Use the Dashboard to understand support workload, service and knowledge

Updated Wednesday, October 28, 2026 · Marcel Gravelle

The Supportbench Dashboard brings case workload, response performance, customer feedback, team activity and knowledge reuse into one configurable workspace. Use it to spot a question that needs attention, then investigate the underlying cases or records before deciding what to change. In B2B support, that connection matters: a small number of complex cases can carry substantial customer commitments even when overall volumes look healthy.

At a glance

For: support supervisors, operations leads, knowledge managers and agents with Dashboard access. Applies to: Supportbench Core.

Start here: Dashboard

Choose the decision your Dashboard should support · Check access, scope and freshness first · Open and configure your first useful Dashboard · Match picker names to the displayed cards · Arrange, resize and recover your layout · Cases: understand volume, status and resolution · Experience: separate feedback, predictions and service measures · Team: distinguish workload, contribution and presence · Knowledge: connect activity to verified reuse · Run a practical 15-minute support review · Troubleshoot empty, inconsistent or unsaved results

Choose the decision your Dashboard should support

Operating review	Start with	Follow through
Morning workload review	Case Volume, Case States and Top Issues.	Identify the queue, customer commitment or repeated symptom needing an owner. Open the appropriate case view for the actual work.
Service-quality review	First Response Time, SLA Performance and Customer Satisfaction.	Check which population and period each card measures, then review representative cases and customer feedback.
Team capacity conversation	Agent Activity and Agent Status, alongside the case mix.	Discuss workload and availability with context. Activity counts and online duration do not measure the difficulty or quality of the work.
Knowledge improvement review	Top Articles, Article History and Top Article Contributors.	Inspect the articles, their current accuracy, audience and reuse opportunities. High use is a reason to maintain a guide carefully, not proof that it is correct.

Choose a small set of cards that supports your regular decisions. Adding every widget can make the important exception harder to notice. A supervisor may need workload and commitments first; a knowledge lead may put reuse and maintenance first. These are layout recommendations, not mandatory product presets.

Check access, scope and freshness first

Your agent role must permit Dashboard access. An account can also restrict an agent to their own Dashboard data. Team, Agent and Division choices are constrained by the permitted scope; changing a selector does not grant additional record access. Ask an administrator to review the specific reporting permission when the page or a needed choice is unavailable.

A Dashboard is current when requested

Core explicitly shows Current when requested · no automatic refresh. The page is not a continuously updating wallboard. Choose Refresh when you need a new snapshot; changing filters or a card’s supported date controls also reloads the relevant data. Record the scope and refresh time when using a screenshot in a meeting.

There is no universal date selector that makes every widget use the same reporting period. Most cards use their own fixed rolling windows. Case Volume and Agent Status have their own controls. Read the card and the reference below before comparing numbers.

Open and configure your first useful Dashboard

Open Dashboard from the agent navigation. If it is unavailable, confirm your role’s reporting access.
Review Team, Agent and Division. Select the operational scope you intend to discuss. Changing Team refreshes the eligible Agent choices; recheck Agent afterward.
Choose Add widgets. Use Search widgets… or the category filter: All categories, Cases, Experience, Team or Knowledge.
On the required card, choose + Add widget. The picker marks an included card ✓ Added. Add the other cards needed for the review, then choose Done.
Arrange the cards using their drag grip or the movement actions in the widget menu. Put the question you check first near the top. Use the supported size controls to make charts readable.
Choose Refresh, then check that each card displays data or a meaningful empty state rather than a loading or error message. Review each card’s period and unit.
Leave and reopen Dashboard to verify that your personal layout was saved. Your layout belongs to your signed-in agent, even when the Agent filter displays another person’s data.

Core maintains your personal widget layout; this procedure does not create named dashboards for different customers or teams. Changing Team or Agent changes the data scope, not the owner of the layout. Use a saved case view or scheduled report when the review requires an explicitly defined customer-specific list.

Core Add widgets dialog with Search widgets, category filter, widget descriptions and Added or Add widget controls. — Use Search widgets or a category to find a card. Added marks a card already in your personal layout. Picker names can differ from the shorter titles displayed on the Dashboard; the mapping below connects them.

Match picker names to the displayed cards

Some names in Add widgets are longer or use the established catalogue wording. Use this mapping when the name in a screenshot or card does not exactly match the picker.

Add widgets name	Displayed card
Case volume and satisfaction	Case Volume
Case same day closure	Same Day Closure
Top 10 issues	Top Issues
Cases created by activity type	Activities Created
Activity word cloud	Keyword Cloud
First response	First Response Time
Customer NPS score	Net Promoter Score
Sentiment scores	Customer Sentiment
Case status summary	Case States
Knowledge base history	Article History

The remaining names correspond directly apart from capitalization: Customer satisfaction, Agent activity, SLA performance, Top article contributors, Top articles, Agent status and MTTR. Read the description and category in the picker to confirm the intended widget.

