A collection of multi-agent POC implementations for financial services — 34 use cases across 7 domains, all built on one shared foundation of infrastructure and backend code.
FSI Foundry provides two things:
- Shared foundations — reusable deployment adapters, framework base classes, Terraform modules, and Docker configurations that work across all use cases
- Use case implementations — FSI-specific multi-agent applications, each implemented in two frameworks (LangGraph/LangChain and Strands) and deployable to three patterns (EC2/ALB, Step Functions, AgentCore)
graph TB
subgraph Deployment["Deployment Layer"]
Adapters["FastAPI | Lambda | AgentCore"]
Registry[Agent Registry]
Adapters --> Registry
end
subgraph UseCases["Use Cases (34)"]
Banking["Banking (8)"]
Payments["Payments (3)"]
Risk["Risk and Compliance (5)"]
Capital["Capital Markets (9)"]
Insurance["Insurance (3)"]
Ops["Operations (3)"]
Mod["Modernization (3)"]
end
subgraph Foundations["Foundations"]
Base["Base Classes: LangGraph | Strands"]
Tools["Shared Tools"]
IaC["IaC: EC2 | Step Functions | AgentCore"]
end
Registry --> UseCases
UseCases --> Base
Base --> Tools
Base --> IaC
34 use cases across 7 FSI domains. Each use case includes LangGraph/LangChain and Strands implementations with sample data.
Banking (8)
| Use Case | Agents |
|---|---|
| KYC Risk Assessment | Credit Analyst, Compliance Officer |
| Customer Service | Inquiry Handler, Transaction Specialist, Product Advisor |
| Customer Chatbot | Conversation Manager, Account Agent, Transaction Agent |
| Customer Support | Ticket Classifier, Resolution Agent, Escalation Agent |
| Document Search | Document Indexer, Search Agent |
| AI Assistant | Task Router, Data Lookup Agent, Report Generator |
| Corporate Sales | Lead Scorer, Opportunity Analyst, Pitch Preparer |
| Agentic Commerce | Offer Engine, Fulfillment Agent, Product Matcher |
Payments (3)
| Use Case | Agents |
|---|---|
| Agentic Payments | Payment Validator, Routing Agent, Reconciliation Agent |
| Payment Operations | Exception Handler, Settlement Agent |
| Fraud Detection | Transaction Monitor, Pattern Analyst, Alert Generator |
Risk and Compliance (5)
| Use Case | Agents |
|---|---|
| Document Processing | Document Classifier, Data Extractor, Validation Agent |
| Credit Risk Assessment | Financial Analyst, Risk Scorer, Portfolio Analyst |
| Compliance Investigation | Evidence Gatherer, Pattern Matcher, Regulatory Mapper |
| Adverse Media Screening | Media Screener, Sentiment Analyst, Risk Signal Extractor |
| Market Surveillance | Trade Pattern Analyst, Communication Monitor, Surveillance Alert Generator |
Capital Markets (9)
| Use Case | Agents |
|---|---|
| Investment Advisory | Portfolio Analyst, Market Researcher, Client Profiler |
| Earnings Summarization | Transcript Processor, Metric Extractor, Sentiment Analyst |
| Economic Research | Data Aggregator, Trend Analyst, Research Writer |
| Email Triage | Email Classifier, Action Extractor |
| Trading Assistant | Market Analyst, Trade Idea Generator, Execution Planner |
| Research Credit Memo | Data Gatherer, Credit Analyst, Memo Writer |
| Investment Management | Allocation Optimizer, Rebalancing Agent, Performance Attributor |
| Data Analytics | Data Explorer, Statistical Analyst, Insight Generator |
| Trading Insights | Signal Generator, Cross Asset Analyst, Scenario Modeler |
Insurance (3)
| Use Case | Agents |
|---|---|
| Customer Engagement | Churn Predictor, Outreach Agent, Policy Optimizer |
| Claims Management | Claims Intake Agent, Damage Assessor, Settlement Recommender |
| Life Insurance Agent | Needs Analyst, Product Matcher, Underwriting Assistant |
Operations (3)
| Use Case | Agents |
|---|---|
| Call Center Analytics | Call Monitor, Agent Performance Analyst, Operations Insight Generator |
| Post Call Analytics | Transcription Processor, Sentiment Analyst, Action Extractor |
| Call Summarization | Key Point Extractor, Summary Generator |
Modernization (3)
| Use Case | Agents |
|---|---|
| Legacy Migration | Code Analyzer, Migration Planner, Conversion Agent |
| Code Generation | Requirement Analyst, Code Scaffolder, Test Generator |
| Mainframe Migration | Mainframe Analyzer, Business Rule Extractor, Cloud Code Generator |
Deploy the same code to any of three patterns without changes.
| Pattern | Infrastructure | Best For |
|---|---|---|
| EC2 + ALB | Traditional compute | Development, debugging |
| Step Functions | Serverless orchestration | Event-driven workflows |
| AgentCore | Bedrock AgentCore Runtime | Scalable deployments |
Each use case is implemented in two frameworks:
| Framework | Description |
|---|---|
| LangChain + LangGraph | State-based workflow orchestration |
| Strands | AWS Strands Agents SDK |
Custom frameworks are supported via the base classes in foundations/src/base/.
