Skip to content
Justin's Digital Garden

Agentic AI Frameworks

Use Case

Section titled “Use Case”

Agent processes information and sends response on queries and workflows.

Multiple agents work on the flows and a framework helps manage the agents.

About Frameworks

Section titled “About Frameworks”

Agentic frameworks are tool kits for building agentic systems [fn:1]. Framework helps with the with the plan, act, and iterate actions.

Framework features:

  • Integration
  • Monitoring
  • Task
  • Communications

Select a framework based on the use case.

Different Frameworks For Different Systems

Section titled “Different Frameworks For Different Systems”

Linear workflows

Section titled “Linear workflows”
  • Predictable
  • Sequence
  • Reliable, controlled
  • Examples
    • Use case: Customer support
    • Frameworks: LangChain, LlamaIndex, LangGraph

Autonomous agents

Section titled “Autonomous agents”
  • Multiple agents working together
  • Examples
    • Use case: open ended problems, software development, business analysis
    • Frameworks: AutoGen, BabyAGI, CrewAi

Role based agent

Section titled “Role based agent”
  • Multiple agents, similar to autonmous agents, each agent has role working together constrained by role
  • Examples:
    • Use case: team work for research, analysis, generative
    • Frameworks: CrewAI, AutoGen, ChatDev (software development)

Production orchestration

Section titled “Production orchestration”
  • Real work systems
  • Integration of API, database, business workflows
  • Examples:
    • Use case: monitoring and operations and acting on issues
    • Framework: Agent framework (Semantic Kernel and AutoGen), LangGraph

Rapid prototyping

Section titled “Rapid prototyping”
  • Validate ideas
  • Testing
  • Examples:
    • Use case: Prototype creation like user interface creation, user flows and development
    • Frameworks: Langlow, Flowise

References

Section titled “References”