shift.
Pair the rollout with a presence and status discipline conversation. The dashboard is only as honest as the reps' status, and that is a process problem you cannot fix with configuration.
Before go-live, log in as each supervisor persona and confirm both the report entity privilege and the business-unit slice. Two separate failure modes, both silent, both look like "the dashboard is empty".
Push back if someone asks for real-time analytics on a project with no unified routing. The answer is a routing design change, not a reporting change, and it is far cheaper to say so in discovery.
Set the expectation that preview reports are for evaluation. Written into the design document, it saves an argument later.

Worth revisiting when Intent group and User group reach GA, or if the refresh behaviour on the work-item reports changes.