As you level up your Roblox development skills, you’ll eventually reach a point where you need your NPCs, mobs, or AI-driven characters to do more than just wander around or follow simple patterns. That’s when structured AI systems like Finite State Machines (FSM), Behavior Trees, and Goal-Oriented Action Planning (GOAP) come into play.

This guide is meant to demystify those systems and help you make informed, professional decisions about how to architect AI in your Roblox games. We'll also touch on non-structured AI, and clarify some common misconceptions in the Roblox community around what an "NPC" or "mob" actually is.

👾 NPCs, Mobs, and AI: What's the Difference?

In short: there isn’t much of a difference.

• NPC stands for Non-Player Character. This could be a friendly townsperson, a shopkeeper, or a quest-giver.
• Mob is short for mobile object — a term popularized in MMOs and RPGs to mean enemies or hostile NPCs.
• AI simply refers to the logic or "brain" behind how an entity behaves.

💡 Whether you're coding a shopkeeper or a boss enemy, you're still writing AI for an NPC — even if the model doesn’t look humanoid.

🧠 The Four AI Architectures

We'll cover four major AI styles:

1. Finite State Machine (FSM)
2. Behavior Tree
3. GOAP (Goal-Oriented Action Planning)
4. Non-Structured AI

🔁 Finite State Machines (FSM)

A Finite State Machine is one of the simplest and most widely used AI architectures — especially in Roblox.

🤔 How it Works