Skip to main content

Use Case Scenario

note

This scenario continues directly from CRM Use Case Scenario, where AUS Bank built a single Service Cases register across card disputes, fraud alerts, loan servicing, and account access. Nothing in that scenario changes here. This page picks up right where its Outcomes left off.

Background​

The Service Cases register solved AUS Bank's reporting problem: every card dispute, fraud alert, loan servicing request, and account access issue now lived in one place, with a resolution-time export the compliance team could pull in minutes instead of a week. But a register is only as good as what happens after a case is logged into it, and six months into the rollout, a second pattern emerged.

"We solved the reporting gap, but we traded it for a quieter one. A fraud alert could sit untouched for an entire shift if the agent it was assigned to was out sick, and nobody would know until the case showed up as overdue on next week's export. By then the customer had usually called back, angry, asking why nobody had followed up."

Monica, Contact Center Director

The register told Monica's team what had happened. It did not yet tell them what needed attention right now. That gap was the opening for Automations: the same Service Cases module already configured for the CRM scenario, now wired to two workflows that watch it continuously instead of waiting for someone to look.

Alerting the Agent the Moment a Case Lands​

The first workflow reacts the instant a case is created, rather than waiting for anyone to notice it. Monica's team built it as an Event-Based Workflow, since nothing about it runs on a timer: it exists entirely to respond to one specific thing happening in Service Cases.

  1. Trigger: On Service Cases Created
  2. Action: Send Email Notification to Assigned Agent
  3. Condition: Is Case Type Fraud Alert?
  4. Delay: Wait the matching SLA window
  5. Action: Get Record Service Cases By ID
  6. Condition: Still unresolved?

The agent assigned to a case gets an email notification the moment it lands, before they have even finished their next call. The Condition step after that splits the SLA itself: a Fraud Alert gets a much shorter window than a routine Account Access request, since the two carry genuinely different urgency. The Delay node holds the workflow for exactly that window, then the workflow re-reads the case with a fresh Get Record Service Cases By ID action, since the agent may well have already resolved it, and checks its current state with a second Condition.

"A fraud alert and a password reset request are not the same kind of urgent," Monica says. "Giving them the same SLA either rushes the password reset or leaves the fraud alert waiting too long. Splitting the delay by Case Type was the obvious fix once we saw it laid out as a workflow."

Watching Every Open Case, Not Just New Ones​

The first workflow only ever watches a case it saw get created. A case that was already open before the workflow existed, or one whose single Delay check happened to land while an agent was mid-resolution and then stalled afterward, would never be looked at again. Monica's team built a second workflow to close that gap: a Schedule-Based Workflow that sweeps every open case once a day, independent of when each one was created.

  1. Trigger: Scheduled Trigger, once daily
  2. Action: Get All Records Service Cases By Filter (Status = Open)
  3. List: Processing Item Pattern: All
  4. Formula: DATE_DIFF(Created At, NOW)
  5. Condition: Age value computed?

A Scheduled Trigger fires once every night. The workflow pulls every case still marked Open with Get All Records Service Cases By Filter, then hands the full result set to a List node set to its All pattern, so every single open case is walked in turn rather than just the first or last. A Formula node computes each case's age with DATE_DIFF against NOW, and a Condition step after it checks that an age value actually came back.

warning

The Condition node's rule builder only offers string operators on a Formula field, with no greater-than comparison available. So this final check confirms that Case Age Value exists, not that it has actually crossed a numeric SLA threshold for that case's Case Type. A case one hour old and a case one month old both pass it the same way today. The full technical walkthrough in the companion Configuration guide covers this gap, and the workaround, in detail.

In practice this still closes most of the gap Monica's team was worried about: every open case gets surfaced to the escalation step nightly, rather than only the ones that happen to be aged past a specific threshold, so nothing stalls silently forever. Tightening it to a genuine age cutoff is the next refinement, once the platform exposes a numeric comparison operator on Formula output.

Routing an Escalation to the Right Supervisor​

Both workflows eventually reach the same problem: once a case is confirmed overdue, who actually gets the escalation? Service Cases spans four service lines, and Contact Center Supervisor is not one single person. It is a role held by Monica and three service line leads together. A plain Condition node branching on Case Type could route to the right person for four fixed values, but Monica's team wanted the logic to hold even as new Case Type values are added later, without reopening the workflow each time.

That is the gap a Script node filled: a short piece of JavaScript that reads the case's Case Type, looks it up against the service line leads, and returns the correct person to notify, falling back to Monica herself if a Case Type has no specific lead assigned yet.

"A Condition node is exactly right when there are two or three paths," Monica explains. "Once routing turns into a lookup table that is going to keep growing, the Script node is just the more honest tool for the job. We would rather maintain one small function than a Condition node with eight branches stacked on top of each other."

The escalation itself goes out as both an email and an SMS to whichever supervisor the script resolves, through the same Email and SMS Actions already used for the agent's initial alert.

Watching the Automations Run​

Two workflows now run continuously against live customer cases, which made visibility into their behavior a requirement from day one, not an afterthought added later.

  • Automation Monitoring is where Monica's team checks, at a glance, that both workflows are active and have run recently, rather than silently failing without anyone noticing, the same blind spot the register itself was built to close.
  • Execution History is where a specific run gets investigated, for example confirming exactly when a particular fraud alert's escalation fired, and which supervisor the Script node resolved it to.
  • Registry Health Check is where the team confirms the On Service Cases Created trigger is still correctly registered against the module, a check that matters more here than on most workflows, since the entire first workflow depends on that one registration staying intact.

Controlling Who Can Touch It​

An automation that pages a supervisor is only trustworthy if the escalation logic itself cannot be quietly changed by someone outside that decision. Monica's team restricted editing of both workflows, through Automations Permissions, to the same Contact Center Supervisor role already defined for Service Cases itself: Monica and the three service line leads. Service Agents can see that the workflows exist and read their Execution History for cases assigned to them, but cannot open the Designer and change a Delay window, a Condition, or the Script node's routing logic.

Outcomes​

Measured against the quieter gap the register's own success had exposed, the two workflows closed it within the first full week of running.

MeasureBeforeAfter
Time for an agent to learn a new case is assigned to themUntil next list refreshImmediate, by email
Cases discovered overdue only at the weekly exportRoutineEscalated same day
Stalled cases caught by the nightly sweep, first month0, no sweep existed23 caught and reassigned
People who can edit the escalation logicAnyone with Designer access4, Supervisor role only

For AUS Bank, the register answered what had happened. The two workflows now answer what needs attention right now, closing the loop the original regulatory review first exposed.

Where Each Part Is Documented​

Every stage of this scenario corresponds to one or more of the configuration guides already written for Automations. This page serves as the map between the narrative and the reference material. The pages below are where each capability is explained in complete, step by step detail.

GuideUsed in
Automations OverviewBackground
Automations Getting StartedBackground
Automations Workflow BuilderBoth Workflows
Event-Based WorkflowsAlerting the Agent
Schedule-Based WorkflowsWatching Every Open Case
Trigger CategoriesBoth Workflows
Action CategoriesBoth Workflows
Condition NodeAlerting the Agent
Delay NodeAlerting the Agent
List NodeWatching Every Open Case
Formula NodeWatching Every Open Case
Script NodeRouting an Escalation
Automation MonitoringWatching It Run
Execution HistoryWatching It Run
Registry Health CheckWatching It Run
Automations PermissionsControlling Access