Owlchemy Labs is integrating Artificial Agency’s agentic AI Behavior Engine into Dimensional Double Shift to power a new single-player mode featuring Totally Expendable Multi-dimensional Personnel (T.E.M.P.s) that collaborate with the human player.

Brian Tanner, CEO of Artificial Agency, emphasizes a game-native approach and agent empathy to align agent perception with in-game reality, enabling responsive and adaptive gameplay at practical costs.

Andrew Eiche, CEO of Owlchemy Labs, notes that the integration preserves multiplayer teamwork dynamics while making the experience accessible to solo players, and enables voice-to-action control through agents.

Industry context places the video game market on a path toward $300 billion by 2029, with intense competition and a surge of new releases in 2025, while persistent agents still face economic and performance hurdles.

Dimensional Double Shift launched on Meta Quest in 2024 and on Samsung Galaxy XR in 2025, with multiple expansions through 2026 and reaching a milestone of one million players.

Gamescom playtest faced connectivity issues due to Wi‑Fi traffic, but developers expect latency improvements from better conditions and potential on-device AI processing to reduce delays.

Dimensional Double Shift Single Player Mode is planned to ship in the first half of 2027.

Artificial Agency was founded in 2023 by Brian Tanner and Alex Kearney, both with DeepMind backgrounds; it raised $16 million in 2024 from investors including Radical Ventures and Toyota Ventures, and is based in Edmonton, Canada.

Artificial Agency positions this as the first shipped game built with its Behavior Engine, intended to prove viability and scalability of agentic behavior in modern gaming and beyond.

Artificial Agency trained a small, game-specific model with a reflex architecture to separate fast reactions from deeper reasoning, reducing latency and inference costs and enabling potential on-device processing for broader platform support.

The Gamescom demo showcased AI agents performing tasks in a cooperative coffee-shop scenario with natural-language requests, improviseable actions, and personality traits, responding to player input without scripted sequences.

The collaboration is framed as a milestone showing agentic behavior can redefine game design and player experience, serving as proof of concept for commercially viable agentic gameplay.