- AWS Account with Bedrock access (Claude models enabled)
- AWS CLI >= 2.28.9
- Terraform >= 1.0
- Python >= 3.11
- Docker with buildx support
# Interactive deployment (recommended)
./scripts/main/deploy.sh
# Direct deployment
export USE_CASE_ID="agentic_payments"
export FRAMEWORK="langchain_langgraph"
./scripts/deploy/full/deploy_agentcore.sh
# Test
./scripts/main/test.shDeploy multiple use cases to the same AWS account with complete isolation via Terraform workspaces:
# Deploy KYC with LangGraph
USE_CASE_ID=kyc_banking FRAMEWORK=langchain_langgraph \
./scripts/deploy/full/deploy_agentcore.sh
# Deploy Payments with Strands (same account, separate resources)
USE_CASE_ID=agentic_payments FRAMEWORK=strands \
./scripts/deploy/full/deploy_agentcore.shReusable components shared across all use cases.
| Adapter | File | Pattern |
|---|---|---|
| FastAPI | foundations/src/adapters/fastapi_adapter.py |
EC2, local development |
| Lambda | foundations/src/adapters/lambda_adapter.py |
Step Functions |
| AgentCore | foundations/src/adapters/agentcore_adapter.py |
Bedrock AgentCore Runtime |
Central registration for all use cases. Each use case registers itself on import:
from base.registry import register_agent, RegisteredAgent
register_agent("my_use_case", RegisteredAgent(
entry_point=run_agent,
request_model=MyRequest,
response_model=MyResponse,
))Framework-specific base classes for building agents and orchestrators:
foundations/src/base/langgraph/— LangGraph state-based orchestrationfoundations/src/base/langchain/— LangChain agent compositionfoundations/src/base/strands/— Strands SDK integration
Terraform modules for all deployment patterns:
foundations/iac/shared/— Per-use-case resources (S3, IAM)foundations/iac/shared_networking/— Shared VPC across use casesfoundations/iac/ec2/— EC2 + ALBfoundations/iac/step_functions/— Step Functions + Lambdafoundations/iac/agentcore/— AgentCore Runtime
# Required
USE_CASE_ID=agentic_payments # Which use case to deploy
FRAMEWORK=langchain_langgraph # Which framework implementation
AWS_REGION=us-east-1 # Target AWS region
# Optional
DEPLOYMENT_MODE=agentcore # fastapi | agentcore | lambda
APP_ENV=production # development | production
LOG_LEVEL=INFO # DEBUG | INFO | WARNING | ERROR
BEDROCK_MODEL_ID=... # Override default Claude modelUse underscores (not hyphens) for Python compatibility. The foundations automatically convert to hyphens for AWS resource names.
- Create use case directory:
use_cases/my_use_case/src/langchain_langgraph/ - Implement agents and orchestrator
- Register with the agent registry in
__init__.py - Add sample data to
data/samples/my_use_case/ - Update the registry in
data/registry/offerings.json
| Resource | Description |
|---|---|
| Architecture Overview | How FSI Foundry works |
| Deployment Guide | Step-by-step deployment |
| EC2 Deployment | EC2 + ALB pattern |
| Step Functions Deployment | Serverless orchestration |
| AgentCore Deployment | Bedrock AgentCore |
| Security | Threat model, AI security, AWS service security |
| Adding Use Cases | Build your own |
| Local Development | Development setup |
| Testing | Testing strategies |
| Cleanup | Resource teardown |
| Issue | Solution |
|---|---|
| VPC limit reached | Use shared VPC — see setup guide |
| AgentCore not available in region | Use us-east-1, us-east-2, or us-west-2 |
| Docker image build fails | Ensure Docker buildx is installed and use correct architecture |
| Terraform workspace conflicts | Each use case + framework combination needs a unique workspace |
FSI Foundry implements defense-in-depth security:
- IAM least privilege — minimal permissions per resource
- Encryption at rest — S3, CloudWatch Logs
- Network isolation — VPC, security groups
- No hardcoded credentials — IAM roles only
- Input validation — Pydantic models for all inputs
Apache License 2.0 — see LICENSE for details.