BANK-FACING DISCLOSURE · INDEPENDENT REVIEW
One portfolio.
Different use boundaries.
This register separates RegTech Nexus AI’s public tools, calculators, research and responsible-AI materials so a bank can evaluate each item on its own purpose, evidence boundary and technical maturity.
PUBLIC-BUILD POSTURE
What a bank can rely on today.
The published build is intentionally transparent about its limits. It is suitable for learning, structured discussion, synthetic-data exercises and independent recomputation. Production reliance on live or near-live customer, transaction or confidential data requires a materially different hosted and governed deployment.
SEPARATE COMPONENT REVIEW
Portfolio register
| Component | Primary purpose | Public-build boundary | Appropriate current posture |
|---|---|---|---|
| TradeGuardTBML, transaction monitoring and document review | Organises potential trade-finance and financial-crime indicators. | Synthetic/anonymised inputs; deterministic browser rules; no live screening or bank connection. | Sandboxed pilot / awareness only No production AML, approval, freeze, STR or case decision. |
| ICMS Readiness Toolkit24-item, 0–3 self-assessment | Structures discussion around governance, Three Lines, MIS, TBML and corrective action. | Indicative and self-reported; browser-local; not a complete clause-by-clause assurance review. | Advisory readiness aid Evidence and independent assurance remain with the bank. |
| MD/CEO KPI Calculator30 KPIs across five lenses | Organises Board-approved baseline, target, actual and evidence inputs. | Phase 1 manual calculator; public data cannot complete all KPIs; indicative output only. | Board discussion aid Use the prescribed official template and verified evidence for appraisal. |
| PayScale 2026 CalculatorBasic-pay transition scenarios | Models user-entered pay and phase assumptions. | Planning calculator; not a government pay-fixation portal and not a bank payroll system. | Administrative planning only Verify the applicable instrument, grade, effective date and assumptions. |
| Regulatory Control & Policy Gap AnalyzerPolicy-to-control review and action planning | Maps a non-sensitive policy excerpt to an original source-mapped control pack and produces an evidence gap action plan. | Browser-local public pilot; deterministic review prompts; no server upload, confidential document store, legal conclusion or regulatory certification. | Sandboxed review aid Validate applicability, current sources, operating evidence and accountable professional judgment. |
| AI Assistant / GPTResearch and drafting support | Helps with research, issue spotting, drafting and structured questions. | Not a legal or regulatory authority; output may be incomplete, outdated or wrong. | Research only Primary-source verification and human approval are mandatory. |
| Frameworks & ToolkitsWorking models and practical templates | Provides adaptable control, review and governance structures. | Advisory content; not automatically mapped to a bank’s policies, owners or evidence. | Adapt and validate Map each recommendation to an approved control objective. |
| Publications & PlaybooksTraining and governance material | Supports workshops, professional learning and management discussion. | Educational material; not regulatory instruction or legal advice. | Training / awareness Use alongside approved manuals and authoritative sources. |
| Regulatory ExplainersSimplified interpretation | Makes complex requirements easier to discuss and teach. | Must be checked for jurisdiction, date, scope and primary-source wording. | Orientation only Do not use as sole evidence for a compliance conclusion. |
| Research & InsightsRegulatory and risk analysis | Horizon scanning and structured professional commentary. | Opinion and analysis may change; publication date and source list matter. | Horizon scanning Subject to source, method and update review. |
| Responsible-AI materialsAI readiness and governance | Introduces explainability, evidence, bias, permissions and human checkpoints. | Conceptual governance baseline; no claim of a tested bank control environment. | Governance starting point Convert principles into tested controls and owners. |
| Agentic-AI governancePermissions, evidence, checkpoints and stop controls | Structures safe discussion of AI-agent mandates and actions. | No production agent deployment, RBAC or kill-switch service is provided by the page. | Design and challenge aid Implementation evidence is required before approval. |
| Briefing and contact channelsInformation distribution | Receives enquiries and requests for regulatory updates. | Email-based contact fallback; no claim of a bank-grade subscription or case-management service. | Communication channel Do not send confidential customer or transaction data. |
TECHNICAL AND OPERATIONAL BOUNDARY
Transparent limits are part of the product.
Browser-local public build
Published calculator and toolkit inputs are processed in the browser. The public build does not present a backend case database, institutional login, RBAC, server-side audit trail or bank-system integration.
No live screening claim
The public tools do not claim live sanctions, PEP, adverse-media, SWIFT, core-banking or market-price feeds. A displayed score is not a verified external fact.
Inspectable rules
Browser-side logic can be inspected or altered by a user. This is acceptable for a public educational build, but not for an institutional decision-influencing workflow without server-side execution and change control.
Evidence before reliance
Every bank use case requires source hierarchy, version pinning, independent recomputation, evidence retention, authorised review and a documented decision owner.
Enterprise AI controls are not public claims
Hybrid retrieval, RAG grounding, immutable audit logs, private deployment, PII/MNPI filtering, model validation, SOC 2/ISO evidence, stress testing and outage fallbacks belong in a future institutional due-diligence pack; the public pages do not claim to provide them today.
CONDITIONS PRECEDENT
What should happen before a bank relies on an output.
- Use synthetic data first. Do not enter customer names, account data, transaction details, confidential documents or MNPI into the public build.
- Pin the exact build. Record the product, version, date, source assumptions and change-log entry used for every exercise or pilot.
- Recompute independently. Test boundary values, missing inputs, conflicting indicators, critical overrides and rounding behaviour against an approved reference.
- Map the output. Link every result to the bank’s applicable law, circular, policy, control objective, owner, evidence and escalation route.
- Escalate institutional use. Any use beyond training or sandbox discussion requires an information-security review, privacy review, vendor due diligence, access control, auditability, retention and incident-response design.
- Keep the human decision. The platform must not be the sole basis for customer approval, transaction rejection, account action, STR decision, regulatory submission or Board appraisal conclusion.
GOVERNANCE CONTACT
Questions, corrections and responsible-use requests.
RegTech Nexus AI welcomes corrections to source links, version information, scope statements and product-boundary disclosures. For a responsible-use or security question, use the public contact channel with SECURITY, PRIVACY or CORRECTION in the subject line.