Reshaping the game - 2026 report
The agentic AI revolution: from scripted puppets to autonomous actors
Agentic AI has moved from concept to shipped games - autonomous teammates, adaptive adversaries and self-testing systems now operate inside live titles.
While generative AI provides the static assets for a game, agentic AI provides the dynamic behaviour. This second pillar of the AI revolution marks a fundamental evolution from AI as a tool that generates content to AI as an autonomous system that acts within the game world.
The core concept: perception, reasoning, and action
At the heart of an agentic system is a continuous operational cycle of perception, reasoning and action, allowing for its independent behaviour.
- Perception: The agent collects data from its environment, such as the current game state, the player's position and inventory, or the behaviour of other in-game entities.
- Reasoning: It analyses that data, updates its objectives and formulates a plan of action to achieve its long-term goals. For example, it may evaluate whether a game is becoming too difficult or too easy for a particular player.
- Execution and learning: The agent executes a chosen action, monitors the outcome and adjusts its future behaviour based on what it learns. This could include lowering the difficulty level, by weakening enemy non-playable characters, when it detects that a player is becoming frustrated.
For those interested in the application of agentic AI in industries beyond gaming, we recommend our separate agentic AI guide.
Key application scenarios
The impact of agentic AI becomes clearer when viewed through practical applications. What the first edition of this guide described as an emerging possibility is, by 2026, live in commercially released games.
Emergent narratives and living NPCs
One of the most significant applications of agentic AI is the creation of more dynamic characters and stories. Instead of relying solely on pre-scripted narratives, storie can emerge dynamically from the actions of autonomous characters. NPCs are no longer static quest dispensers with limited dialogue trees. Equipped with persistent memories, they can remember past interactions, pursue independent goals, and form relationships with others. This can drive unscripted events and create a world that feels genuinely alive. A major life-simulation title now features co-playable characters designed to plan, act, and reflect, illustrating how living NPCs have moved on from experimental concepts.
The autonomous teammate and the adaptive adversary
Agentic AI is also being used to create companions and opponents that react intelligently to changing circumstances.
- Autonomous teammates
AI-controlled squad members can communicate with players in natural language, make independent tactical decisions, and coordinate with the player. A leading battle-royale title already features an AI teammate that talks, shares loot, and drives, while a major mobile action game runs a comparable on-device AI for millions of weekly players.
- Adaptive adversaries
This technology creates enemies that break the mould of predictable, pattern-based encounters. A recently launched MMORPG ships bosses that learn from players' tactics across encounters, adapting their strategies to provide an escalating challenge with each battle.
The “sentient” game world and intelligent backend systems
The use of agentic AI is also being applied to the operation of the game itself, making it more responsive and intelligent.
- Dynamic difficulty adjustment
The game itself acts as an agent. It perceives a player's performance in real time and subtly adjusts the level of challenge to keep the player in a state of "flow", which has been shown to increase engagement and retention.
- A world that adapts to the player
Beyond difficulty, an agentic system can perceive a player's individual playstyle and dynamically alter the game world to match it — more hidden paths for explorers, more spontaneous encounters for combat-oriented players — making the world itself feel as if it is evolving in response to the player's actions.
- Intelligent quality assurance
During development, AI agents are deployed to test games. Unlike rigid scripts, they explore game worlds with human-like curiosity, actively trying to break systems and identify bugs, freeing human QA teams to focus on subjective feedback.
- Player creation tools
Agentic capabilities are increasingly reaching players directly. In 2026, a leading UGC platform equipped its entire community with agentic tools, featuring a planning mode that analyses game code and turns natural-language prompts into editable action plans, an early sign that agentic creation will not remain a professional privilege (Section 5 on moderation consequences).
Contacts
© 2026 Bird & Bird