Future Roadmap for AI Agents for Beginners: 3 New Lessons Coming Soon

The Microsoft AI Agents for Beginners repository has three lessons explicitly marked "Coming Soon" in the main README, covering scalable deployment, local agent execution, and security practices, with the modular structure designed to support additional enterprise topics beyond the current 15-lesson curriculum.

The AI Agents for Beginners course by Microsoft provides a comprehensive learning path through self-contained lessons combining written guides, video tutorials, and Jupyter notebooks. While the current curriculum already covers 15 lessons ranging from fundamentals to advanced agent patterns, the repository's future roadmap documented in README.md outlines specific upcoming modules that will bridge the gap between prototype and production environments.

Planned Lessons: The "Coming Soon" Roadmap

The maintainers have designated three critical topics as Coming Soon in the lesson table located at /cache/repos/github.com/microsoft/ai-agents-for-beginners/main/README.md lines 108-115. These modules target specific operational gaps in the current curriculum.

Deploying Scalable Agents

This lesson will address containerisation, serverless architectures, Kubernetes orchestration, and Azure AI Foundry V2 integration. The focus moves from running single agents locally to deploying horizontally-scalable agent services that can handle production workloads. According to the roadmap, this module will demonstrate how to package agents using the Microsoft Agent Framework (MAF) for cloud-native environments.

Creating Local AI Agents

Targeting privacy-conscious and offline scenarios, this module covers running agents without cloud dependencies using tools like Ollama, LM Studio, or on-device inference frameworks. This lesson addresses use cases requiring data sovereignty or low-latency local execution rather than API-based models.

Securing AI Agents

The security-focused module will establish threat modeling, compliance requirements, and production best practices specifically for agentic AI systems. This includes securing agent-to-tool communications, managing credential storage, and implementing guardrails for autonomous agent behaviors.

Sample Code: Scalable Agent Deployment

The repository already contains placeholder infrastructure for the Deploying Scalable Agents lesson. Below is the sample implementation from 16-deploying-scalable-agents/16-python-agent-framework.ipynb demonstrating how to initialise a Microsoft Agent Framework provider for Azure AI Foundry:

# 16-deploying-scalable-agents/16-python-agent-framework.ipynb – First Cell

from agent_framework import Agent, AzureAIProjectAgentProvider
import os

# Load environment (these vars are defined in .env.example)

project_endpoint = os.getenv("AZURE_AI_PROJECT_ENDPOINT")
deployment_name = os.getenv("AZURE_AI_MODEL_DEPLOYMENT_NAME")

# Initialise provider that talks to the Azure AI Foundry Agent Service V2

provider = AzureAIProjectAgentProvider(
    endpoint=project_endpoint,
    deployment_name=deployment_name,
)

# Create a simple travel-assistant agent

travel_agent = Agent(
    name="ScalableTravelAgent",
    system_prompt="You are a helpful travel-assistant that can scale horizontally.",
    provider=provider,
)

print(f"Agent {travel_agent.name} ready – will run in a Docker-based deployment.")

This snippet resides in the 16-deploying-scalable-agents/code_samples/ directory and references environment variables established in the 00-course-setup/.env.example file.

Repository Structure Supporting Roadmap Expansion

The project's architecture facilitates seamless addition of new roadmap items. Key directories include:

• ##-lesson-xx/ (e.g., 01-intro-to-ai-agents/) — Each numbered lesson contains a README.md, code_samples/ folder with notebooks, and supporting images.
• .github/workflows/ — CI pipelines handle automated translation updates and lint checks when new lessons are added.
• translations/ — Auto-generated language packs for 50+ locales sync automatically via the co-op-translator workflow.
• AGENTS.md — Contains detailed architecture descriptions and setup commands referenced by new lessons.

Potential Beyond the Current Roadmap

While the three "Coming Soon" lessons represent the immediate future roadmap, the repository structure supports additional enterprise-focused extensions:

• Multi-Cloud Agent Orchestration — Extending beyond Azure to AWS and GCP implementations.
• Agent-to-Agent Protocols at Scale — Production implementations of MCP (Model Context Protocol), A2A (Agent-to-Agent), and NLWeb standards.
• Observability & Monitoring — Telemetry, tracing, and logging for distributed agent systems.

How to Contribute to the Roadmap

Contributors can accelerate the roadmap through the following process outlined in the repository's contribution guidelines:

1. Open an Issue — Propose a new lesson detailing specific gaps in the current curriculum.
2. Fork and Branch — Create a branch named feature/lesson-xx-<topic> following existing naming conventions.
3. Add Lesson Directory — Follow the established pattern: <number>-<slug>/README.md, code_samples/, and images/ subdirectories.
4. Create Notebook — Use the MAF Python template (*-python-agent-framework.ipynb) for code samples.
5. Update Master Table — Edit README.md to replace "Coming Soon" placeholders with actual lesson links.
6. Submit PR — The CI pipeline will automatically run translation updates and validation checks.

Summary

• The AI Agents for Beginners future roadmap** explicitly lists three upcoming lessons: Deploying Scalable Agents, Creating Local AI Agents, and Securing AI Agents.
• Scalable deployment focuses on containerisation, Kubernetes, and Azure AI Foundry V2 integration.
• Local agent execution will cover Ollama, LM Studio, and on-device inference without cloud dependencies.
• The repository structure in ##-lesson-xx/ directories and .github/workflows/ supports seamless curriculum expansion.
• Contributors can propose new roadmap items via GitHub issues following the CONTRIBUTING.md guidelines.

Frequently Asked Questions

When will the "Coming Soon" lessons be officially released?

Microsoft has not committed to specific release dates for the roadmap items. The repository uses a community-driven development model where contributors can accelerate timelines by submitting pull requests that fulfill the lesson requirements outlined in the README placeholders.

What specific technologies will the local AI agents lesson cover?

According to the roadmap documentation, the Creating Local AI Agents module will demonstrate running agents using Ollama, LM Studio, and other on-device inference frameworks. These tools enable running open-source models locally without requiring cloud API calls or internet connectivity.

Can I suggest topics beyond the three listed roadmap items?

Yes. The repository welcomes proposals for additional lessons covering Multi-Cloud Agent Orchestration, Agent-to-Agent Protocols at Scale (MCP, A2A, NLWeb), and Observability & Monitoring. Submit these suggestions through GitHub issues to discuss alignment with the project's educational goals.

Is the Microsoft Agent Framework (MAF) required for the scalable deployment lesson?

The sample code in 16-deploying-scalable-agents/16-python-agent-framework.ipynb uses the Microsoft Agent Framework with the AzureAIProjectAgentProvider class to interface with Azure AI Foundry. While the lesson may cover alternative frameworks, the reference implementation specifically demonstrates MAF integration with Docker-based deployments.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}