How to beginner · 3 min read

Semantic Kernel concepts explained

Quick answer

The semantic_kernel is a Python SDK that provides a modular framework to build AI applications by combining kernels, skills, and AI services. A kernel manages AI models and orchestrates skills, which are reusable units of AI logic, enabling structured and composable AI workflows.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install semantic-kernel openai

Setup

Install the semantic-kernel Python package and set your OpenAI API key as an environment variable.

Run pip install semantic-kernel openai
Set environment variable OPENAI_API_KEY with your OpenAI API key

bash

pip install semantic-kernel openai

Step by step

This example shows how to create a Kernel, add an OpenAI chat completion service, define a simple skill with a prompt, and invoke it.

python

import os
import semantic_kernel as sk
from semantic_kernel.connectors.ai.open_ai import OpenAIChatCompletion

# Initialize kernel
kernel = sk.Kernel()

# Add OpenAI chat completion service
kernel.add_service(OpenAIChatCompletion(
    service_id="chat",
    api_key=os.environ["OPENAI_API_KEY"],
    ai_model_id="gpt-4o-mini"
))

# Define a simple skill function
prompt = "Tell me a joke about {topic}."
skill = kernel.create_semantic_function(
    prompt_template=prompt,
    skill_name="JokeSkill",
    function_name="TellJoke"
)

# Invoke the skill
result = skill.invoke_async({"topic": "programming"}).result()
print("Joke:", result)

output

Joke: Why do programmers prefer dark mode? Because light attracts bugs!

Common variations

You can use different AI services like Azure OpenAI or Anthropic by adding their connectors to the kernel. Skills can be composed into workflows, and you can use synchronous or asynchronous invocation. You can also load skills from external files or directories for modularity.

python

import asyncio

# Async invocation example
async def async_example():
    result = await skill.invoke_async({"topic": "AI"})
    print("Async joke:", result)

asyncio.run(async_example())

output

Async joke: Why did the AI go to art school? Because it loved to draw conclusions!

Troubleshooting

If you see authentication errors, verify your OPENAI_API_KEY environment variable is set correctly.
If prompts fail, check your prompt template syntax and ensure the AI model ID matches available models.
For network issues, confirm internet connectivity and firewall settings.

✅

Key Takeaways

Kernel is the core orchestrator managing AI services and skills.
Skills are reusable AI prompt functions that encapsulate logic.
You can add multiple AI services to the kernel for flexible model usage.
Invoke skills synchronously or asynchronously for integration flexibility.
Proper environment setup and API keys are essential for smooth operation.

Verified 2026-04 · gpt-4o-mini

Verify ↗