​
Healthcare Voice Agent Monitoring with TraceRoot.AI

This guide explains how to monitor your existing healthcare voice agent built with LangChain and LangGraph using TraceRoot.AI’s telemetry collection capabilities. The TraceRoot.AI platform not only offers a user-friendly interface for visualizing tracing, logging, and metrics data but also features a customizable AI agent that can resolve related issues or answer questions by linking directly to the source code.

​
1. Setup Healthcare Voice Agent Project

git clone https://github.com/traceroot-ai/traceroot-examples.git
cd traceroot-examples
python3 -m venv venv
source venv/bin/activate
pip install -r examples/healthcare_voice_agent/requirements.txt

​
2. Setup TraceRoot.AI

To get started, we need to set up TraceRoot.AI. 
pip install traceroot==0.0.4a5
Set required credentials given by TraceRoot.AI. Put the credentials and information in a yaml file called .traceroot-config.yaml in the root of your project as described in the Quickstart Here is an example of a .traceroot-config.yaml file: .traceroot-config.yaml

​
3. Healthcare Voice Agent System Introduction

In this project, we have a healthcare voice agent system that is built with LangChain and LangGraph for a coding agent. Heathcare Voice Agent System Then system contains five parts:
  1. A speech to text agent that transform the patient’s question to text.
  2. A planning agent the plans the response in general.
  3. A doctor search agent that schedules potential appointments with the patient.
  4. An response agent that takes in the output of planning agent and the doctor appointments agent and general the final response to patients.
  5. A text to speech agent that transforms the final response to voice.
You can run the project by: 
python examples/healthcare_voice_agent/create_input_audio.py  # Generate test input
python examples/healthcare_voice_agent/main.py  # Run the system

​
4. Heathercare Voice Agent System Pain Points

  1. The healthcare voice agent system is complex and hard to debug. Every agent’s input and output are very messy and hard to read.
  2. It’s hard to understand the multi-agent system tool usage such as the code execution in this healthcare voice agent system.
  3. It’s unknown which part of the system has the highest latency.
  4. It’s inconvenient to correlate the logs and traces with the code source.

​
5. TraceRoot.AI Solution

By just adding TraceRoot.AI tracing and logging to the healthcare voice agent system with several lines of code 
logger = traceroot.get_logger()

@traceroot.trace(TraceOptions(trace_params=True, trace_return_value=True))
def func():
    logger.info("Hello, world!")
    pass
We can get the following benefits:
  1. The TraceRoot.AI provide an easy to use SDK to track all tracing, logging and metrics data with just a few lines of code change. examples/healthcare_voice_agent/main.py
  2. TraceRoot.AI provides a user-friendly UI to visualize the tracing, logging and metrics data.
You can view the whole tracing and logging in a structured way Tracing and Logging UI 1 You can click the Show Code button to view the code that is related to the logging. Tracing and Logging UI 2 You can also view which agent has the highest latency. Tracing and Logging UI 3
  1. TraceRoot.AI has a customized AI agent to help you summarize, analyze, and debug all your logs and traces by resolving aforementioned pain points directly.
Tracing and Logging UI 4 You can also ask TraceRoot.AI agent to analyze the final response generated by the multi-agent system Tracing and Logging UI 5