From reading to action
A practical path through this guide.
Select a step to explore what to do and why it matters.
Inventory the estate
List workflows, schedules, owners, connected applications and criticality.
An automation maintenance plan should define what is monitored, who investigates failures, and when support is available. A monthly retainer without these boundaries leaves both sides guessing. Begin with an inventory of the workflows covered by the agreement.
Define the supported estate
For every workflow, record its business purpose, owner, connected applications, expected schedule, credential owner, and criticality. Include links to run history and recovery instructions. Mark experimental or retired workflows so they do not receive the same attention as revenue-critical integrations.
Agree how a new workflow enters the plan. Otherwise a retainer that started with two integrations can gradually inherit twenty. Vendor subscription fees, major redesigns, and emergency work outside agreed hours should have an explicit treatment.
Make the routine work visible
A useful maintenance cycle checks failed runs, unusual completion delays, credential warnings, queue growth, and unexpected usage. It also verifies outcomes with a small reconciliation sample. A dashboard can look healthy while a mapping sends leads to the wrong team.
Produce a short monthly report covering incidents, records recovered, recurring causes, usage changes, and proposed improvements. Distinguish “no alerts received” from “health verified.” The report should explain what evidence was checked.
Agree response expectations
- Support days, hours, and timezone.
- How incidents are reported and which channel is monitored.
- Severity definitions tied to business impact.
- Response targets and the difference between acknowledgement and resolution.
- Access requirements and escalation contacts at the client.
Avoid promising resolution times for vendor outages you cannot control. Instead, agree an investigation and communication process, plus temporary workarounds where feasible. Identify when the client must approve pausing a workflow or replaying a backlog.
Maintain recovery capacity
Periodically rehearse a controlled failure and restore. Verify that the backup configuration exists, the responsible person can access it, and replay instructions still match current fields. Remove former staff access and rotate credentials through the organization's normal process.
Review the scope when transaction volume, connected tools, or business criticality changes. A flow that once handled occasional enquiries may need stronger monitoring after it becomes the main sales intake path. Maintenance should adapt to actual usage rather than remain a static checklist.
Further reading
Use automation monitoring metrics to choose the report's measurements. The maintenance offer starts with agreed support hours and response expectations.
A few practical answers
Frequently asked questions
What should an automation maintenance agreement cover?
Name the supported workflows, owners, monitoring tasks, support hours, reporting channel and severity definitions. Explain how additional workflows enter the agreement.
Is acknowledgement the same as incident resolution?
No. Acknowledgement confirms someone has received the issue. Investigation and resolution depend on the cause, available access and external providers; distinguish these expectations.
What should an automation support agreement say about response times?
State the covered workflows, support hours and timezone, severity levels, escalation contact and response target. Distinguish an acknowledgement from a repair, identify exclusions and provider dependencies, and agree the reporting cadence and change allowance before the plan starts.
