AI statement

1. What NORD is

NORD is an AI system that V.O.F. NORD SOLUTIONS offers under its own name. It uses general-purpose AI models from specialised providers (such as Anthropic) and adds instructions, tools, connections to your services, your company knowledge and safety limits. In AI Act terms we are the provider of the AI system NORD; the model builders are providers of the underlying models. You, as a business using NORD, are the deployer.

2. Risk class under the AI Act

• Not a prohibited AI practice (Art. 5): NORD does not manipulate, score people socially, recognise emotions at work or identify people biometrically.
• Not a high-risk system (Art. 6, Annex III) for its intended use: administration, communication, knowledge lookup, reports and invoice checks for businesses. NORD makes no decisions about people.
• It is an AI system that communicates with people and generates text, so the transparency rules of Article 50 apply; see §3.

That is why our terms exclude use for high-risk purposes, such as recruitment or evaluating employees, creditworthiness of individuals, access to education or essential services, and anything biometric. Anyone who uses NORD for those purposes acts outside its intended purpose and carries the AI Act obligations that then apply.

3. Transparency

• In the app, every conversation states that you are working with an AI and that NORD can make mistakes.
• A message NORD sends on your behalf comes from your account; you are the sender. Where the recipient may reasonably expect to know that an AI helped write it, say so. We recommend mentioning AI use in customer contact by default.
• Images NORD creates come from image models that add their own identifying marking (such as metadata or a watermark); NORD does not remove it.
• If NORD makes calls on your behalf in the future (not yet in the product), it will state at the start of every call that it is an AI.

4. People stay in charge

• Outward actions: in a conversation NORD asks for your confirmation before sending anything on your behalf. Goals and scheduled tasks have a hard block: any outward action not explicitly allowed waits for your approval.
• You can stop an answer or task at any time.
• Every goal has a budget and a maximum number of steps; above that NORD stops.
• The invoice check in Finance gives a verdict with reasons; a person decides and can correct it, and NORD learns from those corrections.
• You can view, edit and delete what NORD remembers about you in the app.

5. What NORD cannot do

• NORD can make mistakes: invent facts that sound convincing, use outdated information, misunderstand something or miscalculate. Check what matters.
• NORD does not give legal, tax, financial or medical advice. Use its output as help, not as a replacement for an expert.
• Models can carry biases from their training data. Do not use NORD to judge people.
• NORD only knows what is in its knowledge, what you provide or what it looks up. Answers about your business are as good as those sources.

6. Which models NORD uses

NORD does not train AI models on your data. Which data goes to a model and where is in the privacy policy.

7. AI literacy

Since 2 February 2025 the AI Act (Art. 4) requires those who provide or deploy AI to ensure the people working with it understand enough of it. Everyone at NORD works under an internal AI literacy policy. For your team: this statement, the explanations in the app and our onboarding help provide the basics; feel free to ask for a short tailored briefing.

8. Report something

Does NORD do something that is wrong, harmful or feels off? Report it via the feedback button in the app, or email hi@nordsolutions.nl. For personal data: privacy@nordsolutions.nl. We review every report and act on serious incidents immediately.

9. Changes

We update this statement when NORD, the models it uses or the rules change. The date at the top shows when that last happened.