Skip to content

Platform Story

Why we built a governed risk-results platform instead of another risk engine.

The Problem Worth Solving

Risk computation is rarely the bottleneck. Most institutions can already calculate SA-CCR, CVA, market risk, IRRBB, capital, and stress. What proves brittle — expensively, and often under regulatory scrutiny — is everything around the calculation:

  • making intake quality repeatable, not heroic
  • governing which portfolios, scopes, and configurations actually ran
  • controlling execution so results are reproducible
  • producing evidence and lineage an auditor will accept
  • explaining why a number moved between two runs

These are operating problems, and they are where one-off batch estates quietly fail. A number you cannot reproduce, evidence, or explain is a number you cannot defend.

The Shift: The Result Is The Product

Vannarho RaaS changes the operating model by making the governed result — not the raw batch job — the product. You consume stable APIs, governed workbench flows, curated analytical products, and explainable review surfaces. You do not inherit the operational fragility of running the calculation yourself.

This is a deliberate position:

It is not a hosted risk engine, and it is not rented compute. The value is the complete governed lifecycle that turns a calculation into a defensible, reproducible, explainable outcome.

The Governed Lifecycle

flowchart LR
    A[Intake] --> B[Governance]
    B --> C[Run planning]
    C --> D[Deterministic execution]
    D --> E[Evidence]
    E --> F[Analytical truth]
    F --> G[Result serving]
    G --> H[Explainability]

Each stage is explicit, inspectable, and recorded. That is what lets the platform answer, for any material result: what was submitted, what was approved, what executed, what evidence exists, what was published, and why it changed.

Why Google Cloud

The platform needs a coherent operating substrate for all of the following at once, and Google Cloud provides it as first-party services:

  • stable northbound APIs with policy control
  • strong tenant isolation through projects and IAM
  • a durable analytical backbone for canonical facts and serving
  • isolated, deterministic execution for heavy regulated analytics
  • first-party web delivery and identity
  • managed asynchronous orchestration
  • governed AI and explainability over analytical facts

Deterministic mathematics stay isolated from the control plane. The control plane is operational, not computational — it governs execution; it does not change the maths.

What This Means For You

  • You get regulator-grade outcomes without building and operating the industrialisation layer yourself.
  • You get defensibility by default — reproducibility, evidence, and lineage are properties of the platform, not afterthoughts.
  • You get a clear path to expand — from standard tenants today toward richer explanation, reporting, and multi-geo operation as your needs grow.

Where The Platform Is Today

The governed lifecycle above is available today, end to end, for the core regulatory and economic analytics. Richer narrative explanation, reporting products, and the federated multi-geo model are on the roadmap and build on this proven core. See Capability Maturity And Status.