Kibo / For the EU AI Act
The EU AI Act, system by system.
Every AI system you build or use, classified with its reasoning recorded, and the obligations of its class and your role turned into tasks with owners, evidence and deadlines. You know where you stand before anyone asks.
An inventory with its reasoning.
Kibo keeps an inventory of the AI systems your company builds or uses: their purpose, provider, your role, the data they use and the people they affect. Each is classified through a short set of questions, and the obligations that follow are tracked to the end.
Records it works on:
- AI systems
- Risk classes
- Classification reasoning
- Obligations
- Owners
- Evidence
- Deadlines
Obligations with owners, not a reading list.
The obligations of each system's class and your role, such as AI literacy, risk management, logging, transparency, human oversight or technical documentation, become tasks with an owner, evidence and a deadline. Classification and dates are worked out by the system, never guessed, and a readiness score shows where you stand.
Approved once. Enforced every time.
Bringing a new AI system into use is a policy: the AI governance lead approves, then the data protection officer when it uses personal data. It is approved once, versioned and enforced by code.
Sovereign by design.
On your own servers, or hosted for you in a region of your choice, with an open-licence model. Your inventory and evidence are never sent to a third-party AI provider, and no telemetry is sent to anyone.
More: sovereignty · where Kibo runs · open model
Predictable cost.
Every system, owner and piece of evidence is included in the same annual licence as the rest of the company, with no per-seat, per-token or per-module pricing.
More: total cost of ownership
An audit trail by construction.
Every classification, change and piece of evidence is in the audit record with a name and a time. Approvals are enforced by code, and nothing is changed without confirmation.
More: governance · security · policy as code