Register your first AI system
Add an AI system, answer the guided questions and behavioural flags, and get a classification with the exact obligations that apply.
Everything in Veritome hangs off an AI system. Before the Act applies to anything, you have to know what you have — so your first job is to build an inventory. Registration is the front door of the 6-phase journey (Classify → Scope → Implement → Assess → Register → Monitor): it captures what the system is and who you are in the value chain, and hands those answers to the engine that decides what you must do.
Step by step
- Open AI Systems and click Register a system.
- On the intro screen, press Start registration. The wizard autosaves your draft to this browser, so you can close it and return with Resume registration without losing your answers.
- In the details panel at the top, fill in the identity of the system:
- System name — a recognisable label (e.g. ResumeRank). Required.
- What does it do? — a one- or two-line purpose (e.g. résumé screening built on Mistral, used by our recruiters). This is the text Aria reads if you turn on assist, so make it descriptive.
- Vendor, Department, Version — optional context for your register and for later documentation.
- Owner — optionally assign an accountable person now from your team; you can also assign owners per obligation later.
- Work through the five guided steps — Role → Model → Prohibited → Risk → Review (see How classification decides your obligations).
- On the Review step, confirm the declaration and press Create. The system lands on your register and the engine generates its obligations.
Tip: register a system even if you're not sure it's in scope. Running it through classification is the cleanest way to find out — and a system you've assessed and excluded is a far better answer to a regulator than one you never looked at.
The behavioural questions — and what each one drives
A handful of yes/no answers change which obligations apply. The engine reads them as behavioural flags on the system, so answer them honestly — each one switches specific duties on or off. They're captured across the wizard steps, not all on one screen.
| Question | Where it's asked | What it drives |
|---|---|---|
| Placed under your own name? | Role step (Art. 25) | If you're a deployer, distributor or importer who brands the system as your own — or substantially modifies it — Art. 25(1) upgrades you to provider, the heaviest obligation set. |
| Workplace AI? | Risk step — the Employment Annex III domain | Marks the system as used to recruit, screen, evaluate or monitor workers (Annex III §4). Adds Art. 26(7) worker-information and Art. 26(6) logging duties for deployers. |
| Chatbot / interacts with people? | Risk step — transparency capabilities | Triggers Art. 50(1) transparency: you must tell people they are interacting with an AI system. |
| Emotion recognition? | Risk step — transparency capabilities | Inferring emotions from biometric data. Triggers Art. 50(3) disclosure — and, in workplace or education contexts, may be a prohibited practice under Art. 5. |
| Generates content? | Risk step — transparency capabilities | Produces synthetic text, image, audio or video. Triggers Art. 50(2) machine-readable marking and Art. 50(4) deepfake / synthetic-content labelling. |
These flags are independent of the risk tier. A minimal-risk chatbot still owes Art. 50 transparency; a high-risk hiring tool that also generates content owes both its Annex III duties and the labelling duties. The engine layers them.
Your inventory
Each registered system gets a card on the AI Systems page showing its risk tier, your role, its compliance score, and where it sits in the 6-phase journey. This inventory is itself a compliance artifact: regulators expect you to know, and be able to show, every AI system in scope. From a card you can open the system, reclassify it, or jump into any obligation tab.
Once registered, the system's classification dossier opens automatically with a sealed, tamper-evident first entry — your audit trail starts on day one.