Amazon Lex is a managed AWS service for adding conversational interfaces to applications through voice and text. Built on the same technology that powers Alexa, it supports chatbots, voice assistants, machine learning-powered IVR and virtual agents. Developers can describe a bot in natural language to generate a baseline version, use the automated chatbot designer to build from existing conversation transcripts, or work in the Visual Conversation Builder. Billing follows usage: a speech request costs $0.004 and a text request costs $0.00075, with no upfront commitment or minimum fee.
Developers and solution architects already working in AWS are the natural users, since bots plug into the rest of the AWS ecosystem. Contact center managers use it to automate routine calls and chats and route customers to the right human agent. Enterprise IT and HR teams can build assistants that answer frequently asked questions about technical support, benefits or finance, and product teams can add natural language task completion to existing apps. Business analysts can contribute to bot design through the visual builder, so conversation design does not rest only on engineers.
Customer service teams deploy self-service bots across IVR, chat and SMS so callers resolve queries without waiting for an agent. Published results include 80% IVR containment at National Australian Bank, a 65% increase in self-service rate at nib Health Insurance, and a 30% reduction in agent call volume at WaFd Bank. Intelligent routing bots collect context before handing customers to the right agent through Amazon Connect. Applications can let users make a booking, raise a ticket, generate a report or submit an application by voice or chat, and bots can be deployed to Facebook Messenger, Slack, Kik and Twilio SMS.
