Artificial Agency Secures $16M to Revolutionize Game NPCs

Artificial Agency, a startup founded by former Google DeepMind researchers, has emerged from stealth mode with $16 million in funding. The Edmonton-based company aims to move beyond traditional, script-heavy non-playable characters (NPCs) by deploying an AI-driven behavior engine designed to make virtual interactions more dynamic and responsive.

The company secured $12 million in a seed round co-led by Radical Ventures and Toyota Ventures, following a previous $4 million pre-seed investment from Radical Ventures. Additional participants in the latest round included Flying Fish, Kaya, BDC Deep Tech, and TIRTA Ventures.

Moving Beyond Static Scripts

Modern gaming typically relies on rigid decision trees and pre-written dialogue for NPCs, which often leads to repetitive and predictable gameplay. Artificial Agency seeks to change this by shifting the developer’s role from a writer of specific outcomes to a “stage manager.”

Under this new framework, developers assign NPCs specific motivations, goals, and rules. The engine then calculates how the character should act in real-time based on player behavior. This system can be integrated into existing titles or utilized as the foundation for entirely new games.

Demonstrating AI Autonomy

In a demonstration using Minecraft, the startup showcased an NPC named “Aaron” that was tasked with being helpful. Beyond simple conversation, the character autonomously gathered specific equipment—such as armor and food—from storage chests to assist the player in a mining task. When the player mentioned a dietary restriction, the NPC adjusted its offer, swapping bread for cooked chicken without needing custom programming for that specific scenario.

Co-founder and CEO Brian Tanner notes that this level of interaction is becoming a priority for the industry:

• Industry Shift: Tanner believes dynamic responses will be “table stakes” in the gaming industry within a few years.
• Cost Efficiency: While a five-minute demo currently costs roughly $1 in inference fees, Tanner expects that to drop to one cent or less within a year due to GPU and model optimizations.
• Model Usage: The startup currently utilizes open-source models, including Meta’s Llama 3, to power its behavior engine.

Industry Adoption and Challenges

Despite the potential, the integration of generative AI into AAA games faces skepticism. Ethical concerns and the risk of “hallucinations” have made some studios hesitant. According to the 2024 State of the Game Industry report, only 21% of developers anticipate a positive impact from generative AI, with 42% expressing ethical concerns.

However, Daniel Mulet, an investor at Radical Ventures, argues that the demand exists because developers are tired of building these complex systems from scratch. “Once you see like 20, 30 groups that are trying to build this themselves, there is an opportunity to build a platform and make it available to everyone,” Mulet said.

Artificial Agency has confirmed it is currently working with “several notable AAA studios” and expects its technology to reach the market in 2025. Whether these AI features will be included in standard game prices or require a premium subscription remains an open question for the industry.