Amazon Lex

Overview

Amazon Lex is an AWS service for adding voice and text conversations to applications. It combines automatic speech recognition with natural language understanding, so a bot can process spoken requests as well as text. Common applications include self-service voice assistants, informational and transactional bots, enterprise productivity tools, and device-control bots.

Lex includes tools for designing and evaluating bots, options for connecting them to application logic and identity services, and routes for deployment across devices and messaging services. Its use of AWS integrations and pay-as-you-go request pricing makes the requirements of the intended bot, channels, and workload important factors in evaluating it.

Explore our Chatbot Builders and Conversational AI Platforms lists for more options.

Key features

Bot design and evaluation

The Automated Chatbot Designer analyzes conversation transcripts and proposes an initial design with intents and slot types. Test Workbench lets builders create and run test sets, then inspect aggregated performance results and detailed insights. These tools address two distinct stages: shaping a bot from existing conversations and assessing its behavior against defined tests.

Knowledge-based responses

Lex supports conversational FAQ responses through an intent that queries an authorized knowledge source and foundation models from Amazon Bedrock. This provides a way to draw on connected knowledge when responding to questions.

Integrations and channels

For backend business logic, Lex integrates with AWS Lambda; Amazon Cognito supports user authentication, and Amazon Polly provides text-to-speech. Lex also natively integrates with Amazon Connect. AWS says pre-built integrations are available for other contact center solutions.

Bots can be deployed to mobile devices and chat services including Facebook Messenger, Slack, Kik, and Twilio SMS. The listed platforms are Android, iOS, API, and web, with multichannel deployment.

Pricing

Amazon Lex is paid, with a listed starting price of $0.00 and a free plan. Text requests and speech requests are listed as free, billed pay only for what you use; each text input counts as a text request and each speech input as a speech request. Pricing examples also show free for a text request and a speech request, with no upfront commitment or minimum fee.

The Automated Chatbot Designer is listed at $0.50 a month, while the pricing note specifies a charge of $0.50 per minute of training time for transcript analysis and intent identification. The AWS Free Tier is listed as free: it is available for six months after account creation. New AWS customers receive up to $200 in AWS Free Tier credits, which apply to eligible services including Lex and must be used within 12 months of account creation.

Platforms

The listed platforms are Android, iOS, API, and web. Lex supports deployment across multiple channels, including mobile devices and the named chat services. Amazon Lex V2 supports multiple languages and locales, although some features are limited to particular languages or locales.

For transport security, Lex V2 clients must support TLS 1.0 or later; AWS recommends TLS 1.2 or later. Amazon Lex V2 is HIPAA eligible, and AWS describes it as PCI, SOC, and ISO compliant.

Who it's for

Lex may suit teams building voice or text self-service assistants, informational or transactional bots, enterprise productivity bots, and device-control experiences. Its transcript-based bot design, testing tools, AWS service integrations, and contact-center connections are relevant to projects that need those capabilities. Teams should also account for supported language and locale limits on some features, as well as their channel and security requirements.

Pros and cons

• Pros: Supports voice and text interactions, with tools for transcript-informed bot design and test evaluation.
• Pros: Connects with AWS Lambda, Cognito, Polly, and Amazon Connect, and supports several chat and mobile deployment channels.
• Pros: Offers a free plan and pay-as-you-go request pricing with no upfront commitment or minimum fee in the listed examples.
• Cons: Some Lex V2 features are limited to particular languages or locales.
• Cons: The Automated Chatbot Designer has a training-time charge; the listed monthly figure and per-minute pricing note use different billing descriptions.

Alternatives

Other options to consider include IBM watsonx Assistant, Yellow.ai, AiSensy, Chat Data, Botpress, DocsBot AI, Rasa, and Gupshup.

Verdict

Amazon Lex brings voice and text bot building, transcript-based design suggestions, test evaluation, knowledge-based FAQ responses, and AWS integrations together in one service. Its listed request pricing starts at zero, but the Designer has a training-time charge and the Free Tier credits have an eligibility window. Lex is worth considering when its channels, language coverage, security requirements, and AWS connections align with the application being built.