DELPHI
Severe-but-plausible scenario narratives with quantified loss ranges, a two-pass challenger workflow, and a full governance and audit trail.
The Lab
Working prototypes that blueprint what modern operational risk tooling should look like for mid-tier banks.
Every tool in this lab started as a gap I hit inside a bank: a regulatory expectation with no tooling behind it, a committee process held together by spreadsheets and goodwill, a question from an examiner that took three weeks to answer because the evidence lived in nine places.
The premise is simple. Operational risk tooling stagnated at the register stage while the discipline moved to resilience: services, tolerances, dependencies, testing, evidence. AI makes the missing instruments buildable, but only when the builder understands what a regulator will probe and what effective challenge has to capture. So I build them: as working prototypes and blueprints, not products.
Two builds are live. The rest of AEGIS, the twelve-module suite, is in research, sequenced roughly by how much pain each gap causes. Everything here is published to be examined, challenged, and cited; the full roadmap, with architecture decisions and evaluation notes, is available on request.
Live builds
ORBIT
An operational resilience workbench for mid-tier banks: important business services, impact tolerances, dependency mapping, scenario testing, board-ready reporting.
In research: the rest of the platform
Third-Party Dependency Mapper
Nth-party dependency chains for critical services: who actually sits under your important business services, and where concentration hides.
RCSA Copilot
Drafting and challenging risk and control self-assessments: consistency checks across units, stale-assessment detection, challenge prompts.
RegChange Sentinel
Regulatory-change tracking mapped to obligations and controls, so a new guideline lands as a delta, not a reading assignment.
Impact Tolerance Calibrator
Data-driven impact tolerances for important business services, in the metrics a board can actually approve.
Control Rationalizer
Deduplication and gap analysis across bloated control libraries: which controls actually map to which risks, and which are theatre.
External Loss Intelligence
Mining public loss events and enforcement actions into structured scenario inputs and comparator benchmarks.
Issues Management
One register for self-identified, audit, and regulatory issues across the bank: lifecycle workflows from action plan to validated closure, and issue exposure aggregated across every taxonomy node.
Leading-Indicator Intelligence
Which key risk indicators actually lead losses, and which merely decorate the pack: indicator quality grading, backtested thresholds, and honest early warning.
Tail-Event Modeling for Stress Testing
Frequency and severity modeling for operational loss projection under stress, quantified in the shape the CCAR cycle expects.
Loss Event Classification
Consistent, auditable classification of loss events at the point of capture, so the register feeds analysis instead of rework.
The full roadmap (sequencing, architecture decisions, and evaluation notes) is available on request.
Request the roadmap