Generative AI Apps and Agents

Implement generative AI apps and agents by using Foundry

What is “Foundry” in this context?

This refers to Azure AI Foundry, Microsoft’s environment for building generative AI applications and agents.

Think of it as: A workspace where you design, configure, test, and deploy AI-powered apps and copilots

1. Generative AI Applications
Applications that use AI models to generate content:
Text (chatbots, summaries, emails)
Code
Images
Key idea: The app sends a prompt to a model and returns a generated response

2. AI Agents (Agentic AI)
Agents go beyond simple prompts—they can:
Take actions
Use tools or APIs
Perform multi-step tasks
Example: An AI agent that:
Reads emails
Extracts tasks
Creates calendar events

3. Prompt Engineering
You control behaviour using prompts:
System message → defines role (e.g. “You are a helpful assistant”)
User prompt → actual question
Context/grounding data → improves accuracy

4. Grounding (Using Your Data)
Improves responses by connecting AI to trusted data sources:
Documents
Databases
Knowledge bases
Prevents hallucinations and increases relevance

5. Tools and Connectors
Agents can interact with:
APIs
Databases
External systems
Enables real business automation

6. Safety & Responsible AI Controls
Built into Foundry:
Content filtering
Moderation
Usage policies
Ensures outputs are safe and compliant

7. Evaluation & Monitoring
Test prompts and responses
Track performance
Improve over time

8. Deployment
Publish apps and agents
Integrate into:
Websites
Apps
Business workflows