How to Detect Missing Automation Work With a Reconciliation Check

Know what an execution status cannot prove

SitePoint describes workflows that appear sound in a demo but later fire twice, fail to fire or receive a payload shaped differently from the tested one. Treat execution status as one piece of evidence, then define the business outcome that each eligible item should produce.

Sources: How to Design Automation Triggers That Don’t Fail Silently in Production.

SitePoint identifies duplicate delivery, silent failures, schema drift and ordering assumptions as recurring production trigger failure modes. Use the readiness page to define a clear expected outcome before attempting to reconcile completed work.

Sources: How to Design Automation Triggers That Don’t Fail Silently in Production.

• A trigger can fire twice or not at all.
• Input shape can change after testing.
• A run status does not prove the destination result.

Build the reconciliation worksheet

A reconciliation worksheet compares each eligible source item with the destination outcome expected for that same item. Use one row per source item with these columns: source ID, expected outcome, eligibility time, destination evidence, and match status with review note.

In the destination-evidence column, record the destination ID, observed outcome and confirmation reference. Mark the row matched only when the destination evidence corresponds to the source item and intended outcome.

• Keep one row per expected source item.
• Use the same stable identifier on both sides.
• Record destination evidence, match status and review notes.

Match, wait and classify exceptions

Octacer recommends a stable key to check whether work was already completed before doing it again. After the business-chosen processing window, match source and destination records by stable identifier before treating an item as missing.

Sources: Workflow Best Practices for Reliable Automations | Octacer.

Classify exceptions as missing outcome, duplicate outcome, conflicting outcome or untraceable outcome, then assign review rather than replaying automatically. Stop and investigate when evidence is ambiguous because an unrecorded acknowledgement or delayed destination update can make a completed item look absent.

• Match with a stable key.
• Wait for the chosen processing window.
• Classify before replaying work.

Use independent confirmation without overstating it

Hallermann Consulting observed one labelled synthetic enquiry marked sent on its first recorded scheduled-worker attempt, with no recorded delivery error and a matching inbox receipt; n=1, period 2026-10-03 to 2026-10-03, verified from outbox state and message-ID matching. Use a comparable independent receipt, destination record or approved downstream status as outcome evidence when it is available.

Sources: One queued synthetic notification delivered by a scheduled worker.

Measure an Automation Pilot can help choose outcome evidence without relying only on a workflow activity count. That single observation does not prove delivery rates, latency, uptime, retry behaviour, exactly-once delivery or broad error-path coverage.

• Use a separate confirmation where possible.
• Keep the sample limitation visible.
• Do not convert one observation into a reliability claim.

Choose heartbeat or item-level reconciliation

Use a heartbeat or overdue-completion check when work is expected on a regular schedule and an absent run itself needs attention. Use item-level source-to-destination reconciliation for event-driven work because a quiet period can be legitimate while an individual eligible item is still missing.

Use the pilot-controls page to measure whether the completed workflow produced the intended business result after monitoring is installed. Confirm the provider’s current prerequisites in its documentation before changing schedules, queries or replay behaviour that could affect a live service.

• Use heartbeats for regular expected work.
• Use item matching for irregular events.
• Keep both checks tied to a defined outcome.

Source-to-destination reconciliation worksheet

Keep one row for each expected source item. Record its expected result, when it became eligible, the destination identifier and observed evidence, then state the match decision and accountable review note.

The examples are hypothetical worksheet entries, not measured workflow results. Choose the processing window and review frequency according to the workflow’s business consequences.

When this is not the right approach: Do not rely on a reconciliation worksheet alone for a workflow that requires immediate human approval or where delayed review could cause unacceptable harm. Use a controlled human process until safe automated exception handling is defined.

Related guidance

• Logging and Alerting for Small-Business Automation: A Practical Monitoring Plan
• An Actionable Automation Failure Alert Template for Small Businesses