Arrange, resize and recover your layout

Task	How to do it	What to verify
Move a card	Use its drag grip or widget-menu actions Move left, Move right, Move up and Move down.	The reading order supports the review on the screen size you normally use.
Change card size	Use Make wider/narrower/taller/shorter or the southeast resize handle.	Labels and values remain legible; enlarging a card does not change its underlying data.
Remove a card	Open its widget menu and choose Remove widget.	The card leaves your layout. You can add it again from Add widgets.
Save ordinary layout edits	Allow the automatic save after editing; it is debounced by approximately two seconds.	Reopen Dashboard after the save completes to confirm persistence.
Recover a failed save	Use Retry save when the page reports a save failure.	Check the result before closing the page or repeating layout changes.
Resolve another saved version	Choose Load saved layout to use the stored version, or Save my layout over this version only when you intend your current layout to replace it.	Review which version you want to keep; concurrent tabs can otherwise overwrite a deliberate arrangement.
Handle a repaired old layout	Review the displayed arrangement and use Accept displayed layout repair when offered.	Confirm the repaired result before treating it as the new saved layout.

A layout conflict is different from a statistics error. The first concerns the arrangement you are saving; the second concerns the data a card could not load. Keep those recovery steps separate.

Cases: understand volume, status and resolution

Widget	What it measures	How to use it responsibly
Case Volume	Cases opened and closed, with estimated opening backlog and ending inventory by local day. Default range is 14 days; presets include 7, 14, 30, 60, 90 and 180 days, with an optional start date.	Use the trend to identify sustained pressure. Backlog is an estimate and can be affected by incomplete reopen/lifecycle history. Read count and satisfaction axes separately.
Same Day Closure	For cases closed in the last 7 days, the share created and closed on the same local calendar day. Additional within-7-day and within-30-day closure measures use a 90-day closed-case population.	Same calendar day is not within 24 elapsed hours and is not first-contact resolution. A late-night case closed after midnight can fail the same-day condition despite a short duration.
Top Issues	Cases created in the last 7 days, grouped by their current issue and state, with Created, Open and Closed counts.	Look for a common problem to investigate. Current classification and state can change; the card is not a frozen history of how each issue was classified when created.
Activities Created	Seven-day activity-type mix, with overall activity totals for 7 and 30 days and a separate count of cases created over 30 days.	Understand the communication/work mix. An activity is not a case, and the separate case total is not an automatic conversion-rate denominator.
Keyword Cloud	Up to 35 phrases weighted by occurrences on issue-associated cases created in the last 7 days. Eligible closed cases can contribute; forum, Trash and merged records are excluded.	Use repeated phrases as leads for investigation. The words are not a sentiment verdict or clickable search filters.
MTTR	Weekly mean resolution duration, measured in whole hours, for cases created in the trailing 6 months; uses recorded resolution time or closure fallback. Weeks start Sunday.	Read the chart unit: it uses days when any weekly average exceeds 24 hours, otherwise hours. Averages and the case-created cohort can hide unusually long cases; review the distribution and complex customer work separately.
Case States	Current counts by case state for eligible non-Trash, non-merged cases with a positive status ID.	Find current concentrations of waiting or active work. This is a state snapshot, not a trend showing when cases entered each state.

Case-based widgets exclude Trash and merged records where their case population is built. Team attribution also matters: queued cases use their Queue’s team; unqueued cases use the owner’s applicable team membership. Therefore an apparently similar team label in another report does not guarantee an identical population.

Experience: separate feedback, predictions and service measures

Widget	Population and period	Interpretation
Customer Satisfaction	Received survey feedback for 7 days, 30 days and 3 months; separately shows binary AI-predicted CSAT from scored cases created in 7-day and 30-day cohorts.	A customer’s survey response and an AI prediction are different evidence. The AI cohort uses case creation, not the date a prediction was generated. Review them separately.
First Response Time	A 3-day distribution in bands up to 4 hours, over 4–8, over 8–12, over 12–24 and over 24 hours. Mean values use positive recorded first-response durations for cases created in 3-, 7- and 30-day cohorts.	Use this to investigate responsiveness, while checking which cases have a recorded positive value. It is not automatically the same population as every newly opened case or every SLA record.
Net Promoter Score	Valid 0–10 response bands and NPS over 7 days, plus rolling-month NPS and average score. The current compatibility calculation uses all received surveys in its denominator, including surveys without an NPS answer.	Do not recalculate the displayed value using only answered-NPS responses and assume it must match. NPS and mean numerical score are different measures. Check the received-survey and answered-question counts when reconciling another system.
Customer Sentiment	Eligible activities with positive recorded sentiment scores on cases created in 3-, 30- and 60-day cohorts; includes 3-day bands and cohort averages.	The cohort follows case creation, not activity creation. A case with several scored activities can contribute more than once. Sentiment is a signal to read the conversation, not a substitute for it.
SLA Performance	Response-SLA activity rows created in the last 3, 7 and 30 days with both actual and expected duration positive. A hit means actual duration is no greater than expected duration.	This measures qualifying response records, not the percentage of all cases satisfying every service commitment. Resolution promises and other obligations need their own review.

