M38.2 CONNECT THE MECHANISM
Write questions the model can answer
Tally's first list of routing options splits one worried customer across two near-twin labels, and a safety message has nowhere clear to go. Learn to write options, precedence rules, and thresholds that give usable answers.
LESSON OVERVIEW23 min lesson
Lesson overview
Tally's first list of routing options splits one worried customer across two near-twin labels, and a safety message has nowhere clear to go. Learn to write options, precedence rules, and thresholds that give usable answers.
What you’ll explore
- Decision questions come in three types (choice, score, and yes/no). Good options are distinct, cover the cases, and separate an off-topic message from one with too little information, with precedence rules where routes overlap. A top probability and a provider's confidence are different numbers, and a classification is not authorization to act.
GO TO THE SOURCE
Original explanations, connected to the research.
TypeSafe AI documentation (primitives and patterns)TypeSafe AI documentation, confidence formulas for Choice, Score, and NoulType-Safe Is Not Error-Free: A Constrained Decision Head Follows the Option Name, Not the Rubric Bound to It (arXiv 2609.26758, September 2026 preprint)Introducing System One models and Jev (TypeSafe AI blog, September 2026)Laya model card (Convai Innovations, Hugging Face)Suggest a correction
A precise note can make an explanation better.
Choose the scene and describe what needs attention. Download a feedback file to share through a channel you already use. This page does not send feedback or connect you with a reviewer.