OakDai

The intelligence beneath every credit decision

OakDai is a four-layer decisioning AI platform that turns raw data into structured knowledge, and structured knowledge into optimal options — giving Kazuzo the architectural foundation to underwrite, service and manage risk at scale.

OAK

Options. Actions. Knowledge.

The architecture treats intelligence as the interplay between what a system knows about the world and what actions it can take within it — at every level of abstraction and across every time horizon.

Structured world models

Knowledge

A continuously updated knowledge graph of entities, relationships, policies and outcomes — the foundation from which every credit decision derives meaning.

  • Entity & relationship graphs
  • Policy & constraint encoding
  • Predictive outcome models
  • Causal inference chains

Temporally extended actions

Options

Abstract decision pathways across time horizons — from an instant limit change to a multi-quarter portfolio programme — enabling hierarchical decisioning.

  • Hierarchical action abstraction
  • Multi-horizon planning
  • Option composition & chaining
  • Reward & risk modelling

AI-driven optimal selection

Decisioning

Where knowledge meets options: contextual constraints evaluated, learned policies applied, and the highest-value path surfaced with an explanation attached.

  • Contextual policy enforcement
  • Multi-criteria optimisation
  • Explainable AI reasoning
  • Real-time adaptation

Platform architecture

Four layers. One platform. Adopted incrementally.

A complete composable stack from raw ingestion to domain applications — so capability can be added without rebuilding what already works.

01

Data & Signal Layer

The foundation of truth

Ingest, normalise and validate streams from core systems, bureaus, multi-bank aggregation, regulatory oracles, IoT and asset telemetry into one semantic fabric — with lineage and attestation throughout.

02

Intelligence Layer

Where patterns become insight

Feature engineering, behavioural scoring, alternate-data models and early-warning signals that convert raw activity into a live read of capacity, intent and exposure.

03

Decisioning Layer

Options evaluated, policies enforced

Credit policy, Shariah rules, affordability caps and regulatory constraints expressed as configuration — so the rulebook can change without rewriting the lender.

04

Application Layer

Solutions delivered at the edge

Origination journeys, limit management, servicing and collections surfaced wherever the borrower is — branch, app, merchant checkout, dealer floor or employer payroll.

Reference architecture

Technology, Banking-as-a-Service, Open Banking and Digital Banking

Composable platforms, forward-deployed delivery, outcomes that compound. The same substrate serves Kazuzo's own book and the partners who embed it.

Open Banking

Consented account aggregation and payment initiation over instant rails. Multi-bank cash-flow visibility becomes the underwriting substrate; Raast becomes the disbursement and collection channel.

Banking-as-a-Service

Ledger, KYC, limits and compliance exposed as APIs so partners can embed Kazuzo credit into their own journeys — with onboarding, monitoring and reporting handled by the platform.

Digital Banking

Composable front-ends over a common core: customer onboarding, servicing, statements, disputes and self-service collections, all reading from the same decisioning state.

Islamic-native rails

Murabaha, Ijarah, Mudarabah and Diminishing Musharakah modelled in the core ledger, with Shariah governance embedded in product configuration ahead of the 2028 mandate.

Deposit tokens

Programmable commercial bank money, under the credit book

Blockchain representations of bank deposits: 1:1 backed, still on the bank's balance sheet, programmable and available 24/7.

Deposit tokens matter to a lender for three reasons. Settlement stops waiting for a cut-off, so disbursement and collection happen when the event happens. Conditions can be written into the money itself — release against a verified invoice, sweep on a defined trigger, escrow until delivery is confirmed. And collateral and liquidity positions become observable in real time rather than reconstructed at day end.

Aggregated through OakDai's data and signal layer, tokenised flows sit alongside conventional account data in one view — meaning a supply chain programme can be underwritten, funded, settled and monitored on a single consistent record.

Market projections place bank-token transaction volume in the region of $100–140 trillion a year by 2030, with the tokenised deposit market growing from a small base today. Kazuzo treats that as an infrastructure trajectory to be ready for, not a product to chase.

Backing
1:1 against commercial bank deposits
Balance sheet
Remains with the issuing bank
Availability
24/7 settlement and transfer
Programmability
Conditional release, escrow, auto-sweep