For Internal Audit + Compliance · CSCRF-Ready Format

CSCRF + AI advisory crosswalk.

Every Annexure-A directive mapped to its CSCRF control domain. Each row tagged "net-new," "extends," or "amplifies" so you know whether to build something fresh or update existing evidence. Designed for the audit trail, not the marketing deck.

Need help mapping your stack?

Annexure-A Directives

CSCRF Domains

Net-New Obligations

Audit-Ready

Submission Format

Summary view

10 directives. 3 categories of work.

Net-new (3)

M-SOC onboarding (item 6c), SBOM mandate (item 9), Long-term AI plan (item 10)

Extends CSCRF (3)

AI-based VA tooling (item 2), Vendor AI-VD risk (item 3), AI threat scenarios in risk register (item 7)

Amplifies CSCRF (4)

Virtual patching (item 1), Minor-change management (item 4), API security explicit (item 5), ZTNA naming (item 8)

How to use

For each Annexure-A item, you have an existing CSCRF control area to anchor evidence against. Use that as your audit-trail starting point.
The "Net-new" rows are where you need fresh artefacts. Prioritise them in your gap assessment.
The "Amplifies" rows usually need policy or evidence updates, not new programmes.
The "Extends" rows are where you should review your existing programme and add the AI-aware overlay.
Use the right column ("Evidence impact") to scope what your IT committee will ask for.

#	Annexure-A directive	CSCRF domain	Relation	What changes	Evidence impact
01	Patch management + virtual patching	Endpoint Security	amplifies	CSCRF already mandates patch management. The advisory adds explicit endorsement of virtual patching as an interim control where vendor patches are unavailable — driven by AI-driven attackers compressing the disclosure-to-exploitation window.	Add a virtual-patch register to your existing patch-management evidence. Each entry needs CVE reference, control mechanism, retirement criteria.
02	VA using conventional + AI-based tools	Application Security · VAPT	extends	CSCRF mandates VAPT cadence (annual for MIIs, half-yearly for QREs, change-driven for all REs). The advisory extends this with explicit endorsement of AI-based VA tools "where possible." VAPT cadence does not change; AI tooling becomes an evaluation requirement.	Add an AI VA tooling evaluation memo to your VAPT documentation. If AI tools rejected, document rationale. If accepted, document evaluation criteria.
03	Third-party + COTS vendor risk assessment	Vendor / Third-Party Risk	extends	CSCRF requires periodic third-party risk assessment. The advisory specifically requires Exchanges and Depositories to direct empaneled COTS vendors to assess AI-led VD risks — a "shall direct" obligation that flows down the supply chain.	Vendor-management programme must add an AI-VD risk assessment row to vendor scorecards. For Exchanges and Depositories: documented vendor-direction trail.
04	Change management for all changes	Application Security · SDLC	amplifies	CSCRF requires change management for material changes. The advisory amplifies scope to include minor changes — "(including minor changes) should encompass full documentation, thorough impact analysis, structured review, rigorous testing and secure deployment."	Change-management policy must be revised to make minor-change documentation mandatory, not optional. CAB review of last 30 days minor changes for compliance evidence.
05	API security (inventory, authN/Z, rate-limit, whitelist)	Application Security · API	amplifies	CSCRF references API security at high level. The advisory adds four specific sub-items: inventory, strong authN/Z + least privilege, rate limiting + throttling, whitelist-based connections. None are net-new concepts but the explicit enumeration creates audit evidence demand.	API security programme must produce: API inventory artefact, authZ policy doc, rate-limit configuration export, whitelist register. Most REs have these implicitly; explicit documentation needed.
06	SOC monitoring + SOAR/SIEM + M-SOC onboarding	Security Operations · Monitoring	net-new	CSCRF requires SOC for MIIs and QSBs. The advisory adds: (a) explicit SOAR + SIEM integration expectation, (b) "shall expedite" M-SOC onboarding for all eligible REs not yet onboarded, (c) MII obligation to run member-onboarding workshops. (b) and (c) are net-new.	M-SOC eligibility determination, onboarding plan, and (for MIIs) member-engagement programme become required evidence. SOAR playbook documentation needed if not already present.
07	Risk assessment with AI-based threat scenarios	Governance · Risk Management	extends	CSCRF mandates periodic risk assessment with scenario-based testing. The advisory extends scope: "the capability of AI based models may also be considered as one of the risk scenarios." Permissive language ("may"), but the audit expectation is clear.	Risk register must add AI-attacker scenarios. Recalibrate likelihood scores on existing entries where AI capability changes the threat (credential stuffing, API enumeration, business-logic abuse).
08	System hardening + ZTNA	Network Security · Endpoint Security	amplifies	CSCRF mandates secure configuration + least privilege. The advisory specifically names Zero Trust Network Access (ZTNA) — first explicit naming in SEBI cybersecurity guidance. Most REs have ZTNA roadmaps; the advisory makes them programme commitments.	ZTNA architecture document and deployment roadmap become required evidence. CIS-benchmark drift report adds an audit value.
09	Asset inventory + Software Bill of Materials	Asset Management	net-new	CSCRF requires asset inventory. SBOM is net-new in Indian sectoral regulation — the first SEBI mandate for SBOM "for all critical applications including open source stack." This creates a tooling and process obligation most REs have not yet operationalised.	SBOM artefacts (CycloneDX or SPDX format) for every critical application. SCA tool integration. SBOM refresh cadence aligned to release cadence.
10	IT-committee guidance + long-term AI plan	Governance · Cyber Strategy	net-new	CSCRF requires board-approved cyber policy + IT/CSC committee. The advisory adds: (a) "shall seek guidance from IT committees" on AI-VD risk (binding consultation), (b) "need to prepare a long-term plan" for AI in detection and autonomous/agentic mitigation (directory deliverable). (a) is binding; (b) is the most strategic obligation in the entire advisory.	IT-committee minutes recording the consultation. Long-term AI plan as a ratified document. Annual revision cadence.

Need this crosswalk applied to your control matrix?

Two-week scoped engagement: we map your existing CSCRF evidence onto the advisory directives, identify gaps, and produce an IT-committee evidence pack. Pricing on request.

Talk to a Compliance Specialist

CSCRF + AI advisory crosswalk.

10 directives. 3 categories of work.

CSCRF + SEBI AI Advisory Control Crosswalk

Need this crosswalk applied to your control matrix?

Next in this series

SEBI AI advisory checklist

AI advisory readiness assessment

AI defence roadmap template