AI Implementation Guides

Choose between the options

Long-form guides that walk a single decision end to end, with the trade-offs and the evidence to weigh them against.

• How to Choose a Model for AI Agents (A Decision Guide)

  There's no single best model for agents. Choose by your agent's shape and constraints, weigh tool-calling reliability first, and test on your own tasks.

• Fine-Tuning vs. RAG vs. Long-Context: A Cost-and-Constraints Decision Guide

  RAG, long-context, and fine-tuning answer different questions. Choose by data volatility, corpus size, and cost per query, with the math and evidence shown.

• Structured Data for Technical Content: What JSON-LD Does and Doesn't Do

  What Google and Bing actually say about JSON-LD and AI answers, which schema types fit technical articles, a checker script, and what it found on this site.

Explore guide categories

• model-selection

  Choose models by use case, capability, constraints, and implementation trade-offs.

• provider-selection

  Compare provider options, deployment paths, APIs, pricing factors, and operational constraints.

• benchmark-interpretation

  Understand benchmark results, limitations, evaluation context, and practical relevance.

• workflow-implementation

  Plan and implement AI workflows such as RAG, agents, fine-tuning, and evaluation.

• ai-architecture

  Design AI systems, integration patterns, data flows, and deployment architectures.

• governance

  Apply governance, risk, review, and monitoring practices to AI implementation decisions.

Fast reference cards

Short answers to recurring implementation questions. These are the fast lane of the two-speed content system: useful now, and held to a lower evidence bar than the long-form guides above.

• When to add a reranker to a RAG pipeline

  A reranker helps most when retrieval returns relevant passages but ranks them poorly — not when relevant passages are missing entirely.

• Reading context-window numbers honestly

  A large context window is a capacity limit, not proof that a model uses long inputs effectively.

• Routing simple and hard queries to different models

  A router that sends easy requests to a smaller model and hard ones to a stronger model can cut cost and latency without an across-the-board quality loss — when the routing signal is reliable.

• Structured output: JSON mode vs. function calling

  JSON mode constrains output shape; function/tool calling additionally signals intent and arguments. They solve overlapping but distinct problems.

• Choosing an embedding model for retrieval

  Embedding choice should be driven by your domain, language coverage, and update cost — not by a single leaderboard number.

• Prompt-injection basics for retrieval apps

  Any system that feeds retrieved or user-supplied text into a model can carry instructions an attacker planted in that text.

All guides (16)

Every published guide, including the decision guides nested under each area.

• AI Architecture
• Benchmark Interpretation
• Choose A Model FOR Agents
• Choose AI Model FOR RAG
• Choose Managed API OR Self Hosted
• Compare AI Providers
• Design Enterprise RAG Workflow
• Evaluate AI Model Quality
• Fine Tuning vs RAG
• Fine Tuning vs RAG vs Long Context
• Governance
• Interpret AI Benchmarks
• Model Selection
• Provider Selection
• Structured Data FOR Technical Content
• Workflow Implementation

Continue your research

Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to AI Implementation Guides.

• Browse models
• Review providers
• Inspect benchmarks
• Explore workflows
• Learn AI concepts
• Read decision guides
• Review trust and methodology
• Start a decision path