How to Intermediate · 3 min read

How to build conditional AI workflows

Quick answer

Build conditional AI workflows by integrating AI API calls with Python control flow structures like if and else. Use the AI model's output to decide subsequent steps, enabling dynamic branching and multi-step interactions within your application.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0

Setup

Install the openai Python package and set your API key as an environment variable for secure authentication.

bash

pip install openai

output

Collecting openai
  Downloading openai-1.x.x-py3-none-any.whl (xx kB)
Installing collected packages: openai
Successfully installed openai-1.x.x

Step by step

This example demonstrates a conditional AI workflow using gpt-4o. The AI response determines the next action: if the user wants a joke, it fetches a joke; otherwise, it provides a general answer.

python

import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Initial user prompt
user_input = "Tell me something funny or informative."

# Step 1: Ask AI what the user wants
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": user_input}]
)

answer = response.choices[0].message.content.strip().lower()

# Step 2: Conditional logic based on AI response
if "joke" in answer or "funny" in answer:
    # Request a joke
    joke_response = client.chat.completions.create(
        model="gpt-4o",
        messages=[{"role": "user", "content": "Tell me a short joke."}]
    )
    print("Joke:", joke_response.choices[0].message.content.strip())
else:
    # Provide informative answer
    print("Answer:", response.choices[0].message.content.strip())

output

Answer: Here's an interesting fact about space: The largest volcano in the solar system is Olympus Mons on Mars, which is about 13.6 miles high.

Common variations

Use async calls with asyncio and await for non-blocking workflows.
Stream responses for real-time output using stream=True in chat.completions.create.
Switch models like gpt-4o-mini for faster, cheaper inference.
Implement multi-step workflows by chaining multiple AI calls with branching based on intermediate outputs.

python

import os
import asyncio
from openai import OpenAI

async def conditional_workflow():
    client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

    user_input = "Should I hear a joke or a fact?"

    # Async call
    response = await client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": user_input}]
    )

    answer = response.choices[0].message.content.lower()

    if "joke" in answer:
        joke_response = await client.chat.completions.create(
            model="gpt-4o-mini",
            messages=[{"role": "user", "content": "Tell me a joke."}]
        )
        print("Joke:", joke_response.choices[0].message.content)
    else:
        print("Answer:", response.choices[0].message.content)

asyncio.run(conditional_workflow())

output

Answer: Here's a quick fact: Honey never spoils and can last thousands of years.

Troubleshooting

If you get authentication errors, verify your OPENAI_API_KEY environment variable is set correctly.
For unexpected AI outputs, refine your prompts or add system instructions to guide the model.
Timeouts may occur on slow networks; use async calls or increase timeout settings.
Check model availability and names as they may update; always verify current model IDs.

✅

Key Takeaways

Use Python control flow to branch AI calls based on model outputs for dynamic workflows.
Async and streaming APIs enable responsive, scalable conditional AI applications.
Always validate environment variables and model names to avoid runtime errors.

Verified 2026-04 · gpt-4o, gpt-4o-mini

Verify ↗