Risk identification that understands your industry
RiskNote uses Claude to suggest risks relevant to your organisation — your industry, size, goals, and past incidents. No generic lists. You approve or reject every suggestion.
Why AI in risk management?
Traditional risk identification starts with a blank Excel template and a brainstorm. The output is generic: ”data breach”, ”employee turnover” — disconnected from your actual business.
RiskNote pairs your organisation profile (industry, size, role, goals, challenges) with an AI model specialised on ISO 31000. The result: five concrete, industry-specific risk suggestions as a starting point — not a replacement for your expertise, but a jumpstart.
What makes AI analysis better than a template?
Industry-specific terminology
A healthcare org gets ”Patient data exposure via legacy EHR integration”, not ”Risk of data breach”. A municipality gets e-service and citizen-data risks. A manufacturer gets supply chain and production stoppage.
Five distinct suggestions per run
The AI sees what risks you already have and suggests new ones that differ in type, area, and level. No duplicates, no semantic overlap.
You stay in the driver's seat
GDPR Art. 22 and EU AI Act Art. 50 require meaningful human control. Every suggestion must be approved manually before it enters the register. AI helps; the decisions are yours.
No PII sent to the AI provider
RiskNote sends only the risk note's name, description, areas, and your organisation context. Never email, user names, or data from other customers.
Language follows your content
Write in Swedish, get Swedish suggestions. Write in Finnish, get Finnish. Detection runs on the risk note's own content, not your user preference.
How to use AI analysis
1. Fill in your organisation profile once
Industry, size, role, goals, challenges, core assets, regulatory frameworks, suppliers, past incidents. This context is used in every AI run going forward.
2. Create a risk note and describe scope
Give the note a name and description. For example: ”Introducing an AI-based e-service for student absence management.” More specific = better AI suggestions.
3. Run the AI analysis
Click ”Analyse”. Claude generates five risk suggestions with likelihood (1–5) and consequence (1–5). Each has a title and 1–2 sentences of description.
4. Approve, edit, or reject
Review each suggestion. Approve what's relevant, edit what needs nuance, reject what doesn't fit. Only what you approve lands in your register.
How many AI runs are included?
Starter includes 15 AI analyses per month. Pro includes 80 — enough for a consultant with 5–10 concurrent client cases. Business includes 250. All plans have a 7-day free trial, no credit card required.
AI analysis FAQ
Which AI model is used?
Anthropic's Claude Haiku 4.5. The model runs via the Anthropic API (not locally). The model ID is logged on every generated risk so you can audit which version produced a suggestion.
Can I use AI without data leaving the EU?
RiskNote calls Anthropic's standard endpoint, which runs in the US. For customers that require EU data residency for AI inference, contact us — we're evaluating EU-bound options for a future version.
Do rejected suggestions count against my monthly quota?
No. The quota counts runs, not generated or approved risks. One run = one API call where you receive five suggestions. You can reject all five at no extra cost.
Try AI analysis yourself
First risk assessment ready before lunch. No credit card, no commitment.