Why agents add autonomy to LLM apps in LangChain - See It in Action

Why agents add autonomy to LLM apps

📖 Scenario: You are building a simple LangChain app that uses an agent to decide which tool to use automatically. This helps the app act on its own without you telling it every step.

🎯 Goal: Create a LangChain agent that can choose between a calculator and a search tool by itself, showing how agents add autonomy to LLM apps.

📋 What You'll Learn

Create a list of tools with a calculator and a search tool

Set up an agent configuration to use these tools

Use the LangChain agent to decide which tool to run based on user input

Run the agent with a sample question to see autonomous behavior

💡 Why This Matters

🌍 Real World

Agents let apps decide what to do next without waiting for user instructions, making apps smarter and faster.

💼 Career

Understanding agents is key for building advanced AI apps that can act independently, a valuable skill in AI development jobs.

1

Create the tools list

Create a list called tools containing Tool(name='Calculator', func=calculator_function) and Tool(name='Search', func=search_function).

LangChain

# Your code here

Need a hint?

Use the Tool class to create each tool with the exact names and functions.

2

Set up the agent configuration

Create a variable called agent by calling initialize_agent with tools and llm, setting agent='zero-shot-react-description'.

LangChain

tools = [Tool(name='Calculator', func=calculator_function), Tool(name='Search', func=search_function)]
# Your code here

Need a hint?

Use initialize_agent with the exact parameters to create the agent.

3

Use the agent to decide the tool

Call agent.run with the string 'What is 12 plus 30?' and assign the result to response.

LangChain

tools = [Tool(name='Calculator', func=calculator_function), Tool(name='Search', func=search_function)]
agent = initialize_agent(tools, llm, agent='zero-shot-react-description', verbose=True)
# Your code here

Need a hint?

Use agent.run with the exact question string and save the answer in response.

4

Complete the autonomous agent setup

Add a final line to print the response variable to show the agent's answer.

LangChain

tools = [Tool(name='Calculator', func=calculator_function), Tool(name='Search', func=search_function)]
agent = initialize_agent(tools, llm, agent='zero-shot-react-description', verbose=True)
response = agent.run('What is 12 plus 30?')
# Your code here

Need a hint?

Use print(response) to display the agent's autonomous answer.