What you get
Multi-cloud foundation
Kubernetes-first architecture runs on any cloud provider with pluggable components for ingress, storage, DNS, WAF, and APIM. No vendor lock-in.
AI gateway and multi-provider LLMs
Provider-independent gateway with built-in guardrails, semantic caching, and traffic management. Use Azure AI Foundry or bring your own models.
GitOps and automated CI/CD
Argo CD-driven deployments with opinionated workflows: auto-provisioning, PR checks, security scans, build/publish automation, and production releases.
Enterprise observability
Full-stack telemetry with OpenTelemetry, Langfuse agent traces, real-time evaluations, Grafana dashboards, and PagerDuty alerting.
Production-ready starter kits
Pre-wired templates for single agents, multi-agent teams, MCP integrations, and knowledge-based RAG agents. Ship in minutes, not weeks.
Vector storage and RAG infrastructure
PostgreSQL vector database, Redis caching, object storage, Azure Service Bus, and container registries, all configured for RAG workflows.
Banking service integration
Pre-integrated MCP servers for deposits, payments, loans, fraud detection, and customer data. BIAN Coreless adoption in progress.
Enterprise security and compliance
Multi-layer guardrails, PII detection, content safety, RBAC, secure SDLC with scans, sandboxed testing, and full audit trails.
Integration within the Backbase fabric
The Agent Platform resides within the Intelligence Fabric layer, specifically providing Agentic Automation Services. It complements other services in the fabric, including Hybrid AI Services and Data Services, while supporting the Digital Banking Fabric layers above it.The platform adheres to core Backbase principles: Composable architecture, open design, and interoperability.
Reference: Backbase Confluence “Agentic Platform as a Service - Cloud Agnostic Architecture”.
| Fabric | Key services | What you deploy/use | Developer impact |
|---|---|---|---|
| Digital Banking Fabric | Banking Services, Identity, Flow 2.0 | Expose agent APIs via APIM; consume banking/identity/flow APIs | Publish OpenAPI, map to service URL. |
| Intelligence Fabric | Agentic Automation Services (this platform), Hybrid AI Services, Data Services | Agents, tools, knowledge, evaluations | Core build surface for agents and tooling. |
| Integration Fabric | Integration connectors, marketplace partners | Grand Central connectors for core banking systems, payments, and cards | Consume unified APIs for backend integration. |
Core architecture and components
The architecture centers around AI Agents and features a complete ecosystem of supporting services, often referred to as “The Ring”.
AI agents
The core functional units that perform tasks.
Agent creation dev kits
Tools and kits for building and customizing agents.
Knowledge layer (RAG)
Provides knowledge retrieval and context using hybrid search.
Tool calling and MCP
Enables agents to discover and use external tools via the Model Context Protocol.
Agent orchestration
Manages multi-step agent processes and workflows.
AI observability
Provides monitoring, logging, and audit trails for all agent activities.
Evaluation and guardrails
Ensures performance, security, and compliance.
AI gateway
Unified entry point for LLM access and policy enforcement.