A proposal for an AI-governance baseline. Enter the access key you were provided.
J.P. Morgan Private Bank puts AI into research, screening, and client-facing workflows — the surface examiners and limited partners now expect documented and governed. We propose a fixed-scope baseline that inventories your AI estate, scores it across three lenses, and hands you a board- and LP-ready posture — in weeks, not quarters.
Supervisors have made model governance non-optional: model-risk management expectations (SR 11-7 / OCC 2011-12) now extend to AI/ML, the SEC's predictive-data-analytics rulemaking targets conflicts in AI-driven client interactions, and the SEC cyber-disclosure rule covers AI-touched systems. LPs and auditors increasingly ask for a documented AI-governance posture in diligence.
To be clear about what we are not saying: nothing in our preliminary review suggests J.P. Morgan Private Bank's AI behaves improperly. Working from public materials only (June 2026), what we observe is that a published basis — model disclosure, accuracy basis, or human-oversight description for AI in decision-sensitive workflows — is not readily locatable. The remedy for a control gap is fast: document, add the right human-in-the-loop gates, and review — which is exactly what the baseline delivers.
You receive a prioritized remediation roadmap and a posture readout your leadership, owners, and counsel can act on. Where independent attestation is the goal, we prepare the evidence — we never self-certify; formal certification runs through accredited third-party bodies, and the baseline is the on-ramp to ISO/IEC 42001 readiness.
Indicative fee — Regulated / large-cap
$350K+ (custom)Indicative range for an organization your size; fixed once we confirm the AI-system count together in the first working session. Standards mapped: NIST AI RMF · SR 11-7 model risk · NIST CSF 2.0 · ISO/IEC 42001.
| Week | What happens | From your side |
|---|---|---|
| 1 | Working session; access scoping; AI-feature and data-flow inventory begins. | One technical contact; product/ops walkthroughs. |
| 2–3 | Evidence collection; three-lens scoring; sector control review; cost read. | Short interviews with product, security, and compliance owners. |
| 4 | Readout: posture, prioritized roadmap, remediation plan, attestation path. | A leadership readout session. |
Every step has a human decision gate — nothing is assessed, scored, or reported without your context.
We bring the baseline instrument and a draft inventory built from your public footprint; you bring the people who know the work. Ninety minutes confirms scope, fee, and timeline — and you leave with our preliminary observations either way.