Voice Live API: AI Sales Coach
A demo application showcasing AI-powered voice training for sales professionals, built on Azure.
!Voice Live API Salescoach in Action
Overview
Voice Live API Salescoach is a demo application showcasing how AI-based training could be used in sales education using Azure AI services. Practice real-world sales scenarios with AI-powered virtual customers, receive instant feedback on your performance, and improve your sales skills through immersive voice conversations.
Features
- Real-time Voice Conversations - Practice sales calls with AI agents that respond naturally using Azure Voice Live API
- Performance Analysis - Get detailed feedback on your conversation skills
- Pronunciation Assessment - Improve your speaking clarity and confidence with Azure Speech Services
- Scoring System - Track your progress with metrics
Demo
See the Voice Live API Salescoach in action:
https://github.com/user-attachments/assets/904f1555-6981-4780-ae64-c5757337bcad
How It Works
- Choose a Scenario - Select from various industry-specific sales situations
- Start the Conversation - Click the microphone to start your simulation
- Engage with AI - The virtual customer responds realistically based on the scenario
- Receive Feedback - Get instant analysis on your performance including:
Getting Started
Deploy to Azure
The deployment creates Azure AI and Speech resources and a Container App. It deploys
gpt-4.1-mini for conversation analysis and scenario generation. The app does not
require an embedding model. Select a region that supports Voice Live and the model
deployment, such as Sweden Central, and confirm that your subscription has enough
model quota.
The container builds the frontend with Node.js 24 LTS and uses Python 3.11 on
Debian 12 (Bookworm), which the Speech SDK supports.
Container builds verify HTTPS certificates when downloading npm and Python packages.
If your network uses a private certificate authority, configure its trusted CA
certificate instead of disabling verification.
The application uses Voice Live SDK 1.3 with the GA 2026-07-15 API, Speech SDK
1.51, and OpenAI SDK 3 with the GA /openai/v1/ endpoint. Azure OpenAI keys and
Microsoft Entra ID authentication are supported. Token authentication refreshes
credentials automatically.
- Deploy to Azure:
azd up- Access your application:
Local Development
This project includes a dev container for easy setup and a build script for development.
For development without a container, use Python 3.11 or 3.12 and Node.js 24 LTS.
The build script installs the committed npm lockfile with npm ci; it does not
delete or regenerate the lockfile.
TypeScript 7 compiles the frontend. Microsoft's TypeScript 6 compatibility package
provides the JavaScript compiler API required by the current ESLint parser, using
the official side-by-side setup.
- Use Dev Container (Recommended)
- Fill in the .env file
.env.template to .env
- Fill in your Azure AI Foundry and Speech service keys and endpoints (you can run azd provision to create these resources if you haven't already)
- Build and run
# Build the application
./scripts/build.sh
# Start the server
cd backend && python src/app.py
Visit http://localhost:8000 to start training!
Optional Foundry agents
Instruction-based scenarios remain the default. To use Foundry agents, set
USE_AZURE_AI_AGENTS=true and PROJECT_ENDPOINT to an existing Foundry project
endpoint, such as https://.
The deployment templates do not create this project. The application identity
must have the required Foundry access, and scenario models must be deployed in
that project.
The application creates prompt-agent versions with Projects SDK 2 and connects
Voice Live to the returned agent name and version. Agent connections use
Microsoft Entra ID, even when an OpenAI API key is configured. For connections
without a scenario agent, AGENT_NAME selects an existing agent and
AGENT_VERSION optionally selects its version. AZURE_AI_PROJECT_NAME is
inferred from the project endpoint when omitted. AGENT_ID remains available for
legacy agents and is not required for newly created agents.
See the Microsoft Learn Foundry voice-agent quickstart for project and role setup.
Architecture
|
The application leverages multiple Azure AI services to deliver real-time voice-based sales training:
|
Contributors
Contributing
This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.
When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.
This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.
Security
Microsoft takes the security of our software products and services seriously, which includes all source code repositories managed through our GitHub organizations, which include Microsoft, Azure, DotNet, AspNet and Xamarin.
If you believe you have found a security vulnerability in any Microsoft-owned repository that meets Microsoft's definition of a security vulnerability, please report it to us as described in SECURITY.md.
Trademarks
This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies. Any use of third-party trademarks or logos are subject to those third-party's policies.
Made with ❤️ in 🇨🇭