CClaude Cert Prep
CCAR-P · cheatsheet

When To Use What — Decision Cheat-Sheet

The near-duplicate choices the exam loves, with a one-line rule for each.

ChoicePick the first when…Pick the second when…
RAG vs long-contextThe corpus is large, changing, or access-controlledThe context is small, stable, and fits the window
MCP vs direct function-callingYou want standardized, reusable, scoped tool access at scaleIt's a one-off internal call with no reuse
Sync vs Message BatchesIt's in a user's blocking pathIt's bulk and can wait ≤24h (50% cheaper, supports tools)
Single-agent vs multi-agentThe work fits one context windowTruly independent subtasks justify the coordination cost
LLM-as-judge vs human reviewHigh volume, calibrated rubric, cost-sensitiveHigh stakes, low volume, or judge can't be trusted yet
Canary vs blue-greenYou want gradual exposure with fast rollbackYou need instant cutover with a warm standby
Prompt vs deterministic controlIt's guidance the model can reasonably followIt's a must-hold rule — enforce it in code

If you remember one thing

Match the mechanism to the constraint, and put must-hold rules where they can be guaranteed — not in a prompt that only asks.

All CCAR-P references

Independent, unofficial study material from Claude Cert Prep. Not affiliated with Anthropic.