Intraday Performance Report

Status: Public Preview — behaviour may change.

What it does

A grid that sets planned against actual volume, handle time, service level and staffing, interval by interval, for the current day or a past one, split by channel and queue. Supervisors and workforce managers use it to spot coverage gaps while there's still time to move breaks, extend shifts or offer voluntary time off, and to review afterwards how closely the day matched the plan.

Key facts

Prerequisites: the Workforce Management for Customer Service package installed in the Power Platform admin center, Intraday Performance enabled under Workforce Management in Copilot Service admin center, and a shift plan with published shift bookings, a capacity plan and a forecast scenario behind it.
The data chain is planning group, shift plan, capacity plan, forecast scenario, then queues for each selected channel. The planning group sets the interval length; the shift plan supplies Reps Scheduled; the capacity plan supplies handle time, service level target, target time to answer, maximum occupancy and shrinkage; the forecast scenario supplies predicted volume and the available queues.
Overall Staffing is the group to read first. Over/Under is Reps Scheduled minus required representatives including shrinkage: negative means understaffed, zero matches, positive is over, a dash means no requirement for that interval.
Fcst. Vol. is the only true forecast column. Planned handle time, service level, time to answer and occupancy are configured values from the capacity plan, not predictions.
Actual columns refresh every 10 seconds. Forecast, planned, required and scheduled columns refresh every 5 minutes.
Actual volume counts every record that arrives in the interval, abandoned ones included, attributed by arrival time and not by when it was answered.
Service level and time to answer columns aren't available for case records.
History only reaches back to the date the report was enabled. The grid exports to Excel with your filters.

When to use / skip

Use it for live floor management in a centre that already runs forecasting and published shifts. It answers "is this half hour covered?" faster than building your own report, and the negative-run pattern is useful in the post-day review too.

Skip it if you're not using the workforce management package, or if schedules aren't published. With no shift bookings behind it, the staffing half of the report has nothing to show.

Configuration decisions

Interval length on the planning group. The grid rows follow it, so shorter intervals mean more rows and noisier actuals. Choose it with whoever will actually read the grid.
Capacity plan parameters. Handle time, service level target, time to answer, occupancy ceiling and shrinkage drive the required-representatives numbers. Wrong parameters produce a confident-looking gap that isn't real.
Which channels and queues people view by default. Deselecting a queue removes its contribution from Overall Staffing, so a partial view understates the requirement.
Who gets access and when you enable it. History starts at enablement, so turn it on before you need the past-day comparisons.

Gotchas

Deselecting queues changes the totals, not just what's visible. Two supervisors with different filters can reach different conclusions about the same interval.
The interval in progress changes until it closes. Reading its service level mid-interval and acting on it is acting on partial data.
A large negative that repeats across consecutive intervals usually means the schedule doesn't match the shape of the forecast. It's a shift-design problem, not a one-interval problem.
If there's no active planning group, the Planning group filter disappears and you pick the shift plan directly. Training screenshots from one tenant won't match another's.
Forecast scenario drives the available queues. A queue missing from the list is a forecast scenario question, not a report defect.

Consultant notes

Walk supervisors through the Over/Under sign convention and the shrinkage line in the first session. Reading required reps without shrinkage is the common misread and it makes everything look overstaffed.
Before anyone trusts this report, sanity-check the capacity plan against the last month of real handle times. Planned values that were guessed at go-live are the usual cause of "the report is wrong".
Agree what supervisors are allowed to do on a negative interval: move breaks, extend shifts, offer time off. The report tells them there's a gap; it doesn't settle who has the authority to act.
It's a preview. Run it alongside whatever supervisors use today for a few weeks before retiring the old view.

Preview feature — worth revisiting when it reaches general availability.