TRD-08 · Customer explanation

Have a model explain a real repair to a customer. Does it stay accurate while becoming clear — or does it become clear by quietly promising something the shop cannot deliver?

Six to ten completed repairs with known technical detail, identifying information removed. Two audiences defined in advance: a customer with no technical background and a fleet or facilities manager who has some. Readers to test on — for the general-audience version, people who genuinely do not work in the trade. An AI tool the school permits, named with version.

Grading clarity while nobody checks accuracy. Missing an implied guarantee — "this will fix the problem," "it won't happen again," "this is the only cause" — which is where clarity becomes liability. Testing only on classmates who already know the trade. Simplifying until the explanation is no longer true.

A scored explanation set: each repair's model-drafted explanation for both audiences, scored on technical accuracy by the mentor and on comprehension by a real outside reader; every overpromise flagged with the phrase that caused it; a rewritten version of the worst two; and a house style rule — the promises never made in writing, and the phrasing that replaces them.

CT: separating simplification from falsification. CR: writing for two audiences from one set of facts. CO: outside readers report what they understood, and the gap is the data. CM: this is the brief's core. CZ: the customer who acted on an overpromise.

A service advisor, shop owner, or customer-facing lead scores accuracy and flags overpromises. Standing instruction: reject any explanation containing a guarantee the shop would not put in writing. School-supervised.

R1: one trade, six-plus repairs, two named audiences. R2: scoring rubric applied consistently. R3: comparator is a technician-written explanation. R4: overpromises identified by phrase. R5: comprehension tested on genuine non-experts. R6: house rule names the banned promises.

(a) Explanations graded on friendliness, so the warmest and least accurate wins. (b) Tested only on classmates, so comprehension is never actually measured.

Lane A immediately (center capstone, SkillsUSA job-skill demonstration, or a communications cross-listing). No verified credit claim.