📋 SEBI Compliance Mode — Disclosure documentation per SEBI IA/RA Regulations 2025 + Algo Framework Aug 2025. Not legal advice. Consult a SEBI-registered compliance officer.
SEBI Compliance Intelligence
AI Governance Layer — SEBI IA/RA Regulations 2025 + Algo Trading Framework August 2025
SEBI Requirements Scorecard
| Requirement | Regulation | Status | Evidence | Action |
|---|---|---|---|---|
| AI Usage Disclosure to Clients | SEBI Reg. 19(vii) RA Regs + Reg. 15(14) IA Regs, Jan 8 2025 | ✅ Compliant | Auto-generated AI Disclosure Certificate available. 5 AI models documented with Algo IDs, confidence metrics, and human override status. | Download certificate → share at client onboarding. Mandatory by April 30, 2025 for existing clients. |
| Audit Trail & Record Retention — 5 Years Minimum | Regulation 25(1) RA Regs + SEBI Master Circular Jun 27, 2025 | ⚠️ Partial | ⚠️ No events yet — run AI Analyst to begin logging | Export audit log monthly. Blockchain-anchored hash provides 5-year tamper-proof retention evidence. |
| Algorithmic Kill Switch | SEBI Algo Framework Circular SEBI/HO/MIRSD/MIRSD-PoD/P/CIR/2025/0000013, Feb 4 2025 — effective Aug 1 2025 | ✅ Compliant | Kill switch operational — suspends all AI predictions instantly. Logged to immutable audit trail with unique Event ID per SEBI kill switch SOP requirement. | Test kill switch quarterly. SEBI requires documented SOP for activation and restoration procedures. |
| Unique Algo ID per Algorithmic Strategy | SEBI Algo Framework, Feb 4 2025 — Algo ID tagging mandatory from Apr 1 2026 | ✅ Compliant | Last Algo ID: N/A | Format: EVT-{ts}-{rand5} | 4 registered strategies: RISKGRAPH-HMM-01, RISKGRAPH-KM-02, RISKGRAPH-IF-03, RISKGRAPH-OLS-04 | Maintain Algo ID registry. Black-box strategies require SEBI Research Analyst registration. |
| Model Portfolio Methodology Documentation | SEBI RA Master Circular Jun 27 2025 — Annexure-A, Model Portfolio Framework | ✅ Compliant | 19 mathematical formulas documented at /math with step-by-step derivations. Benchmark: NIFTY 50. Horizon: FY2025-26. Rebalancing: on data refresh. Risk disclosures: explicit in /math limitations. | Include /math URL in client agreement methodology section per SEBI model portfolio disclosure requirement. |
| Concentration Risk Monitoring | SEBI Diversification Guidelines + SEBI UPSI Amendment 2025 — unusual activity monitoring | ⚠️ Partial | HDFCBANK eigenvector centrality: 0.84 — exceeds 0.70 threshold. Systemic contagion risk flagged. GSVI: 0.67. | Review high-centrality asset allocation. Consider hedging or reducing HDFCBANK exposure. |
| Conflict of Interest Disclosure (MITC) | SEBI MITC Requirements — Annexure-B, RA Guidelines Jan 8 2025, effective Feb 17 2025 | ✅ Compliant | RiskGraph issues zero buy/sell/hold recommendations. GPT-4o system prompt hard-codes prohibition on directional financial advice. No commission, no trade execution capability. | Include conflict-of-interest statement in MITC. Display mandatory SEBI disclaimer on all output pages. |
| Annual Compliance Audit | Regulation 25(3) RA Regs — audit within 6 months of FY end, submit to RAASB within 1 month | ⚠️ Partial | Full audit log available for ICAI/ICSI/ICMAI member review. Blockchain-anchored entries provide tamper evidence. Final sign-off requires registered auditor. | Export audit log → submit to registered auditor by September 30 annually. Publish findings on platform website per SEBI adverse findings disclosure requirement. |
| AI Accountability — No Autonomous Delegation | SEBI IA Reg. 15(14) — IA solely accountable for security, confidentiality, integrity of client data processed via AI. Jan 8 2025. | ✅ Compliant | All ML outputs are advisory only. No AI system has trade execution capability. Human investor retains full decision authority. Structurally enforced — not merely stated. | Document human oversight process. Include in client onboarding agreement per Reg. 15(14) requirements. |
AI Disclosure Certificate
Auto-generated per SEBI Reg. 19(vii) RA Regs + Reg. 15(14) IA Regs — Jan 8 2025
AI Kill Switch
SEBI Circular SEBI/HO/MIRSD/MIRSD-PoD/P/CIR/2025/0000013, Feb 4 2025 — Kill switch: last line of defence against algorithm malfunction
4 ML models operational | GPT-4o Analyst enabled
SEBI Kill Switch: Armed and Ready
Immediately suspends all AI predictions. Logs KILL_SWITCH_ACTIVATED event with unique Event ID per SEBI Algo Framework SOP.
Blockchain Audit Anchoring
Polygon Amoy testnet — tamper-proof 5-year retention evidence per SEBI Regulation 25(1)
0 events in memory | Not yet blockchain-committed
SEBI 5-year retention: requires periodic anchoring
Writes audit root hash to Polygon Amoy testnet. Publicly verifiable on PolygonScan. ~3 seconds.
Concentration Risk Monitor
Real-time SEBI compliance flags from live ML outputs
✅ No concentration risk violations detected. Portfolio meets SEBI diversification guidelines. All centrality values within acceptable bounds.
The Regulatory Landscape
As of January 8 2025, every registered Investment Adviser and Research Analyst in India must disclose AI tool usage to clients, maintain tamper-proof audit records for 5 years, and implement algorithmic kill switches per SEBI's August 2025 framework. RiskGraph generates this compliance infrastructure automatically — one platform, one click.