Loading...
Agentic Design Patterns
A connected library of patterns and techniques for building reliable AI agents, each with trade-offs, use cases, and an interactive demo.
24 categories · 285 techniques · 1,243 documented use cases
Start here
Six foundational techniques most agent architectures build on.
Browse all categories
Reasoning & control flow
Reasoning Techniques
19Advanced reasoning and thinking techniques
Prompt Chaining
7Multi-step prompt orchestration patterns
Routing
9Dynamic request routing and delegation patterns
Parallelization
5Concurrent execution and parallel processing patterns for AI systems
Planning
9Advanced planning strategies for complex multi-step tasks and goal achievement
Workflow Orchestration
16Stateful, event-driven, and enterprise workflow coordination patterns
Knowledge & memory
Knowledge Retrieval (RAG)
13Information retrieval and augmented generation patterns
Memory Management
16Context management and state persistence patterns
Context Management
19Strategic context window optimization and engineering patterns for AI agents
Context Orchestration
4Patterns for selecting, routing, sizing, and combining context
Knowledge Representation
5Structured knowledge, semantics, validation, and reasoning patterns
Tools, agents & interaction
Tool Use
11External tool integration and function calling patterns
Multi-Agent
15Coordination and communication patterns for multiple AI agents
Agent Economy & Interop Protocols
7Payments, commerce, discovery, identity, and web contracts for the agentic internet
UI/UX & Human-AI Interaction
32Comprehensive user interface, experience, and human-AI collaboration patterns for agentic AI systems
Quality & improvement
Reflection
4Self-evaluation and iterative improvement patterns
Learning and Adaptation
20Dynamic learning and behavioral adaptation patterns
Evaluation and Monitoring
22Performance assessment and system monitoring patterns
Exploration & Discovery
4Patterns for search, experimentation, adaptation, and solution discovery
Prioritization
4Transparent ranking, scheduling, and multi-criteria decision patterns
Interpretability
5Patterns for explaining, inspecting, and validating model and agent behavior
Production hardening
Fault Tolerance Infrastructure
5Infrastructure-level fault tolerance patterns for AI system reliability
Security & Privacy Patterns
26Comprehensive security, privacy, and ethical AI patterns for enterprise deployment
Resource-Aware Optimization
8Cost, latency, energy, compute, and memory optimization patterns
Where to go next
New to agent patterns
- Start with Reasoning Techniques for foundational thinking patterns.
- Explore Prompt Chaining to build multi-step workflows.
- Learn Tool Use to connect agents to real systems.
Designing production systems
- Coordinate work with Multi-Agent patterns.
- Keep long sessions coherent with Context Management.
- Ship safely with Security & Privacy and Fault Tolerance.