Description
Also known as an AI system impact assessment template, this document walks through affected stakeholders, harm severity, and mitigations for a single AI system or use case.
The AI Impact Assessment Template gives you a structured way to evaluate the risks, affected stakeholders, and mitigations of an AI system before it goes live — the document regulators and auditors increasingly expect to see completed for any higher-risk AI use case.
Built to the standard a Big-4 advisor would deliver, this editable Word document includes a worked example alongside the blank template, and is aligned to ISO 42001 and the EU AI Act‘s impact-assessment expectations for high-risk AI systems.
What’s included
- Fully editable Word document (.docx)
- Structured assessment sections: purpose, stakeholders, risk rating, mitigations, sign-off
- A completed worked example alongside the blank template
- Aligned to ISO 42001 and EU AI Act high-risk AI requirements
Who it’s for
Compliance officers, CTOs, and consultants who need a defensible, ready-to-use impact assessment without drafting one from scratch.
Key features of this AI impact assessment
Most teams skip impact assessment entirely, or write one so generic it wouldn’t survive scrutiny from a regulator or a client’s procurement team. This template forces the specific questions that matter — who’s affected, how severely, and what’s actually being done to mitigate it — with a worked example showing what a completed, defensible assessment looks like.
It’s designed to sit downstream of your AI Use Case Evaluation Template if you use one: use cases that clear initial scoring get the full impact assessment treatment here before build sign-off.
Unlike a generic risk assessment template, this one is scoped to the specific harms AI systems introduce — bias, opacity, and automated decision-making — that a standard project risk register doesn’t ask about.
Frequently asked questions
What file format is this in?
A fully editable Microsoft Word (.docx) document, ready to add your branding.
Do I need to complete one for every AI system?
It’s designed for higher-risk AI use cases — the ones with real impact on people’s rights, safety, or opportunities. Lower-risk tools may not need the full treatment.
Does the worked example count as legal advice?
No — it’s an illustrative example to show format and depth, not a substitute for jurisdiction-specific legal review.
Instant digital download. Delivered as an editable Word document. 7-day guarantee — if the document doesn’t fit your needs, let us know.
Related toolkits: this pairs with the AI governance framework template that it sits under, the AI system inventory template to track which systems still need assessing, and the AI decision matrix to assign sign-off once risk is rated. If your AI use involves automated decision-making under Australian privacy law, pair this with the AU Privacy Act ADM Transparency Pack for APP 1.7 compliance.
