Template Library

AI Risk Register Template (FMEA for AI)

An Excel workbook for AI risk: failure modes of AI uses rated by severity, occurrence, and detection before and after controls, the RPN and its band, a severity review flag, actions with owners and due dates, and a dashboard.

The Register lists each failure mode of an AI use with its effect, cause, and existing controls. You rate severity, occurrence, and detection from 1 to 10 using the scales on the Settings sheet, and the risk priority number (RPN) and its band calculate. Any severity at or above the flag level is marked for review whatever its RPN. You then record the recommended action, its control type, owner, and due date, and re-rate the three scores with the control in place.

The Settings sheet holds AI-specific rating scales, editable RPN bands, and a library of common AI failure modes: wrong output stated confidently, misclassification, data leakage, manipulated input, silent vendor change, over-reliance, bias, and runaway action or cost. The Dashboard shows risks by band before and after, the average RPN reduction, the top five residual risks, overdue actions, and actions by control type.

Example risks for complaint classification and a work-order agent at an illustrative plant are included so you can see the formulas working before you replace them.

Download AI Risk Register Template (FMEA for AI) Back to Templates

What Is Included in the Workbook

Sheet	Purpose	What Teams Capture
Start Here	Orientation and workflow	What the workbook does and a numbered workflow with a link to the sheet for each step
Dashboard	Live summary of the register	Risks rated, High before and after, average RPN reduction, severity flags, overdue actions, risks by band, the top five residual risks, and actions by control type
Register	The AI risk register (Excel Table)	AI use, failure mode, effect, cause, controls, S, O, D, RPN, band, review flag, action, control type, owner, due date, status, re-rated scores, overdue flag
Settings	Scales, bands, and lists	Severity, occurrence, and detection scales for AI, RPN bands, the severity flag level, the as-of date, and a library of common AI failure modes

Key Features Inside the Template

FMEA Applied to AI

The familiar severity, occurrence, and detection ratings, with scales worded for AI uses.

Severity First

A review flag marks high-severity failures whatever their RPN, so a low RPN cannot hide a serious risk.

A Library of AI Failure Modes

Twelve common AI failure modes to start the analysis.

Before and After

Re-rate with each control in place and see the RPN reduction and the residual risks.

Best Use Cases

Assessing the risks of a new AI use before a pilot
Keeping a register of AI risks for a governance group
Preparing for an AI governance audit or a customer questionnaire
Showing that controls reduce risk
Teaching FMEA for AI

How to Use the Template Effectively

Check the rating scales and bands on the Settings sheet, and edit them to fit your organization.
List the failure modes for one AI use at a time, using the library as a prompt. Include the people who use it.
Rate severity, occurrence, and detection for the process as it runs today.
Choose controls, starting with those that prevent the failure, then those that catch it. Record the control type, owner, and due date.
Re-rate with the control in place and check that the residual risk is acceptable.
Review the register after any incident, change, or audit finding.

Who Should Use This Template

AI and process governance groups
Quality, risk, and compliance teams
Process owners adopting AI
Students and trainees learning FMEA

Common Mistakes to Avoid

Rating a use as if every control always works
Treating the RPN as a measurement and ignoring severity
Relying on training and a reminder to reviewers as the only control
Never re-rating after controls
Doing the analysis without the people who use the AI

Important Notes

Educational template. It supports an organization's planning and records for the use of AI. It is not legal, regulatory, security, or professional advice, and it does not replace your organization's policies, the laws and standards that apply to you, or qualified people. Check requirements with qualified people. Example data are illustrative.