When we think of intelligent agents, we often jump straight to LLMs. But what if you could build smart, explainable workflows using pure logic and modular tools—no LLM required?
In this quick demo, I used the Microsoft Agent Framework to create a clean, logic-first pipeline using three lightweight agents:
The Agent Chain
Parser Agent – Extracts the numeric value and unit from natural input
Input: "Convert 25 degrees Celsius to Fahrenheit"
Output: { "value": 25, "unit": "Celsius" }
Converter Agent – Applies the formula
F = \left(\frac{9}{5} \cdot C\right) + 32- Output: { "value": 77, "unit": "Fahrenheit" }
Explainer Agent – Generates a human-readable explanation
Output: "25°C is equal to 77°F. This conversion uses the formula F = (9/5 × C) + 32."
Watch the Full Demo
Want to see how this works in action—step by step?
Watch the video here to explore the full agent flow and how you can build similar logic-first systems using Microsoft’s Agent Framework.
https://youtu.be/G2I0JbGZHZI?si=SyCbMOaZEpq6NymH
