Fetch AI, the Cambridge and Menlo Park-based startup led by Humayun Sheikh, a founding investor of DeepMind, announced the launch of three interconnected products aiming to revolutionize the AI agent ecosystem. The offerings include ASI:One, a personal AI orchestration platform; Fetch Business, a brand agent verification and discovery portal; and Agentverse, an open directory hosting over two million autonomous agents.
Together, these tools establish a foundational infrastructure for what Fetch terms the “Agentic Web,” a new layer where consumer and brand AI agents collaborate actively to complete complex workflows, rather than merely providing isolated recommendations.
Addressing Key Limitations in Consumer AI
Current consumer AI systems typically excel at generating recommendations but fall short when tasked with executing multi-step processes that require coordination across different organizations. Fetch AI’s solution centers on enabling secure interoperability between agents from various entities, utilizing verified digital identities and shared contextual information to manage end-to-end task completion.
“We’re building the same foundational platform for AI agents that Google built for websites,” said Humayun Sheikh, Founder and CEO of Fetch AI, highlighting the platform’s ambition to transform AI from passive information retrieval into active task orchestration.
Founding Vision Rooted in DeepMind Experience
Humayun Sheikh founded Fetch AI in 2017, leveraging his unique insights as one of DeepMind’s earliest investors and supporters. He reflected on that period when advanced machine learning research was predominantly confined to large tech labs, noting that his early conviction in agentic systems shaped Fetch’s trajectory.
“Even back in 2013, I recognized that agentic systems—autonomous software agents capable of coordinated action—would be the future,” Sheikh explained. This led to Fetch’s focus on building secure, verifiable infrastructure for multi-agent coordination, emphasizing identity verification and data privacy.
Since inception, Fetch has grown to a 70-person team and raised approximately $60 million. Its platform has engaged over one million users, providing critical data to refine the newly launched products. Sheikh also noted that initial funding came from proceeds linked to DeepMind’s acquisition by Google, enabling early development prior to the mainstream adoption of transformer architectures.
ASI:One—A Specialized Orchestration Layer for Autonomous Agents
At the core of Fetch’s new stack is ASI:One, a language-model-based interface designed to orchestrate multiple autonomous agents rather than handle isolated queries. This “intelligence layer” manages context sharing, task routing, and user preference modeling.
ASI:One stores granular user preferences—such as airline choices, dietary restrictions, budget limits, loyalty programs, and calendar availability—within user-controlled knowledge graphs. When users request complex tasks like trip planning, ASI:One dynamically delegates subtasks to verified agents representing flights, hotels, and restaurants, retrieving actionable outputs such as booking options rather than mere suggestions.
Unlike conventional large language model (LLM) applications relying on API calls or retrieval-augmented generation, ASI:One is architected to coordinate autonomous agents capable of completing transactions across organizational boundaries. Sheikh emphasized the platform’s modular design, combining specialized expert models with agentic systems to enhance execution fidelity.
ASI:One is available in Beta, with a full release expected in early 2026. The platform also offers ASI:One Mobile for iOS and Android, enabling real-time interaction with agents and access to the Agentverse directory while on the move.
Fetch Business—Safeguarding Brand Identity in the Agent Economy
To foster trust and coordination between consumers and companies, Fetch introduced Fetch Business, a portal where organizations can verify their identity and claim official Brand Agent handles (e.g., @Nike or @Hilton), regardless of the underlying agent technology.
This system functions similarly to domain name registration and SSL certification on the web, providing a cryptographically verifiable authenticity badge to protect users from counterfeit or untrusted agents—a significant obstacle to widespread AI agent adoption.
Fetch Business offers low-code tools enabling small businesses to create personalized agents rapidly, integrate real-time APIs such as inventory or CRM systems, and manage social media permissions.
Sheikh explained, “With Fetch, a brand can create an agent in under a minute, personalize it fully, and gain discoverability within the Agentverse ecosystem. Verification persists across platforms, creating a portable identity layer for brand agents.” Verification leverages established web trust mechanisms by requiring domain owners to add code snippets for cryptographic challenges, effectively reusing decades of web security infrastructure.
Agentverse—An Open, Cloud-Agnostic Directory for AI Agents
Completing the launch is Agentverse, an open directory and cloud platform hosting more than two million registered agents spanning sectors such as travel, retail, entertainment, and enterprise services.
Agentverse supplies metadata, capability profiles, and routing logic that ASI:One uses to identify and engage appropriate agents for specific user requests. The platform supports secure inter-agent communication and data exchange, and is framework-agnostic, welcoming agents built on any technology stack.
Sheikh highlighted a critical industry challenge: “Ninety percent of AI agents never get used due to lack of a discovery mechanism. Agentverse solves this by acting like DNS for agents, providing universal discoverability and trust.” He positioned Fetch as building the “Google of agents,” addressing the essential needs of discovery, verification, and interaction within the emerging agent economy.
Agentverse’s cloud-agnostic nature differentiates it from ecosystems tied to proprietary clouds, allowing near-instantaneous agent registration and querying by large language models. The platform also integrates payment capabilities through partners like Visa and Skyfire, supporting micropayments and stablecoins, with user-configurable spending controls.
Implications for AI Industry and Future Developments
Fetch’s integrated platform launch arrives amid a broader shift in consumer AI from static chatbots toward autonomous agents capable of executing complex, multi-step workflows. Current limitations in siloed architectures, weak verification, and poor interoperability have hindered agent adoption.
By providing a unified coordination layer, robust identity verification, and a discovery directory, Fetch aims to overcome these bottlenecks, enabling scalable, secure agent ecosystems. The focus on verified brand identities mirrors early web trust frameworks, addressing risks of counterfeit interactions.
Furthermore, ASI:One’s user preference centralization and multi-agent orchestration contrast with generalized LLM approaches, offering deeper personalization and transactional capabilities. The incorporation of micropayment systems underscores Fetch’s vision for fully autonomous agents performing end-to-end commercial tasks.
Ultimately, Fetch AI’s release of ASI:One, Fetch Business, and Agentverse establishes a comprehensive infrastructure stack designed to underpin the agentic web. Sheikh’s long-held thesis, influenced by lessons from DeepMind’s early days, is that meaningful AI intelligence must be coupled with the ability to act, coordinate, and transact reliably across ecosystems.
Related Articles
Traders Fair Manila 2026 to Unite Global Trading Leaders and Innovators Amid Rapid Financial Market Evolution
Insurance Leader AIG Accelerates Efficiency with Agentic AI and Orchestration Layer
Megacampus Summit Dubai 2026 to Unite Global AI and Tech Leaders in Landmark Event
Cloudflare Accuses Perplexity AI of Ignoring Website Scraping Restrictions
