Build a Voice Travel Concierge with Amazon Bedrock
A new architectural blueprint demonstrates how to integrate a real-time voice travel concierge into airline applications using managed cloud services and speech technology.
Radwan AbdulhadiEditor and technology writer
Introduction to the Voice Travel Concierge Architecture
Airlines routinely provide apps and websites that allow travelers to check flights, pick seats, and manage bookings, but incorporating a natural voice layer opens those tasks to spoken requests. By utilizing services like [Amazon Bedrock AgentCore](https://aws.amazon.com/bedrock/agentcore/), developers can deploy secure AI agents at scale while allowing users to change a seat or verify a flight delay entirely by voice without leaving their current screen.
The underlying architecture runs alongside existing screens rather than replacing them entirely, enabling travelers to effortlessly move between tapping and talking within the same session. Furthermore, the AI layer interfaces with sample airline backend infrastructure featuring synthetic data to accelerate implementation cycles when organizations adapt the pattern to their native systems.
Core Technologies and Real-Time Speech Processing
At the heart of the real-time speech capabilities is [Amazon Nova Sonic](https://aws.amazon.com/nova/models/), a speech-to-speech model built to handle conversational audio streams efficiently. Developers looking to implement advanced capabilities can utilize the [Strands Agents](https://strandsagents.com/) framework alongside [Amazon Nova 2.5 Sonic](https://aws.amazon.com/about-aws/whats-new/2026/10/amazon-nova-2.5-sonic/) to handle multi-turn dialogue effortlessly.
This setup allows travelers to speak requests—such as checking itineraries, updating meal preferences, or modifying seating assignments—while the agent manages tool integration and response generation dynamically. Policy questions are simultaneously addressed using [Amazon Bedrock Knowledge Bases](https://aws.amazon.com/bedrock/knowledge-bases/), which functions as a fully managed retrieval-augmented generation service to ground model answers in accurate documentation.
Backend Integration via Model Context Protocol
To ensure secure and decoupled communication between the AI assistant and internal databases, the solution implements the open standard [Model Context Protocol (MCP)](https://modelcontextprotocol.io/docs/getting-started/intro). Standardized messages travel safely between the agent and corporate backend systems through the AgentCore Gateway.
Amazon Cognito manages user authentication and issues temporary credentials, ensuring secure API access throughout the lifecycle of the session. Meanwhile, the [Amazon Bedrock AgentCore runtime](https://docs.aws.amazon.com/bedrock-agentcore/latest/devguide/agents-tools-runtime.html) hosts the agent securely, applying microVM isolation per session to maintain data protection and system reliability.
Deployment and Infrastructure Automation
The entire solution architecture is structured for automated deployment using the [AWS Cloud Development Kit (AWS CDK)](https://aws.amazon.com/cdk/). Multiple CDK stacks handle distinct system layers, provisioning database tables, serverless functions, REST endpoints, and authentication configurations.
User sessions originate from a web application hosted on AWS Amplify. Once authenticated via [Amazon Cognito](https://aws.amazon.com/cognito/), clients establish a signed WebSocket connection to begin their voice concierge session, allowing the application to process voice commands and deliver contextual spoken replies smoothly.
Operational Management and Live Escalation
Beyond routine itinerary updates and policy queries, the voice concierge accommodates scenarios requiring human intervention. If a traveler requests escalation, the underlying business logic logs the request and issues a reference number and estimated wait time.
System monitoring and security are maintained continuously through comprehensive logging, metric collection, and encryption at rest. This robust operational structure enables enterprises to scale their AI assistants confidently before peak travel periods.
Sources
- AWS Machine Learning BlogBuild a voice travel concierge with Amazon Bedrock AgentCore, Managed Knowledge Base and Nova Sonic
Continue chronologically
Related entity coverage
Related stories
Build Multi-Agent Pipelines with Amazon Bedrock AgentCore
Learn how to build a multi-agent music production pipeline on Amazon Bedrock AgentCore Runtime Instances using shared sessions and GPU infrastructure.
GLM 5.3 Launches on Amazon Bedrock for Coding and Agents
Z.ai's 753B-parameter GLM 5.3 mixture-of-experts model is now available on Amazon Bedrock, bringing advanced coding and long-horizon agentic capabilities.
Agentic Retrieval with LangChain and Amazon Bedrock
Explore building Retrieval Augmented Generation applications using LangChain and Amazon Bedrock Managed Knowledge Base for complex multi-part questions.
Query Claims in Natural Language With Amazon Bedrock
Learn how to build a conversational claims assistant using Amazon Bedrock Knowledge Bases to answer natural-language queries with citations.
