Lex

The intent-and-slot mental model behind Amazon Lex V2 bots via SDK - how structured, schema-driven dialog differs from open-ended LLM chat, and when each fits.

Build and test a minimal Amazon Lex V2 bot with one intent via SDK - create the bot and locale, define an intent with sample utterances, build, publish an alias, and test it with RecognizeText.

Designing structured conversation flows via SDK - custom slot types with enumeration values, slots that attach a type to an intent, required-slot prompts, and resolution strategies.

Sending and receiving runtime messages via SDK with RecognizeText and RecognizeUtterance, and managing sessionState (dialogAction, slots, sessionAttributes) turn by turn.

A checklist for building and running Amazon Lex V2 bots from the SDK safely - intent/slot design, dialog-and-fulfillment reliability, session and cost management, and IAM security.