AGENT launches multi-agent AI system with fixed-supply utility token

Moving from answers to actions

Many current artificial intelligence tools provide users with drafts, suggestions, or one-off answers. However, these tools often require the human user to manually bridge the gap between a high-level strategy and the final technical delivery. AGENT is a project based in Zürich that seeks to solve this specific problem by turning goals into completed work.

Instead of relying on a single conversational model, the project organizes seven specialized AI agents to collaborate on an end-to-end workflow. This structure is intended to move ambitious projects from a simple prompt into a sequence of accountable steps. The goal is to provide a coordinated intelligence where each agent has a defined responsibility within a larger team.

Co-founder Leo Bern explained that the project was designed so that "AI should not stop at the answer." The team wants to connect AI-driven tasks to real economic activity in a way that is transparent. By treating AI as a structured team, the platform aims to serve individuals and businesses who need more than just a chatbot for one-off queries.

How the seven-agent system works

The system operates by breaking down a complex goal into a series of specialized roles. Each agent in the seven-part system handles a specific part of the process, such as research, quality control, or distribution. This separation of roles is meant to create a clear sequence of events rather than leaving tasks in an open-ended chat session.

For example, a founder might ask the system to launch an online shop. In this scenario, the agents would plan the project, research the market, create product listings, and set up the storefront. They would also be responsible for reviewing the quality of the work and driving traffic to the new site. This allows the user to oversee a complex project without managing every minor technical detail themselves.

A key safety and control feature of this system is the requirement for human approval. The project notes that users stay in control of every consequential action, including every payment. If the agents need to access third-party APIs, advertising tools, or other paid services, the founder must approve the transaction from their own wallet before it settles.

The AGENT token economy

To support this ecosystem, the project is introducing a utility token called AGENT. This token is designed to settle payments for paid work across the AI economy, such as data, tools, and digital services. The token has a fixed total supply of 1,000,000,000, and the project has stated there will be no additional minting or burning of the supply.

The payment structure is designed to be highly efficient for service providers. When a payment is completed within the ecosystem, 99% of the funds are directed to the service provider. The remaining 1% is allocated as a protocol fee to support the ongoing development of the AGENT platform. This model aims to create a transparent way for AI agents to interact with the broader economy.

The project has outlined a specific timeline for public participation in this economy. A community presale is scheduled to begin on December 1, 2026, which will be followed by a public sale on March 1, 2027. Together, these two rounds represent 200,000,000 AGENT tokens, which is approximately 20% of the total supply.

Future development and goals

As the system matures, the project plans to deepen the coordination between its seven agents. The team intends to expand third-party integrations to make the payment flows more seamless for real product usage. They also plan to bring the system closer to actual product usage by refining how the agents interact with various digital services.

To encourage early adoption, the project intends to run community programs that include referral initiatives and quests. These programs are part of a broader strategy to build a community around the new multi-agent ecosystem. The project has already published its whitepaper and product overview for public review to provide transparency to potential users.

While the current examples are illustrative, the project notes that actual outcomes will depend on specific user goals and market conditions. The ultimate objective remains connecting AI-driven work to real economic activity. By providing a structured way for agents to execute tasks, AGENT hopes to move AI from a novelty into a functional tool for completed work.