Controlling behavior of entities in funnel sections of a computer-represented environment

Innovations in the area of controlling behavior of computer-controlled entities such as cars in a computer-represented environment are presented herein. In various examples described herein, the behavior of computer-controlled cars is controlled by an artificial intelligence ("AI") behavior engine. The AI behavior engine performs operations to determine if the behavior (driving pattern, etc.) of a computer-controlled car should change for a funnel section. If so, the AI behavior engine switches the computer-controlled car to a trailing mode configuration and selects a candidate car to follow in the funnel section based on various criteria. The AI behavior engine also determines how to follow the candidate car by selecting a racing path and setting rules to determine the speed to use for following the candidate car. Control values are then set. The described innovations also generally apply to other racing scenarios, self-driving/autonomous cars, and other applications with robotic equipment and devices.