Inventories fail in predictable ways: they are built once for a board paper, they list models but miss agents and embedded SaaS AI, and nobody owns keeping them current. VERITRAIL Register is designed around the controls that prevent those failures.
Capabilities
- AI system of record (EI-01). A single inventory spanning in-house models, LLM applications, agents, vendor AI products and AI features in SaaS tools, with record counts and last-updated timestamps.
- Risk classification (EI-02). Classification by business criticality, data sensitivity, decision impact, autonomy and regulatory relevance, producing a suggested tier that a human confirms.
- Named ownership (EI-03). Business and technical owner per system, with owner acceptance recorded.
- System profiles (EI-04). Intended purpose, users, affected parties, context of use and foreseeable misuse.
- Model, prompt and data registries (EI-05, EI-06). Versions, prompts, system instructions, configuration, data sources and lineage.
- Agent and tool registry (EI-07). Every agent, every tool, API, plugin and data store it can reach, and the operations it is permitted to perform — the baseline for runtime least privilege in VERITRAIL Shield.
- Vendor AI inventory (EI-08). Vendor products, embedded features and foundation-model providers, with contract and data-processing references.
- Shadow AI reconciliation (EI-09) and owner attestation (EI-11) to keep the inventory current.
- Use-case intake register (EI-12). Proposed use cases are captured before build or procurement and flow into VERITRAIL Control.
Who it is for
- CIO and CTO offices that need a complete picture of AI in the estate.
- CISOs who need to know which agents can reach which systems and data.
- AI Governance Leads who need the inventory to drive tiering, assessment and evidence.
- Procurement and vendor risk teams responsible for third-party AI.
Outputs
| Output | Use |
|---|---|
| Inventory export (CSV / JSON) with record counts and coverage | Board and audit reporting; evidence for EI-01 |
| Tier distribution and owner coverage metrics | Leadership KPIs (SE-03) |
| Agent tool/permission manifests | Input to runtime policy; access reviews (IS-11) |
| Attestation and reconciliation records | Evidence for EI-09 and EI-11 |
| Per-system profile | Starting point for technical documentation (GL-06) and evidence packs (SE-05) |
Example: an agent record
| Field | Value |
|---|---|
| System | Claims triage assistant (agent) |
| Business owner / technical owner | Head of Claims Operations / Platform Engineering Lead |
| Intended purpose | Summarize incoming claims and propose routing; does not decide claims |
| Decision impact | Advisory |
| Data sensitivity | Personal, including some special-category health data |
| Autonomy | Suggests |
| Tools | Claims system (read), document store (read), ticketing (write — create only) |
| Vendor model | Yes — foundation-model API under enterprise agreement |
| Suggested tier | High (Inferred — special-category data, vulnerable affected parties); awaiting confirmation by AI Risk Lead |
| Required modules | Core, data and privacy, security, generative AI, agentic, fairness and safety, human oversight, vendor, monitoring |
VERITRAIL controls cited
- EI-01 Enterprise AI inventory (system of record)NIST AI RMF: GOVERN 1.6ISO/IEC 42001: A.4.2, 7.5EU AI Act: Art. 26, Art. 49
- EI-02 AI risk classification and tieringNIST AI RMF: MAP 1.1, MAP 1.5, GOVERN 1.3ISO/IEC 42001: 6.1.2, A.5.2EU AI Act: Art. 6
- EI-03 Accountable owner per AI systemNIST AI RMF: GOVERN 2.1ISO/IEC 42001: A.3.2, A.10.2EU AI Act: Art. 26
- EI-07 Agent and tool registryNIST AI RMF: GOVERN 1.6, MAP 4.1ISO/IEC 42001: A.4.4, A.6.2.3EU AI Act: Art. 11
- EI-08 Third-party and vendor AI inventoryNIST AI RMF: GOVERN 6.1ISO/IEC 42001: A.10.3EU AI Act: Art. 25
- EI-09 Shadow AI discoveryNIST AI RMF: GOVERN 1.6ISO/IEC 42001: A.4.2
Framework mappings are indicative VERITRAIL interpretations (Control Library v1.0.0); they are not legal advice and do not establish conformity or compliance.
Limitations
Register is only as complete as the sources and discovery feeding it. VERITRAIL does not perform network or endpoint scanning; it reconciles discovery results you provide. Suggested tiers are Inferred until confirmed. Coverage attestations are User-asserted unless sampled and verified.