Motive Engine - Explainable AI Decisions for NPCs, Factions & Narrative

Your NPC just did something strange. Can you find out why in under a minute?

With hardcoded conditionals and deep behaviour trees, usually not. Motive Engine makes that question answerable by design: every decision comes back with the full reasoning attached - strongest contributors, trait effects, the runner-up, and how close the call was.

Motive Engine is an explainable Utility AI decision layer. You author motives (what an agent cares about), it scores them against live game context and per-agent traits, ranks the alternatives, and returns a response key. Your game decides what that key does. Motive Engine deliberately does not execute behaviour - no movement, pathfinding, animation, or behaviour trees - so it slots in front of the systems you already have (state machines, NavMesh, Unity Behavior graphs, dialogue).

Built for the places where opaque decision logic hurts the most: tactical combat, horror and stealth NPCs, faction pressure, and narrative or dialogue-intent systems. Four sample Decision Lab scenes ship in the package, plus execution-bridge examples for NavMesh movement, Unity Behavior graphs, and Behavior Designer.

The live Motive Debugger window shows every decision as it happens: ranked motives, score gaps, strongest contributors, trait effects, a context snapshot, and scrubbable decision history.

NEW - AI Skills for agentic workflows: the package now ships AI Skill instruction files (SKILL.md) that teach AI coding assistants (Unity AI, Claude Code, Cursor, and similar) exactly when and how to integrate, author, and debug Motive Engine. Point your AI tool at the package's "AI Skills" folder and it will drive your asset instead of reinventing it.

One authored ruleset also runs in Unreal and JavaScript with reference-tested scoring parity - author once, decide identically everywhere.