Amazon Lex is a service for building conversational interfaces into any application using voice and text. With Amazon Lex, you pay only for what you use. There is no upfront commitment or minimum fee. Amazon Lex bots are designed for a request and response interaction or a continuous streaming conversation. With the request and response interaction, each user input (voice or text) is processed as a separate API call. In a streaming conversation, all user inputs across multiple turns are processed in one streaming API call. Pricing for Amazon Lex is based upon these two interaction models.

With streaming conversation, the bot continuously listens and can be designed to respond proactively. For example, you can configure the bot to keep a conversation going when the user needs more time to respond by sending periodic messages such as “Take your time. Let me know once you are ready.” The request and response model is a different user experience where a user input is required as an initiator.

You can try Amazon Lex for free. From the date you get started with Amazon Lex, you can process up to 10,000 text requests and 5,000 speech requests or speech intervals per month for free for the first year. You can also use Automated chatbot designer for two hours of training time per month for first two months.

Request and response interaction pricing

Region
Key terms

Speech request: Each speech input from the user is counted as a speech request.
Text request: Each text input from the user is counted as a text request.

Pricing Example

Consider a bot that processes 8,000 speech requests and 2,000 text requests in one month.

Input requests Cost per request Number of requests Total
8,000 speech requests $0.004 8,000 requests $32.00
2,000 text requests $0.00075 2,000 requests $1.50
Total Amazon Lex charges for the month   $33.50

Streaming conversation pricing

Region

Automated chatbot designer pricing

Region
Key terms

Transcripts: A text file listing conversations between the caller and the agent. Each interaction in the conversation can span multiple lines.
Training time: Time required to analyze transcripts and identify intents.

Pricing examples

Example 1

You are a regional credit union and operate a contact center to help customers with queries and transactions related to their bank accounts. You want create a bot to augment your contact center operations and improve efficiencies. You select the conversation transcripts from customer calls handled by your high performing agents as an input to the automated chatbot designer to create a high-quality bot design. Let’s assume that the calls handled by the top tier agents over a month generate 180K lines of transcripts (assume your contact center has 100 agents, you pick the top 10 high performing agents each answering 20 calls/day; assuming 30 lines of transcript per call, you will have 6K lines daily or 180K lines monthly). The automated chatbot designer takes about 5 hours (or 300 minutes) to analyze the conversation transcripts and surface the design. You are charged for the 300 minutes of training time at $0.50 per minute, leading to total training charges of $150.00 for a month for the 180K lines of transcripts.

Total training time Charges per minute Total charges
300 minutes $0.50 $150.00

Example 2

You are an insurance company with a contact center providing customer support to auto, home, and life insurance policy holders. You want to automate auto insurance conversations with a bot that can help customers with transactions such as making premium payments and filing claims. You use the conversation transcripts from calls with a high customer satisfaction score (CSAT) to ensure high-quality input to the automated chatbot designer. Let’s assume you have 600K lines of transcripts from customer conversations over a week with a high customer satisfaction score (assume a contact center size of 1,500 agents answering a total of 200K calls per week, assume 20K of these calls receive 90% or higher CSAT; with 30 lines per transcript, this results in 600K total lines). The automated chatbot designer takes about 17 hours (or 1,000 minutes) to analyze the conversation transcripts and surface the design. You are charged for 1,080 minutes of training time at $0.50 per minute, leading to total training charges of $540 for the 600K lines of conversation transcripts.

Total training time Charges per minute Total charges
1,000 minutes $0.50 $500.00

Additional pricing resources

AWS Pricing Calculator

Easily calculate your monthly costs with AWS.

Get pricing assistance

Contact AWS specialists to get a personalized quote.