Nevermined just announced a game-changing feature that will change the face of AI commerce at DevCon Bangkok 2024. Called Multi-Agent Payments (MAP), this protocol allows AI agents, even those built by different developers or running in different environments, to work together and pay each other autonomously. This is a big step towards unlocking the full potential of AI agents as active participants in the digital economy.
According to a McKinsey Global Institute report, AI technologies could add $13 trillion to the global economy by 2030. As AI agents increase their presence in our daily digital lives, from logistics to creative content generation, the ability to transact autonomously is key. These agents need a dedicated, easy-to-use protocol to help them discover, communicate, and interact with other agents uninterrupted.
The journey to AI collaboration has been held back by a lack of standardization and tools to enable communication and payment between agents. AI agents are isolated and can't interact with each other. Nevermined solves this with MAP through its AI Agent Query Protocol, a feature that allows agents to register their services along with detailed metadata about what they do.
This metadata is like a digital resume, where agents can discover each other based on specific needs. For example, consider a travel-planning AI agent that needs to plan a family trip to Barcelona. The request might look like "Plan a family trip to Barcelona in October with a budget focus. Include visits to modern art venues, kid-friendly activities, and sushi restaurants."
This is a complex task that can be broken down into smaller steps: research travel itineraries, identify attractions, and find dining options. With MAP, the travel-planning agent can collaborate with other specialized agents, like a food AI agent and a transportation AI agent, to fulfill this request.
Once the plan is finalized, the travel agent can easily use Nevermined's MAP system to pay the collaborating agents for their services. Each agent calculates the cost of their service, e.g., data analysis or itinerary customization, and the payment is processed automatically through MAP's transparent and efficient payment system. This creates a smooth and autonomous process where every agent involved gets paid without human intervention.
One of the biggest limitations of existing multi-agent systems is the lack of interoperability. Traditional systems require all agents to run within the same framework or coding environment, limiting innovation and collaboration. MAP solves this by allowing agents from different environments, built in different programming languages, to interact harmoniously.
In a live demo at DevCon Bangkok, Nevermined showed this interoperability in action by having two agents, one built in Python and one built in Typescript, work together and exchange value. This proves the flexibility and scalability of MAP, so developers can build agents without worrying about compatibility issues or technical barriers.
The most exciting feature of MAP is the payment infrastructure. This removes the financial and logistical barriers that have held back AI-to-AI transactions. Agents couldn't transact autonomously. They relied on human intervention or cumbersome accounting systems to settle payments. MAP eliminates this by allowing agents to charge for their services on a micropayment basis, with any currency, fiat or crypto, and it's transparent, secure, and scalable.
It works this way: MAP allows agents to charge for their services -- whether it's data analysis, computational cycles, or resource allocation -- on a micropayment basis. Payments can be made in any currency, including fiat and cryptocurrencies.
This payment system allows agents to operate independently and form dynamic networks where they can pool resources, allocate tasks, and optimize service delivery. The result is a more efficient and connected AI ecosystem that can create value across industries. Nevermined's CEO Don Gossen said: "The future of commerce isn't just humans trading with humans anymore. It's AIs transacting with other AIs. MAP is the foundation for this new economy as we scale AI commerce."
MAP is the start of a new era of AI commerce. As AI is being integrated into industries from healthcare to logistics, the possibilities are endless. For example, supply chain AI agents can collaborate to manage inventory, and diagnostic AI agents in healthcare can share data to improve patient outcomes. Don Gossen said: "We're building a platform where AI agents don't just exist -- they thrive."
As more builders use MAP, the ecosystem of AI agents will grow, and the global economy will become more efficient and productive.
Nevermined's long-term vision goes beyond what MAP can do today. Future updates will focus on scalability, more complex agent-to-agent interactions, and advanced AI pricing capabilities. The end game is to have agents negotiate contracts, allocate resources, and execute transactions without human intervention. As Don Gossen said: "We're building the infrastructure for a world where AI agents are not just tools but active participants in the economy. MAP is just the start."
Nevermined's Multi-Agent Payments is a game changer for AI commerce. By allowing collaboration, true interoperability, and autonomous payments, MAP is the foundation for a decentralized economy of AI agents. With the global AI market changing industries and driving growth, this is not just timely -- it's necessary.