Module 11 Wrap-up: The Strategic Selection
Hands-on: Use a decision matrix to select the right framework for three real-world business scenarios.
Module 11 Wrap-up: Making the Choice
You have seen the frameworks in action. Now, let’s test your intuition. A great agentic engineer doesn't just build; they Architect.
Hands-on Exercise: The Architectural Challenge
Read the following scenarios and pick which framework (or combination) you would use.
Scenario A: The News Summarizer
Goal: A simple bot that monitors an RSS feed and posts a 1-sentence summary of every new article to a Discord channel.
- Your Choice: ?
- Reason: ?
Scenario B: The HR Compliance Validator
Goal: An assistant that checks employee vacation requests against local labor laws (dynamic rules) and requires a manager's digital signature if the request is over 14 days.
- Your Choice: ?
- Reason: ?
Scenario C: The Ghostwriter for LinkedIn
Goal: A complex system that researches a user's previous 50 posts, finds trending topics in their industry, drafts a new post, critiques the tone for being "Too salesy," and then generates 3 alternate versions.
- Your Choice: ?
- Reason: ?
The Answer Key (The Engineer's Perspective)
Solution A: LangChain or StrandAgents
Why: This is a direct, stateless, non-iterative task. LangChain's existing RSS loaders make this a 10-minute project.
Solution B: LangGraph
Why: This involves Rigid Logic (compliance rules) and Human-in-the-Loop (manager signature). You need the persistence and the "Breakpoint" feature of LangGraph to ensure no laws are accidentally broken.
Solution C: CrewAI
Why: This is a multi-persona task (Researcher, Analyst, Writer, Critic). The "Collaboration" between these personas will produce a much more "Human" LinkedIn post than a single graph node.
Module 11 Summary
- LangChain = Speed & Integration.
- LangGraph = Reliability & Control.
- CrewAI = Performance & Collaboration.
- Hybrid = Enterprise Scale.
Coming Up Next...
In Module 12, we dive into the most discussed problem in AI: Hallucinations. We will learn why they happen and the technical steps to reach 99% accuracy in your agent systems.
Module 11 Checklist
- I can describe the primary "Core Abstraction" for LC, LG, and CAI.
- I identify as a "Control-First" or "Persona-First" developer.
- I understand how to call a CrewAI crew from within a LangGraph node.
- I have chosen a framework for my Capstone Project.
- I understand the relationship between "Cycles" and "Reliability."