IntellectAI

Embedded models are one of six stated architecture properties and the model layer is sold as a named product, yet removing it leaves a full lifecycle wealth platform from onboarding to reporting. Banking software from the parent has shipped for three decades, and the platform underneath does not depend on the intelligence on top.

Every recommendation carries its reasoning, data sources and a confidence level, and advisers keep control of client interactions, surfacing certainty at the point of each decision. Missing are a sampling audit of outputs, an action threshold and any rule for what happens when confidence is low.

A confidence level on every output gives validators a live behavioral signal across the stream rather than one explicable case. Nothing says how that score is calibrated against realized outcomes, and no accuracy, benchmark or validation package is published, so the number is something the model emits rather than a measure of being right.

The head of wealth product and private banking at a Thai bank is quoted by name, and a Q2 2026 analyst landscape lists the platform among 23 notable vendors for agentic automation and compliance monitoring. No adoption figure, productivity measure or revenue effect attaches to any named client.

A three layer knowledge structure separates public market data, the institution's own product and compliance rules, and private client information, with multi tenancy named as a platform property. Presented as architecture rather than a commitment, it never states that one institution's client data stays out of another's outputs or whether models are tuned per tenant.

Deployments across Asia, the Middle East and Europe span very different data protection regimes, and none is addressed publicly. No privacy program, retention position or data subject rights framework covers the client financial data the models read.

Platform and model layer alike publish no certification, attestation report or trust center. A bank evaluating full lifecycle wealth infrastructure across three continents finds nothing to start from before procurement.

Clients carry the authorizations while the vendor holds only the securities disclosure duties of a listed parent, a reporting regime rather than a financial license. Nothing places the model layer before a supervisor, sandbox or regulator run test.

Showing how a recommendation was produced is not evidence that recommendations are even handed across wealth bands or demographics. Portfolio restructuring suggestions and nudges distinguish between clients by design, and no bias testing or fairness evaluation is published.

The adviser sees reasoning, sources and confidence, while the client receiving the advice sees none of it and nothing tells them a model produced the recommendation. Transparency to the intermediary is not recourse for the affected party, and nothing gives the client a contest route or a liability position.

The model layer is named as a distinct platform, a step beyond calling it proprietary, with nothing disclosed about what runs inside. Without a base model, provider or version, an institution cannot record model versions against advisory decisions as model risk oversight expects.

Lifecycle coverage runs from client acquisition through portfolio construction and execution to post trade servicing in one platform. Event driven, microservices based and headless, it is built to be consumed through a client's own front end and to connect to institutions and marketplaces.

Cloud native delivery with multi region and multi tenant operation is named outright, which matters for institutions in Asia and the Middle East under local hosting expectations. No residency regions, on premises path or statement of where the model layer runs relative to client data appears.

The platform is composable and sold by configuration across a wide feature surface, with no published rate, module price or minimum. How units combine into a deployment cost cannot be estimated before a sales engagement.

Relationship managers at retail and private banks, trust banks and wealth managers are served across Asia and the Middle East, with multi currency and multilingual operation built in. The parent's 325 customers span four business lines, and no client count or assets figure is given for the wealth product itself.