Urgent.News

What's breaking now, across thousands of outlets.

AI

Build a voice travel concierge with Amazon Bedrock AgentCore, Managed Knowledge Base and Nova Sonic

Add a voice travel concierge to an airline app with Amazon Bedrock AgentCore, Amazon Nova Sonic for real-time speech, and Amazon Bedrock Knowledge Bases for policy answers. Travelers change seats, check delays, and ask policy questions by voice, while the agent reaches your backend through MCP tools and confirms every change before it writes.

Travelers can now interact with airlines using natural voice commands through a new voice travel concierge application built using Amazon Bedrock AgentCore, Managed Knowledge Base, and Nova Sonic. This voice layer allows users to change seats, check delays, and access policy information by speaking, all within the existing app interface.

The implementation requires careful engineering to handle audio streaming, maintain conversation continuity, and efficiently connect to backend systems. The solution utilizes several Amazon Web Services (AWS) managed services.

Amazon Bedrock AgentCore provides an agentic platform for building, deploying, and operating AI agents securely at scale. It supports various frameworks and models. Amazon Nova Sonic, another Amazon Bedrock service, enables real-time speech processing.

Amazon Bedrock Knowledge Bases offers retrieval augmented generation capabilities, allowing the concierge to answer questions based on airline policy documents. The concierge operates alongside existing screens, enabling seamless switching between tapping and talking during a session.

A sample airline backend is used for testing, with synthetic data. The project is modular, allowing reuse of backend pieces and exposing them to the agent as tools. The process involves deploying the voice AI concierge on AWS using AWS Cloud Development Kit (AWS CDK), building an agent with Strands Agents framework and Amazon Nova 2.5 Sonic, and connecting the agent to backend services through the Model Context Protocol (MCP) via the AgentCore Gateway.

The architecture separates the front end, AI agent, and backend services into distinct layers for independent development and scaling. Amazon Cognito handles user authentication, while Amazon Bedrock AgentCore runtime and Gateway provide secure connectivity to backend endpoints. AWS Lambda processes business logic, and Amazon DynamoDB stores user data. Amazon Bedrock Knowledge Bases addresses policy inquiries, and Amazon SES sends notifications. AWS Amplify hosts the React frontend.

Written by urgent.news from AWS Machine Learning's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at aws.amazon.com →

More in AI

I Built an AI That Tells You to Stop Using AI and Go Outside 🌿

I Built an AI That Tells You to Stop Using AI and Go Outside 🌿 What if the best AI experience was one that made you stop looking at your screen?

  • NatureQuest AI encourages outdoor activities over screen time
  • Open-weight AI model enables local inference and customization
  • Users disconnect phones to engage in screen-free outdoor quests

More from Tuesday 6 October →