Goal-Oriented Action Planning (GOAP) is an approach to game-character AI where, instead of scripting every situation by hand, the designer gives a character a set of goals and a library of actions, each defined by preconditions and effects. At runtime the character searches for a sequence of actions that turns the current world state into one that satisfies a goal. The search is run with A*, the same algorithm normally used for pathfinding, except here it plans through the space of world states rather than map positions.
GOAP was introduced to game development by Jeff Orkin, who built it for Monolith Productions’ 2005 shooter F.E.A.R. while also a graduate student at the MIT Media Lab. Orkin’s GDC 2006 paper “Three States and a Plan: The A.I. of F.E.A.R.” explains that the planning system “most closely resembles the STRIPS planning system from academia” - STRIPS being the 1970 Stanford planner. The key trade he describes: “An FSM tells an A.I. exactly how to behave in every situation. A planning system tells the A.I. what his goals and actions are, and lets the A.I. decide how to sequence actions.”
Orkin’s motivation was practical. F.E.A.R. had “only one A.I. programmer, but there are lots of A.I. characters,” and the complexity of combining behaviors like taking cover, suppressing fire, and flanking was becoming “unmanageable.” Delegating the sequencing to a planner let the small team get emergent, coordinated squad combat without hand-authoring every case. The result became one of the most praised enemy AIs in shooter history and made GOAP a standard reference technique.
Why business readers should care: GOAP is a clean example of declarative versus procedural design - telling a system what you want and the rules of the world, then letting search find the how, rather than enumerating every path yourself.