AI and Behaviour Trees

Unreal includes more built-in AI framework structure than many Unity projects use by default.

Unity Scripted AI vs Unreal AI Framework

Unity	Unreal	Notes
Script on NPC object	`AIController` plus Pawn/Character	Logic and controlled body are more clearly separated.
Custom decision code	Behaviour Tree	Common high-level decision system.
Shared blackboard data script	Blackboard	Stores AI working memory.
NavMesh agent	Navigation system and path following	Unreal nav is tightly integrated with AI controllers.

`AIController`

An AIController possesses a pawn the same way a player controller does.

It often handles:

behaviour tree startup
perception callbacks
target selection
move requests

Behaviour Trees

Behaviour Trees define high-level decision flow using nodes such as:

selectors
sequences
tasks
decorators
services

Blackboards

Blackboards store AI state, such as:

current target
patrol point
last known player location
combat state

Perception

Unreal AI Perception can detect stimuli such as sight and hearing.

This is often easier to scale than custom overlap checks on every NPC.

Unreal commonly uses a nav mesh volume for AI movement.

If AI will not move, check:

nav mesh coverage
possession
movement component setup
behaviour tree state

C++ Integration Points

Common C++ entry points include:

custom AIController classes
custom behaviour tree tasks
perception callbacks
helper functions on the controlled pawn

Common Mistake

A common trap is putting all AI state on the pawn and skipping the controller layer. In Unreal, controller and pawn separation is often worth keeping.