Machine Learning
Amazon Lex
Amazon Lex builds conversational voice and text interfaces using intents, slots, prompts, and fulfillment. It can provide deterministic dialog collection before handing complex language work to a foundation model.
Key points
- Bots manage multi-turn slot elicitation and confirmation
- Lambda hooks validate input and perform fulfillment
- Speech recognition and text-to-speech support voice experiences
When to use it
- Collect verified account details before a support assistant responds
- Route contact-center intents to bounded GenAI workflows
Exam tips
- Choose Lex for structured conversational intent flows and Bedrock for open-ended generation
- Validate slots server-side; a recognized utterance is not authorization for a business action