Automation Risk Checklist
A checklist to identify what could go wrong quietly—before you automate something that creates hidden failure.
Interactive worksheet
Nothing is sent anywhere. Your inputs stay in your browser unless you choose to copy or download a summary.
Flag risks in plain English. The goal is not “perfect automation.” The goal is to prevent quiet failure and unclear accountability.
This is internal guidance only. It’s based on the risks you flagged—not on “how advanced” the automation sounds.
Preview summary (for copy/paste)
Automation Risk Checklist — Summary Estimated complexity (qualitative): Low Operational risk: - (No risks flagged.) Data risk: - (No risks flagged.) Reputational risk: - (No risks flagged.) Compliance / accountability risk: - (No risks flagged.) Human trust erosion: - (No risks flagged.) Reminder: this is a quiet‑failure checklist, not a feature list. If you can’t name an owner and a fallback, pause.
This preview is local to your browser.
This checklist is designed to slow decisions. If any “serious risk” items are unclear, consider keeping a human in the loop longer.
What this helps you decide
- Surface silent failure risks before they become customer-facing.
- Clarify accountability: who notices, who owns, who can override.
- Decide where automation should stop and humans must remain.
When to use it
- You’re automating intake, follow‑ups, routing, or reporting.
- Failure would not be obvious immediately (quiet drift).
- You operate in a compliance-sensitive or reputation-sensitive environment.
The framework
Operational risk
- If this fails silently, who notices—and how quickly?
- What is the manual fallback if automation stops today?
- What’s the maximum acceptable delay before a human intervenes?
Data risk
- What data is required, and what happens if it’s missing or wrong?
- What assumptions does the automation make about field formats and completeness?
- How do you audit inputs/outputs without digging through logs for hours?
Reputational risk
- Could the automation send the wrong message to the wrong person?
- Could it look careless or disrespectful in a relationship-based context?
- What is the ‘worst plausible’ customer experience this could create?
Compliance / accountability risk
- Who is accountable for decisions and communications?
- Is there a clear override point before an irreversible action happens?
- Can you explain and justify outputs if challenged?
Human trust erosion
- Will this make the team trust the system more—or distrust it quietly?
- Will it reduce mental load, or add ‘babysitting’ work?
- What happens when it’s wrong—do people have a safe way to correct it?
Common mistakes
- Only testing “happy paths” and ignoring edge cases.
- Shipping automation without a defined owner and escalation path.
- Treating monitoring as optional (it isn’t).
What this does NOT answer
- Whether you should automate a judgment-heavy or moral decision (you shouldn’t).
- Which tool is best; this is about failure modes, not features.
This checklist is part of our audit process. If you want a calm, scoped risk review of a specific workflow, request an audit.