Skip to main content

Agentic Banking documentation

​
What do you want to build?

Connect core banking and third-party systemsGrand CentralBuild and deploy AI agentsBackbase Agent Platform
Design chat and voice experiencesConversational Banking
Build and operate integrationsGrand Central iPaaSMonitor connectors and manage accessGrand Central Console
Model semantic banking dataBackbase Nexus

New to the platform? Start with the Agentic Banking stack to see how the layers fit together.

Read the platform overview
Agentic Banking extends the Backbase platform with infrastructure for AI agents that integrate with bank systems and run under governance controls. Digital Banking covers online and mobile banking, onboarding, employee workspaces, and marketplace capabilities. Banks run AI agents for customer servicing, sales, and operations in Banking OS, the operating environment for governed execution. Employees and agents collaborate on the same Unified Frontline (the Interaction layer in Banking OS). Customers use mobile, web, and conversational channels. This documentation covers the products in that stack: the Backbase Agent Platform, Grand Central connectivity, Backbase Nexus semantic services, Conversational Banking, and related runtime capabilities. The stack has five documentation layers. Each layer maps to a product area. The integration layer appears under the Grand Central tab (Connectors and APIs, iPaaS, and Console). The layers align with the Banking OS model in the Banking OS layers section.
Diagram of the Agentic Banking stack. Grand Central iPaaS feeds Grand Central Connectors and Unified APIs in the integration layer. Connectors and Unified APIs feed the Backbase Agent Platform in the intelligence layer and Grand Central Console in the operations layer. Backbase Nexus in the semantic layer feeds the Backbase Agent Platform, which feeds Conversational Banking in the experience layer.

​
Banking OS layers

Banking OS organizes Agentic Banking into six layers in a fixed sequence. Each layer describes a distinct job in governed agent execution. The table maps each Banking OS layer to its documentation. The future of banking is not defined by the smartest models alone. It is defined by the operating system that puts models to work with orchestration, control, semantic truth, intelligence, and governed connectivity.

​
Capability areas

Agentic Banking delivers outcome-oriented capabilities across the bank. Natural language is becoming the front door to operational execution: the entry point from intent to resolution.
Conversational Banking, relationship intelligence, customer operations, and bank operations are part of the Agentic Banking roadmap. Documentation for these capability areas is published as the products reach general availability.

​
Governed execution

In regulated banking, agents cannot run without controls. Agentic Banking treats the platform as a control plane for governed execution:
  • Workflow control owns the regulated path from intent to resolution, so AI is applied inside bank workflows rather than beside them.
  • Semantic layer captures proprietary operational context: customer state, process state, decision history, evidence, exceptions, policies, outcomes, and feedback loops.
  • Governance and compliance enforce authorization, audit, and policy before agents act on production systems.
  • Systems of action move beyond recording work to executing and proving resolved work, enabling outcome-oriented value.
  • Vertical specialization encodes banking rules, edge cases, and buyer trust requirements that horizontal AI tools do not carry.
Grand Central supplies connectivity. The Backbase Agent Platform supplies orchestration and intelligence. Backbase Sentinel supplies control and policy evaluation. Backbase Nexus supplies the semantic layer. Conversational Banking supplies the interaction surface where customers and employees meet agent capabilities.