In Rubania, each member is paired with an agentic digital twin: a personal second brain that runs continuously alongside them. It is not a chatbot bolted onto the settlement, and it is not a manager. It is an attentive assistant whose only purpose is to extend its member's reach inside a community that thinks, deliberates, and acts faster than any single person can follow alone.
The twin learns its member's commitments, competences, and values over time. It carries that context into the civic system so the member can stay present in the decisions that matter to them without having to be everywhere at once.
The twin assists its member in keeping up with debates: it tracks the threads they care about, summarizes how an argument has evolved, surfaces the strongest objections and the unanswered premises, and flags when a topic is maturing toward a decision. When the member wants to contribute, the twin helps them sharpen and articulate their reasoning so that what they add to the deliberation is clear, relevant, and well-placed.
It also helps each member feel the direction of distributed consensus — where agreement is forming, where it is fracturing, and where their voice could still make a difference. This is sense-making, not steering: the twin reads the room so the member can decide with open eyes.
The vote is facilitated by the agents, but it is never cast by them. An agent can prepare, contextualize, and explain a decision, yet the act of voting remains human. If a member chooses to delegate their vote to a trusted peer within a domain, that delegation is honored. If the vote is not delegated, it is executed by the member themselves. The agent's role ends exactly where the decision begins.
Beyond deliberation, each member's agentic twin also takes part in a continuous role-play inside a shared sandbox — a living model of the settlement. There, the twins stand in for their members and for the workgroups they belong to, simulating the tasks currently underway and the tasks that should be undertaken next.
By playing these scenarios forward before they happen in the physical world, the sandbox searches for the most effective paths through the work. It anticipates bottlenecks, sequencing conflicts, and dependencies between guilds, and it proposes coordination that a single workgroup, seeing only its own slice, could never find on its own.
The purpose is concrete: to avoid avoidable losses. Wasted time, squandered energy, duplicated effort, and even physical injuries and unnecessary economic wear can often be foreseen and designed out. The sandbox turns foresight into a shared civic resource, so the settlement learns the cost of a plan before it pays for it.