Claude Certified Architect – Foundations Practice Question 11: Routing Cheaper Model
Claude Certified Architect, Foundations · Test 2 · question 1 of 10
A support platform receives three kinds of messages: simple FAQ lookups, billing disputes that need account data, and technical troubleshooting. A single large prompt with every instruction and a top-tier model handles all of them, but it is slow and costly for the simple FAQ traffic, and quality on billing disputes suffers because the prompt is bloated. The team wants to improve cost, latency and quality without letting the model decide its own control flow. Which design should the architect choose?
Show answer and explanation
Answer:
A is correct. Routing is a workflow pattern: an LLM call (or a classifier) first classifies the input, then the request is sent to a focused downstream prompt or model. It allows separation of concerns, so each prompt is shorter and tuned to one category, and it lets easy questions go to a smaller, cheaper, faster model while hard ones go to a more capable one. The control flow stays in your code.
B is wrong because a fully autonomous agent gives up predictability and adds cost and latency; the categories are well defined, so a fixed workflow fits better. C is wrong because running all three prompts multiplies cost on every message and adds a selection step, the opposite of the goal (that is closer to parallelization, not routing). D is wrong because it keeps the bloated prompt and the top-tier model on every request, so cost and prompt dilution remain.
Reference: Building effective agents - Routing