AI you can govern, price and prove.
Every AI call in RiskVantis goes through one gateway. It applies your choice of provider and model, your zero-data-retention policy and your budgets, and it keeps Restricted data out. Every call is recorded in an append-only usage ledger.
| Time | Actor | Action |
|---|---|---|
| 10:42 | AI · Policy review | Reviewed Access Policy vs ISO 27001 |
| 10:38 | J. Mehta | Approved exception R-1031 |
| 10:21 | AI · Assistant | Created finding F-318 (confirmed) |
| 09:57 | S. Ortiz | Mapped evidence E-904 |
| 09:40 | AI · Mapping | Proposed 12 control mappings |
Illustration with sample data
What AI Governance, Models & Budgets does
Choose your AI provider for the whole installation: Anthropic, Claude on Microsoft Foundry (billed through Azure), OpenAI or Google Gemini. Administrators can choose a different model for each AI feature, and every change is audited. When zero data retention is required, calls go only to a provider whose agreement is recorded, and models that need longer retention are swapped automatically. Prompts are editable, with a diff and second-person approval. Content classified Restricted is never sent to an AI provider.
AI Budget & Usage meters every call through one gateway. Tokens are priced from a versioned price list and recorded in an append-only ledger that never stores prompts or answers. Set company or organisation budgets with module and feature allocations, thresholds and policies: monitor, soft limit, hard limit or approval required. Forecasts, anomaly detection and an AI Billing & Usage dashboard show where AI time and money go.
Key capabilities
Your provider
Anthropic, Claude on Microsoft Foundry, OpenAI or Google Gemini.
Model per feature
Pick the right model for each AI feature; every change is audited.
Zero data retention
Calls only to providers with a recorded agreement when required.
Restricted stays out
Restricted records are never sent to an AI provider.
Budgets & policies
Company or organisation budgets, allocations, thresholds and exhaustion policies.
Usage ledger
Append-only, priced per call, with forecasts and anomaly detection.
The business case.
Efficiency, multi-framework alignment, reduced complexity and streamlined processes: what AI Governance, Models & Budgets changes for your programme.
AI oversight evidence
Provider, model, prompt approvals and usage are all on record.
AI cost under control
Budgets, cost approval and per-feature models stop surprise invoices.
Classification-aware AI
Restricted content never leaves for an AI provider.
No shadow AI
Every AI feature is registered, tagged and metered in one place.
Efficiency figures compare the effort of each workflow cycle in RiskVantis with the same work done manually or in a traditional GRC tool. Your result depends on your starting point; we will model it with you in a demo.
From setup to outcome.
Choose provider
Connect Anthropic, Foundry, OpenAI or Gemini.
Set policy
Zero data retention, prompt approvals and per-feature models.
Budget
Allocate budgets and choose what happens at the limit.
Monitor
Watch usage, forecasts and anomalies on the dashboard.
Supports your obligations.
AI Governance, Models & Budgets contributes evidence and coverage to these frameworks and regulations.
Common questions.
Can we bring our own AI provider account?
Yes. You choose the provider for your installation and supply its credentials; Claude on Microsoft Foundry is billed through your Azure subscription.
Do budgets block AI by default?
No. Budgets monitor by default; only hard-limit and approval policies refuse calls, and only when you configure them.
Are prompts stored in the ledger?
No. The ledger keeps metadata only: feature, model, tokens, cost and who asked.
Works better together.
R.O.S AI Assistant
R.O.S is the one AI assistant in RiskVantis. It answers how-to questions, searches the records you can see, runs ready-made reports and drafts changes, and it acts only…
Learn more →AI & Intelligence360° View & Executive Signals
RiskVantis connects regulations, requirements, controls, policies, evidence, risks, findings, exceptions, assets and vendors. The 360° View answers five questions about…
Learn more →Platform, Security & IntegrationsSecurity, Privacy & Data Protection
A dedicated environment for each customer, encryption at rest, a tamper-evident audit log, deny-by-default APIs and data classification that keeps Restricted content…
Learn more →See AI Governance, Models & Budgets in action.
Book a 30-minute walkthrough tailored to your frameworks, vendors and team.