Romulusby New OdysseyTry demo
← Romulus overview

Use cases

Where durable execution helps.

Explore how three agent workloads recover and what their workflows need to preserve along the way.

◈   Support operationsExample

TKT-2418 · Billing question

  • Ticket receivedDone
  • Customer context loadedSaved
  • Priority scoredRecovered
  • Queue assignmentRouted
Priority
High
Queue
Customer support

01 / Customer operations

Keep a ticket moving after a crash.

Recover an interrupted support workflow with its customer context and routing decisions intact.

Explore customer operations
◈   CRM synchronisationExample

OP-4C01 · Update customer record

  • Customer record resolvedSaved
  • Update preparedSaved
  • Provider responseTimed out
  • Operation lookupApplied
Operation ID
op_4c01
Write attempts
1

02 / CRM synchronisation

Recover a write after the response disappears.

Look up a customer record update by its original operation ID and recover the confirmed result.

Explore crm synchronisation
◈   Job operationsExample

JOB-B772 · Monthly reconciliation

  • Job acceptedSaved
  • Correlation ID recordedSaved
  • Completion signalReceived
  • Report retrievedDone
Event
report.completed
Correlation
job_b772

03 / Background jobs

Wait for the result without holding a worker.

Keep the job ID and continuation point while an external process runs. Resume on its completion signal.

Explore background jobs

Choose what to save.

Customer operations
Save the ticket context and the routing decision before writing to the CRM. If the write times out, reconcile it with the provider before submitting a second update.
CRM synchronisation
Save the intended field update and its operation ID. If the response disappears, look up the existing result before deciding whether to retry.
Background jobs
Capture the job ID and expected completion event before releasing the worker. Give the wait a deadline and define how to handle duplicate completion events.