Event-native substrate orchestrating systems, people, and processes. Contextually aware, architecturally resilient.
The AFI Platform is the banking orchestrator: the event-native infrastructure that delivers banking orchestration, unifying systems, people, and processes on a single substrate. Open to existing systems, cloud, and partners, with no lock-in, it takes institutions from pilot to production in weeks.
Infrastructure your institution can put customers on, build revenue on, and take to the regulator. From day one.
Three principles the platform is built on, not layered on.
Apache Kafka carries the event stream; Apache Iceberg holds the queryable history. Every domain emits events natively. No CDC tap. No nightly extract.
Start with one of the adjacent software categories, add several, or run them all. They unify at the Orchestration Layer, not in an integration backlog.
SOC 2 Type II attested. Audit trails are replays of the event log. Primitives that support the GENIUS Act, EU AI Act, IPR, and ISO 20022.
The AFI Platform · The Banking Orchestrator
Connects the institution's core and the systems around it, unifies them on one live context, and powers the digital channels above.
Canonical Model
Payment & FX Orchestration
The Three Layers
Each layer is independently deployable. State stays shared through one event-native substrate.
The bank's human-facing surfaces: end-customer experiences under the bank's brand (XYB invisible), operator experiences through the XYB Console. Apps, workspaces, and portals all read from the same live event stream.
Where the platform orchestrates business processes, systems, and compliance as one package. Apache Kafka streams every transaction as an event; Apache Iceberg provides warehouse-grade ACID transactions and time-travel on the data lake.
The base the platform runs on: microservices on Kubernetes with Docker, Helm, Terraform, and ArgoCD. Cloud-portable across AWS, Google Cloud, and IBM Cloud for Financial Services.
The two technologies that carry state within the Orchestration Layer: one holds the live stream, one holds the queryable history.
Apache Kafka · Event stream
Apache Kafka captures every transaction live, enabling AI agents to reason in real time across every adjacent system.
Apache Iceberg · Unified table
Apache Iceberg brings warehouse-grade ACID transactions and time travel to the data lake. Raw event stream and analytical workloads stay separated, so neither compromises the other.
Unlike legacy ETL, this streaming approach orchestrates real-time and historical data across every system the bank depends on.
The AFI Platform is the substrate. Solutions run on it, each with the platform's guarantees intact.
Where experiences render.
Where events unify.
Where it all runs.