Wasted Potential Studios

We build intelligent systems by building the worlds they must understand.

Wasted Potential Studios is an Illinois-based technology studio working with projects worldwide. We developed ISABELLA to support adaptive agents, persistent context, and structured behavior across complex simulations and custom AI products.

Our point of view

AI becomes meaningful when it has to live inside a real system.

Intelligence is more than a response. It is the relationship between context, experience, decision-making, action, and the environment that gives those things consequence.

That perspective guides our work on ISABELLA and on custom client systems. We begin with the problem and its boundaries, choose technology in service of that problem, and build behavior that the surrounding product can observe and test.

A person working with a responsive digital environment
Our work connects human intent, intelligent behavior, and the systems where outcomes take shape.

Our internal proving ground

Medieval Life: Chronicles of Annora

Annora is Wasted Potential Studios' persistent-world project and an internal environment for confronting the practical questions behind adaptive agents.

Why a living world matters

A persistent simulation forces agent behavior to connect with changing world state, individual experience, social context, available resources, and the consequences of prior choices. Those pressures make architecture decisions visible.

What carries beyond games

The same core questions appear in training environments, operational simulations, digital experiences, and enterprise workflows: what changed, what should be remembered, who needs to know, and what action is permitted next?

Annora informs our engineering approach; it is not presented as validation for every external use case. Each client system requires its own scope, evidence, and evaluation.

How we build

Three principles shape the work.

01 · Start with the system

Understand the whole environment

We map users, goals, data, constraints, interfaces, permissions, and failure conditions before deciding what role a model should play.

02 · Make behavior observable

Build for evaluation

Actions, state, and decision boundaries should be visible enough for domain experts and engineers to test what the system is actually doing.

03 · Fit the environment

Choose technology in context

Model, memory, infrastructure, and deployment choices should reflect the project's hardware, privacy, performance, integration, and maintenance realities.

Work with WPS

Bring us a consequential problem.

If your organization needs intelligence that must remember, adapt, integrate, or operate within firm boundaries, we would like to understand the challenge.