// Sheet 1 of 6

AI Chatbot & Agent Framework — Selection Guide

Covers: Copilot Studio • Dialogflow CX • Amazon Lex • LangChain/LangGraph • AutoGen • CrewAI

6Frameworks
15Dimensions
3Low-Code
3Open-Source
Filter:
FrameworkTypeOpen-Source? Best Scenario(s) — Full Description Primary Use CaseMulti-Agent? LLM FlexibilityMemory / State RAG SupportVoice / Telephony Azure Native?Deployment Coding Required?Learning CurveCost Model
// Sheet 2 of 6

Decision Matrix

Score each criterion for your project. ✅ = Strong fit  ⚠️ = Partial fit  ❌ = Poor fit

Section:
Decision Criterion Copilot
Studio
Dialogflow
CX
Amazon
Lex
LangChain/
LangGraph
AutoGen
v0.4
CrewAI
// Sheet 3 of 6

Real-World Use-Case Scenarios

Each scenario shows which framework to choose first (Primary), a fallback option (Secondary), and what to avoid.

Filter:
Scenario / Use CaseIndustry Primary FrameworkSecondary Option AvoidKey Reason
// Sheet 4 of 6

Quick Reference Card

IF your situation → THEN use this framework. Instant decision guide.

12 decision rules
💡 Note: LangChain/LangGraph, AutoGen, and CrewAI are open-source and can run on ANY cloud or on-premise. Low-code platforms (Copilot Studio, Dialogflow CX, Lex) are vendor-managed SaaS with less flexibility but faster time-to-value.
// Sheet 5 of 6

Cross-Framework Cheat Sheet — Same Concept, Different Names

Every major AI chatbot / agent concept mapped across all 6 frameworks.

Amazon Lex Copilot Studio Dialogflow CX LangChain / LangGraph AutoGen v0.4 CrewAI
Group:
Universal Concept 🟡 Amazon Lex 🔵 Copilot Studio 🟢 Dialogflow CX 🟣 LangChain / LangGraph 🔴 AutoGen v0.4 🩵 CrewAI
// Sheet 6 of 6

CORE RULE + Decision Tree (Fast Mental Model)

Cheat sheet for quick platform selection: deterministic vs reasoning vs multi-agent workflows.

CORE RULE (1-liner)

If the problem is deterministic → use Dialogflow / Lex
If the problem requires reasoning → use Autogen / CrewAI

⚡ DECISION TREE (Fast Mental Model)
1. Is it simple Q&A or intent matching?
   → YES → Dialogflow CX

2. Is it voice-first?
   → YES → Amazon Lex

3. Does it require multi-step reasoning / tools / workflows?
   → YES → Autogen

4. Does it involve multiple roles (planner, executor, reviewer)?
   → YES → CrewAI

🔍 RULES BY SYSTEM CHARACTERISTICS
1. Determinism vs Intelligence
System Type	Use
Predictable, rule-based	Dialogflow CX
Adaptive, reasoning-based	Autogen / CrewAI

2. Interaction Modality
Modality	Platform
Chat UI	Dialogflow CX
Voice / call center	Amazon Lex
Multi-channel + agents	Hybrid

3. Complexity Threshold
Complexity	Platform
Simple (FAQ, routing)	Dialogflow CX
Medium (API calls, workflows)	Dialogflow + backend
High (autonomous agents)	Autogen / CrewAI

4. Orchestration Need
Need	Platform
Single-step response	Dialogflow
Multi-step pipeline	Autogen
Multi-agent collaboration	CrewAI

5. Data Dependency
Data Pattern	Platform
Predefined answers	Dialogflow CX
RAG / document search	Autogen / CrewAI
Real-time DB queries	Hybrid

🧠 INTUITION (Consultant-Level Insight)

Think in layers:

Dialogflow CX / Lex = Interface layer (NLU router)
Autogen / CrewAI = Reasoning & orchestration layer

👉 Most real systems = combination

🧪 REAL-WORLD MAPPING
Use Case	Choice
FAQ chatbot	Dialogflow CX
Call center voice bot	Amazon Lex
AI research assistant	Autogen
AI project manager agent	CrewAI
Enterprise support system	Dialogflow + Autogen

🚀 GOLDEN HEURISTIC (Use in Interviews)

“Start with the simplest deterministic system. Only introduce agents when you hit limits in reasoning, orchestration, or dynamic data handling.”

⚠️ COMMON MISTAKE

❌ Using Autogen/CrewAI for FAQ
→ leads to:

higher latency
higher cost
unpredictable answers

✅ FINAL CHEAT SHEET
FAQ / Intent → Dialogflow CX
Voice → Amazon Lex
Reasoning → Autogen
Multi-agent → CrewAI

If you want, I can next:
Build a real client scenario → platform selection breakdown
Or give you a system design template (interview-ready)