Do not join unlike percentages

A closure percentage, a response-SLA percentage and a satisfaction percentage describe different events and populations. Put their definitions beside a leadership summary. A combined “team quality score” needs an explicitly agreed methodology; the Dashboard does not make those measures interchangeable.

Team: distinguish workload, contribution and presence

Widget	Definition	Operational caution
Agent Activity	Seven-day counts for created cases, closed cases, activities and articles, alongside current open workload. Attribution varies: case creation can use creator/unqueued ownership; closed/current work uses unqueued ownership; activities and articles use their creators.	Do not read every column as work completed by one person under a single ownership rule. Compare the underlying records and specialist involvement before using counts in a performance discussion.
Agent Status	Online, Away and Total durations across the selected inclusive Start and End dates; defaults to the current local month. Intervals are clipped to the selected period, and an open interval runs to request time.	This is accumulated status duration, not a live presence roster or logged case effort. Division does not filter this widget. Stored interval inconsistencies are flagged rather than silently repaired.

A specialist may spend substantial time on one difficult customer issue while another agent handles many short requests. Discuss case complexity, customer impact, quality and available capacity alongside counts. Use recorded activity time for an effort question; do not substitute online duration.

Knowledge: connect activity to verified reuse

Widget	Definition	Next useful action
Top Article Contributors	Top 10 contributors ranked by article-created audit events over 30 days, showing 7- and 30-day counts. The displayed daily average uses the displayed top-10 total over 30 days.	Recognize contribution while also reviewing accuracy, useful reuse and maintenance. The daily average is not an average per contributor.
Article History	Daily audited article edits, case links and article creation from the local date 14 days ago through today. Today is a partial day.	Look for a healthy capture-and-improve process. These are events, not necessarily different articles or successful customer outcomes.
Top Articles	Top 10 articles ranked first by 30-day case links, then views. Eligibility includes articles with content or draft content outside Trash.	Open frequently reused content and check its status, audience and accuracy. Ranking does not certify that an article is reviewed, published or publicly accessible.

For an equipment failure affecting several customer organizations, a frequently linked resolution guide is a good candidate for careful review. Check supported models/versions, prerequisites, verification steps and escalation conditions. Keep private customer details in cases, and publish reusable instructions through the appropriate KCS workflow.

Run a practical 15-minute support review

Set the scope. Choose the intended Team and Division, check Agent, refresh and note the time. Decide whether this is a team workload review or a customer-specific review that needs a case view.
Identify one signal. Suppose Case Volume shows a growing estimated backlog and Top Issues shows repeated connection failures. Do not assume the chart has identified one common defect.
Read the evidence. Open a relevant case view and inspect representative cases, affected assets or versions, ownership, customer impact and the next promised update. Separate repeated wording from a verified shared cause.
Check service and customer context. Look at the qualifying response-SLA records and actual survey feedback. A healthy overall percentage can coexist with a high-impact customer escalation.
Assign the operational response. Record a named case/task owner, investigation scope and next checkpoint. Use the escalation process when the risk or dependency requires it. Moving a Dashboard card does not assign work.
Improve reusable knowledge. If investigation confirms a repeatable resolution, improve or create the appropriate reviewed guide. Check that the intended audience can find it.
Verify next time. Revisit the action and the same clearly described measures. Record changed filters, periods or classifications before interpreting a difference as improvement.

The benefit is a shorter path from a broad signal to accountable work and reusable knowledge. Measure that improvement in your own process; a widget does not guarantee a particular productivity or satisfaction gain.

Troubleshoot empty, inconsistent or unsaved results

Symptom	Likely checks and corrective action
The figures do not change while the page is open	Choose Refresh. Core does not automatically refresh this Dashboard.
A widget says statistics could not be loaded	Use Retry this widget or the widget-menu Retry action. Record the card, scope and time if it repeatedly fails; a failed load is not zero performance.
A card is empty or shows a dash	Check the period, filters, role and whether qualifying records exist. Missing/not-applicable data is not necessarily a numerical zero.
Agent Status does not change with Division	That widget does not apply Division filtering. Use its Team/Agent scope and date controls appropriately.
Two cards disagree	Compare date windows, event versus case-created populations, current versus historical state, attribution and units. Do not force unlike metrics to reconcile.
An exported case list differs	Compare the export timestamp, case filters, access and measured event. Dashboard aggregates and live lists need not have the same population.
The layout changed in another tab	Use the offered layout-conflict choices after deciding which saved/current layout should remain.
A card is unavailable	Search the picker and category filter, check whether it is already added, then review feature/role availability.
A metric appears implausible	Record scope, time and the specific definition; inspect a small set of underlying records and report the reproducible discrepancy. Do not change case dates solely to make a chart look correct.

Was this article helpful?