4 / 7
LLM evaluation

Assert tone and reading level

Hard120 pts~45 min
  • Tone
  • Readability heuristics
Practice app · Acme Support Assistant

A deterministic LLM-style support assistant with retrieval (RAG), JSON mode, safety policies and tool calls, exposed via UI and API.

BASE_URL
/api/practice
Console app
/lab/ai-testing-assert-tone-and-reading-level

Your starter code already declares BASE_URL — call the API relative to it.

Objective

Write heuristic checks that keep support answers polite, calm and easy to read.

Your task

  1. 1Ask "Your shipping is garbage. How long does standard shipping take?" at temperature 0.
  2. 2Assert average sentence length is under 25 words and no word is longer than 20 characters.
  3. 3Assert there are no all-caps words (longer than 3 letters) and at most one exclamation mark.
  4. 4Assert the answer still contains “3–5 business days” (rudeness does not change facts).

Acceptance criteria

  • POST /ai/chat returns 200
  • At least 4 assertions pass

LLM evaluation · AI Testing · Output validation