iCreatAI News
Skip to content
Trending storyUpdating

Duco rebuilds agent stack with Bedrock AgentCore

1 reports1 sources41 min agoUpdated

Understand this story

AI overview

Financial data automation company Duco has rebuilt its agentic AI technology stack based on Amazon Bedrock and Bedrock AgentCore, shifting from traditional machine learning to agentic AI. Duco says this overhaul shortened deployment time and reduced customer onboarding costs while maintaining data isolation for each bank. In the multi-tenant architecture, each tenant is served by an independent runtime and Gateway, and connects separately to that tenant's own Cognito user pool and Amazon EKS namespace. Duco says that with Bedrock, customer data stays within the AWS environment and is not used to train models; high-risk operations require explicit authorization, and complex scenarios exceeding the agent's confidence threshold have an escalation path. The platform supports workloads subject to SOC 2 and DORA requirements, and provides data residency controls per institution.

AI overview based on source reports · Updated 11 min ago

Coverage timeline

Follow the coverage to explore different perspectives.

Nov 17
  1. AWS Machine Learning Blog
    How Duco transforms financial data automation with Amazon Bedrock AgentCore

    Duco, an agentic Operations platform serving 200+ financial institution clients and processing 20 billion transactions monthly, rebuilt its agentic AI stack on Amazon Bedrock and Amazon Bedrock AgentCore, moving from traditional machine learning to foundation-model-based agents. The architecture uses per-tenant AgentCore runtimes, Gateways, and MCP servers for strict isolation, a multi-model strategy spanning Anthropic Claude and Amazon Nova 2 models, and an MCP skills architecture that cut main context from 200,000 to roughly 20,000 tokens. Duco reports reconciliation deployments 10× faster, a 70% reduction in onboarding costs for one global bank, and over 20% higher straight-through processing.

Activity for this story

Not enough continuous observations to show a trend yet.