IBM defines autonomous agents as follows:
“Intelligent agents are software entities that carry out some set of operations on behalf of a user or another program with some degree of independence or autonomy, and in so doing, employ some knowledge or representation of the user’s goals or desires.”
Building on that, Fetch.ai’s autonomous economic agents always act with high degrees of autonomy and with the goal of generating economic value for their owner. In this article, we will look at the building blocks of Fetch.ai’s AEAs with a focus on their ability to serve their owners’ business and industrial needs.
What makes Fetch.ai’s Autonomous Economic Agents unique?
Structurally, agents are composed of three core components:
- Runtime — for an agent to deliver value, the Runtime fetches packages from a set of resources. Once those packages are acquired, the Runtime will execute their code as either scheduled tasks and events or messages.
- A set of resources consisting of skills, protocols, connections, and contracts which the agent uses in code form. In this context, the resources act as a registry for code through which modularity and reusability are introduced.
- The final structural building block is the wallet. It is a rather simple data structure containing the private keys of the autonomous agent, allowing it to append digital signatures to transactions distributed on decentralized ledgers.
In simple terms, agents proactively execute a given business logic and react to changes in their environment. They represent modular, composable code abstractions that can reuse existing packages to achieve their goals.
Additionally, the AEA framework developed by Fetch.ai, being a software framework, has significantly influenced the structure of agents. Agents can contribute to the connectivity of previously siloed data when implemented through it. Moreover, data with a short life cycle that would otherwise add no value will now leverage rapid decision-making in a privacy and ownership-preserving manner.
Furthermore, thanks to the minimally-assuming structure of agents, they have the flexibility to be programmed with both simple (if-then-else) and complex (deep-learning algorithm) logic. That means they can also be leveraged as a tool to massively increase the automation of complex processes’.
Therefore, agents truly represent a unique invention, offering not only automation but genuine autonomy to those that create and deploy them.
Applications of AEAs
Application areas for agent-based systems range from the automation of simple tasks such as temperature monitoring to projects involving multi-party coordination and negotiation. Supply chain management’s jobs like sourcing, procurement, physical distribution, and material handling, as well as tracking and tracing can all be meaningfully optimized thanks to agents’ ability to search, discover, and negotiate. What is more, Fetch.ai’s AEAs achieve all this without the need to sacrifice ownership or privacy of data.
Some of the other potential application areas of AEAs are:
- Autonomous risk and treasury management
- Automating and optimizing commerce
- Order fulfillment, synchronization and tracking
- Dynamic price assignment
- Search and discovery by connecting with existing web 2.0 systems
- Machine-to-machine communication
Conclusion
Autonomously mitigating risk, negotiating, or consistently benefiting from asynchronous and profitable opportunities — are just a few examples of how an autonomous economic agent can deliver value. Fetch.ai’s AEAs are the next evolutionary step of the Internet combined with distributed ledger technology and enhanced by machine learning and deep learning. While their potential in various application domains is still being explored, agent-based systems truly represent an immense opportunity for both companies and people who dare to pioneer.
