Compliance
One core engine. Any regulator plugs in.
Regulatory timeline
Three regulators, six moves, eighteen months — Ordinis ships a plugin for each.
Feb 2025
SEBI — AI/ML liability rule
Jun 2025
SEBI — AI/ML consultation
Aug 2025
RBI — FREE-AI report
May 2026
SEBI — AI-risk circular
Jun 2026
RBI — draft MRM guidance
Jun 2026 →
IRDAI — framework forming
What these circulars require, and how Ordinis covers it
A register of every AI model and agent running in production
AI asset registry — every model risk-tiered at onboarding, with the tier traveling with every routing decision.
Evidence of validation before deployment, and of what changed after it
Validation gating blocks unvalidated models from serving traffic; model integrity monitoring flags changes after deployment.
A record of who approved what, and the ability to stop a model without stopping the business
A hash-chained audit trail records every approval; the kill switch suspends one model without halting the workflow it sits inside.
One core engine, any regulator plugs in
RBI
Available now
PAN/Aadhaar/IFSC · HITL escalation
SEBI
Available now
MNPI/ISIN detection · PIT Regs
IRDAI
Coming soon
Insurance data-privacy modules
One core engine, the same governance and DLP foundation — the regulator-specific ruleset is what plugs in on top.
Regulator plugins, in detail
RBI
Available nowBanking & NBFC lending
PAN/Aadhaar/IFSC detection, HITL escalation for high-risk decisions. Maps to the RBI FREE-AI Framework and draft MRM guidance.
SEBI
Available nowCapital markets & brokerages
MNPI/ISIN detection with an always-block hook on top of the core output firewall. Maps to SEBI PIT Regulations and the AI-risk circular.
IRDAI
Coming soonInsurance
Insurance-specific data-privacy modules, on the roadmap as IRDAI's framework forms.
Mapped controls
SEBI AI/ML Liability Rule
Feb 2025 — market intermediaries
RBI FREE-AI Framework
Aug 2025 advisory, philosophy-setting
SEBI AI-Risk Circular
May 2026 — sharpest control requirements yet
RBI Draft MRM Guidance
Jun 2026 — model register, validation evidence, approval record
DPDP Act 2023
Purpose limitation + data minimization layer
What you hand an examiner
Hash-chained audit trail
Every request logged with SHA-256 hash chaining — the report an examiner can be handed directly.
AI disclosure headers
Every response carries a machine-readable AI-disclosure header, on every path.
Explainability log
Routing rationale, validation status, and risk tier recorded per request.
Examiner mode
72-hour, time-limited examiner tokens for a dedicated regulatory examination portal.