Keeps the inventory
One record per AI system: owner, purpose, components, data and risk tier. GOVERN 1.6 asks for “mechanisms” to inventory AI systems, and this is the mechanism.
AI inventoryWhat the software category does, how it differs from the framework it serves, and four checks to make before you buy one.
An AI governance platform is software that keeps an organization’s record of its AI systems and turns governance policy into repeatable work: inventory, risk and impact assessment, framework mapping and audit evidence.
It does not replace the governance framework. It runs that framework across more systems than a spreadsheet can track.
One record per AI system: owner, purpose, components, data and risk tier. GOVERN 1.6 asks for “mechanisms” to inventory AI systems, and this is the mechanism.
AI inventoryRisk and impact assessments on one scale, with sign-off recorded against the system. ISO/IEC 42001 treats them as separate clauses, 6.1.2 and 6.1.4.
AI risk assessmentLinks each system and control to the EU AI Act, NIST AI RMF or ISO/IEC 42001 at article, subcategory or clause level, so one piece of evidence can answer several.
NIST AI RMFExports what auditors, customers and regulators ask for: the inventory, completed assessments and the documents behind them, each with a date and an owner.
AI audit| AI governance framework | AI governance platform | |
|---|---|---|
| What it is | Policies, roles and processes | Software |
| Where it comes from | Written by you, often modeled on the NIST AI RMF or ISO/IEC 42001 | Bought or built |
| Question it answers | What must be true before AI is used? | Is it true today, system by system? |
| Output | Policy, risk tiers and required evidence | Inventory, assessments, mappings and exportable records |
| Fails when | Nobody follows it | It automates a process nobody defined |
Intake forms capture what people declare. Discovery from code, cloud and network activity finds what they did not. Ask which the product does, and from which sources.
Models you build, models you call by API, agents, and AI inside vendor products. A record limited to in-house models leaves third-party AI ungoverned.
A claimed mapping to the EU AI Act, NIST AI RMF or ISO/IEC 42001 should reach the article, subcategory or clause, and show the evidence attached to each.
Check the export formats, and whether a customer or auditor can read the evidence without a seat in your platform.
Straight answers, kept current. Last updated .
Ask us anything →No. A governance, risk and compliance tool manages controls, policies and attestations across the whole business, usually through questionnaires. An AI governance platform is built around the AI system as the unit of record: its models, data, vendors and behavior. The two work together when AI evidence is passed into the GRC system your auditors already use.
No law names a tool. The EU AI Act sets duties, such as a risk management system for high-risk AI under Article 9 and human oversight under Article 14, and leaves the means to you. A platform earns its place when the number of systems, vendors and frameworks outgrows manual tracking.
Cranium ONE is an AI security and governance platform. It generates AI‑BOMs from code repositories, lets customers map AI systems to the EU AI Act, NIST AI RMF and ISO 42001, scores compliance on demand for the NIST AI RMF Starter Pack, and shares evidence through AI Card transparency reports. It does not document a use case intake or approval workflow.
Primary sources for the figures and frameworks cited on this page. Last verified .
See how Cranium AI discovers, observes, governs, secures and proves every AI model, agent and vendor in your enterprise.
Cookie Policy · Privacy Policy
Global Privacy Control detected. Opt-out request honored: marketing tags stay off in this browser.