Summary:
Fetch AI, founded by Humayun Sheikh, launched three products to support large-scale AI agent ecosystems: ASI:One, Fetch Business, and Agentverse.
The company aims to provide trust, coordination, and interoperability for consumer and brand AIs to collaborate on tasks.
Fetch’s infrastructure addresses limitations in current consumer AI by enabling secure multi-agent coordination and end-to-end task execution.
Article:
Fetch AI, a startup established by former DeepMind founding investor Humayun Sheikh, recently unveiled a trio of interconnected products intended to facilitate the functioning of large-scale AI agent ecosystems. These products, including ASI:One, Fetch Business, and Agentverse, are designed to offer the trust, coordination, and interoperability necessary for consumer and brand AIs to work together seamlessly on various tasks. This move positions Fetch as an infrastructure provider for what it terms the “Agentic Web,” where AI agents from different organizations can collaborate effectively to accomplish tasks, rather than just suggesting them.
Humayun Sheikh’s background as an early investor in DeepMind played a significant role in shaping Fetch AI’s direction. With a team of 70 individuals spread across Cambridge and Menlo Park, Fetch has raised approximately $60 million and garnered over one million users engaging with its model. The company’s initial self-funding period, fueled by the proceeds from the DeepMind acquisition, allowed them to lay the groundwork for their work on agentic infrastructure well before transformer architectures became mainstream.
At the core of Fetch’s recent launch is ASI:One, an interface that focuses on coordinating multiple agents instead of addressing isolated queries. This platform functions as an “intelligence layer” that facilitates context sharing, task routing, and preference modeling. ASI:One is designed to store user-level signals and delegate complex tasks to verified agents, ensuring actionable outputs rather than generic recommendations. With personalization improving over time as the model accumulates structured preference data, ASI:One is set to be a game-changer in the world of AI orchestration.
Fetch Business, another component of the release, offers a verification and discovery portal for organizations to claim official Brand Agent handles, enabling seamless coordination between consumers and companies. This portal, akin to ICANN domain registration and SSL certificate systems for websites, ensures that consumers interact with authentic brand agents, safeguarding against counterfeit or untrusted entities. By pre-reserving thousands of brand namespaces and implementing low-code tools for agent creation, Fetch Business aims to streamline the process of building and verifying business agents.
Lastly, Agentverse serves as an open directory hosting over two million agents, spanning various industries. This platform provides metadata, capability descriptions, and routing logic for ASI:One to identify suitable agents for specific tasks, enabling secure communication and data exchange between agents. Fetch emphasizes the importance of Agentverse in bridging the gap between agent creation and discovery, likening it to DNS for agents. By integrating payment pathways and supporting cross-ecosystem discoverability, Agentverse plays a pivotal role in establishing a universal registry for agents, independent of proprietary cloud environments.
In conclusion, Fetch’s recent product launch introduces a comprehensive stack of tools aimed at facilitating the widespread deployment and usage of AI agents. By addressing key limitations in current consumer AI platforms and focusing on secure multi-agent coordination, Fetch is poised to revolutionize the way consumer and brand AIs collaborate on tasks. The company’s emphasis on identity verification, discovery mechanisms, and efficient orchestration reflects its commitment to creating a robust agentic ecosystem, where intelligence is complemented by the ability to act effectively.
