AI Agent Framework Comparison & Selection Playbook

📋 What It Is

An 8-tab interactive decision workbook that transforms Chapter 2's framework landscape — LangChain, LangGraph, CrewAI, AutoGen, Semantic Kernel, AWS Bedrock, Google ADK, Azure AI, and 9 more — into a structured selection process your team can execute in a single architecture session.

This isn't a static comparison chart. It's an operational instrument with 89 live formulas that connect your project constraints, cloud provider, compliance requirements, and team maturity to framework selections to an auto-generated Architecture Decision Record.

The core problem it solves: teams either default to whatever framework the lead used at their last company, or spend weeks building comparison spreadsheets from blog posts. They compare features without connecting them to constraints — selecting LangGraph without realizing their team scores 2/5 on graph concepts. This workbook makes those mismatches visible before a line of code is written.

Introduces the 5-Layer Agent Technology Stack — clarifying protocols (MCP, A2A) vs. model capabilities vs. frameworks vs. platforms vs. products. Includes 4 domain-specific worked examples (SupplyChainIQ, RegAssist, TalentRadar, AuditFlow).

👥 Who It's For

• Solution architects choosing a framework stack — need to justify why LangGraph over CrewAI with constraint alignment and compatibility data
• Engineering leads running architecture sessions — need a structured process that turns debate into documented decisions
• Platform engineering teams evaluating cloud-platform agent services (Bedrock AgentCore, Vertex AI, Azure AI) against open-source alternatives
• AI strategists and consultants recommending framework stacks to clients — need the ADR as a deliverable artifact
• Enterprise architects validating the proposed framework matches cloud provider, compliance posture, and team maturity
• C-level executives (CAIO/CIO/CTO) validating the framework investment matches organizational constraints
• GRC and compliance reviewers verifying frameworks meet enterprise security and audit requirements

⏱ When to Use It

• New agent project kick-off — evaluate the landscape against your constraints while switching costs are zero
• Framework migration — when a prototype framework needs replacement. The Migration Effort Matrix shows the cost of each path
• Cloud provider commitment — evaluating managed platforms (Bedrock, Vertex, Azure) vs. open-source. The Compatibility Matrix shows which combinations work
• Architecture review — retroactively capture and validate undocumented framework decisions
• Annual technology refresh — reassess whether your stack still matches needs as the ecosystem evolves rapidly
• Build-vs-buy evaluation — open-source framework (Layer 3) vs. managed platform (Layer 4) vs. model API (Layer 2)
• Executive review — present the Selection Summary as your Architecture Decision Record

📦 What It Produces

• Architecture Decision Record — auto-calculated stack composition, tier health checks (✓/⚠/🔴), priority breakdown, 10 constraint alignment checks, and 8 architecture requirement coverage checks
• Tier Selection Health Check — dynamic warnings when selections exceed recommended ranges (e.g., "🔴 3+ orchestrators is unusual and expensive")
• Framework × Platform Compatibility Matrix — 14-row cross-reference: ✓ Native, ⚙ Custom integration, ✗ Incompatible
• Enterprise Readiness Scorecard — 25 operational criteria across Security, Scalability, Observability, Cost Management, and Vendor Lock-in Risk
• 5-Layer Agent Technology Stack — visual taxonomy preventing the common mistake of comparing incomparable technologies
• 8 Framework Combination Recipes — proven multi-framework stack patterns (The Starter, The Enterprise Standard, The AWS/GCP/Microsoft Path, etc.)

🚀 How to Use It — Quickstart

• Step 1. Open Project Context. Fill in cloud provider, deployment model, rate 8 constraints (1–5), check 8 architecture requirements (Yes/No). These drive all downstream analysis.
• Step 2. Review the 5-Layer Stack diagram. Understand the difference between protocols, model capabilities, frameworks, platforms, and products.
• Step 3. Switch to Framework Comparison. Walk through 3 tiers: Orchestration frameworks, Retrieval/Knowledge frameworks, Cloud platform agents. For each, review ratings, set Selected/Priority.
• Step 4. Open Enterprise Readiness. Assess your top 2–3 candidates against 25 operational criteria. Flag Critical gaps.
• Step 5. Check Integration & Ecosystem. Verify model provider compatibility and review the 8 Framework Combination Recipes.
• Step 6. Open Selection Summary. Review tier health checks, constraint alignment, enterprise readiness scorecard. Resolve ⚠ gaps before committing.

👁 Preview — What's Inside

17 Frameworks Across 3 Tiers

4 Domain-Specific Worked Examples

📝 Version History