Solution samples for industry verticals

Production-style sample projects that show how to build, operate, and evaluate AI workflows with LangSmith and LangChain.

This repository is built for teams moving from prototype to production. Each sample models a real business workflow and demonstrates implementation patterns you can adapt to your own stack.

🔗 Documentation & SDK Links

• LangChain docs: https://docs.langchain.com/oss/python/langchain/
• LangChain Python API reference: https://reference.langchain.com/python/
• LangGraph docs: https://docs.langchain.com/oss/python/langgraph/
• Deep Agents overview: https://docs.langchain.com/oss/python/deepagents/overview
• Deep Agents Python API reference: https://reference.langchain.com/python/deepagents/
• LangSmith docs: https://docs.smith.langchain.com/

👥 Who This Is For

• Engineering leaders evaluating production patterns for AI systems
• Product and platform teams aligning AI features with business KPIs
• Developers who want runnable, domain-oriented reference implementations

📦 What You Will Find Here

• End-to-end sample applications by industry/domain
• LangGraph and LangChain orchestration patterns
• Optional LangSmith tracing and observability setup
• Runnable example code with local development instructions

💡 Why These Samples Matter

• Reduce time-to-first-production workflow with proven project scaffolding
• Improve reliability with traceable execution and debugging via LangSmith
• Demonstrate reusable patterns for multi-step and multi-agent systems
• Connect technical implementation to business outcomes (speed, quality, cost)

🗂️ Repository Layout

langsmith-samples/
├── supply-chain/
│   └── supply-chain-optimization/   # Available now
├── financial-services/              # Available now
│   ├── document-research-agent/
│   └── fraud-detection-agent/
├── healthcare-life-sciences/        # Coming soon
└── retail-ecommerce/                # Coming soon

🧪 Samples

Domain	Sample	Status	Description
Supply Chain	supply-chain/supply-chain-optimization	Available	Multi-agent optimization for demand forecasting, inventory, and logistics coordination.
Financial Services	financial-services/document-research-agent, financial-services/fraud-detection-agent	Available	Document research plus fraud detection/AML workflows with AWS mapping, Deep Agents options, risk scoring, and escalation.
Healthcare & Life Sciences	-	Coming soon	Planned examples for care operations and decision support workflows.
Retail & Ecommerce	-	Coming soon	Planned examples for planning, merchandising, and fulfillment workflows.

🚀 Quick Start

✅ Prerequisites

• Python 3.10 or newer
• pip
• An LLM API key (current sample uses Anthropic)
• (Optional but recommended) LangSmith API key for traces

⚙️ Run the Supply Chain sample

cd supply-chain/supply-chain-optimization
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env

Add credentials to .env, then run:

python example.py

For sample-specific details, see supply-chain/supply-chain-optimization/README.md.

💳 Run Financial Services samples

# Use Case 1
cd financial-services/document-research-agent
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
python run_demo.py --query "What changed in KYC onboarding requirements between 3.1 and 3.2?"

# Use Case 2
cd ../fraud-detection-agent
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
python run_demo.py

🎓 What You Can Learn from the Current Sample

From supply-chain/supply-chain-optimization, you can learn how to:

• Coordinate specialized agents through an orchestration layer
• Model domain state with typed schemas and structured outputs
• Incorporate external signals into planning and decisioning
• Stream workflow execution step-by-step for real-time visibility
• Instrument end-to-end traces for operational monitoring

📈 Configure LangSmith Tracing (Recommended)

Set these environment variables in your .env file to capture traces:

LANGCHAIN_API_KEY=your_langsmith_api_key_here
LANGCHAIN_TRACING_V2=true
LANGCHAIN_PROJECT=your_project_name

Learn more:

• LangSmith docs: https://docs.smith.langchain.com/
• LangSmith product overview: https://www.langchain.com/langsmith

🧭 How to Use These Samples

• Start with one sample and run it unmodified
• Map the sample architecture to your own system boundaries
• Review how state, tools, and orchestration are modeled
• Enable tracing and inspect execution runs in LangSmith
• Adapt schemas, prompts, and routing logic to your domain data
• Define success metrics before extending a sample in production

🤝 Contributing

Issues and pull requests are welcome. If you are proposing a new sample, include:

• The business problem and target user
• Architecture overview and key design decisions
• Setup instructions and runnable example inputs
• Expected outputs and observability/tracing guidance

📄 License

This repository is licensed under the MIT License. See LICENSE.