Agentic AI and the Workforce of Tomorrow: Marcello Mari Introduces AI Dojo at HODL 2025

At HODL 2025 in Dubai—a global gathering for visionaries in blockchain, artificial intelligence, and decentralized finance—Marcello Mari, Founder and CEO of Singularity DAO, delivered a visionary forecast that extends far beyond crypto markets. As a trailblazer in the convergence of AI and blockchain, Mari revealed his latest endeavor: AI Dojo, a platform built to evaluate and certify the emerging AI agent workforce.

With a background deeply rooted in decentralized technology, including his foundational role in SingularityNet and the Artificial Super Intelligence Alliance, Mari is no stranger to frontier innovation. But at HODL 2025, he pushed the conversation forward, arguing that the future belongs to agentic AI—and that we urgently need ways to trust and verify these digital workers.


The Rise of Agentic AI: A New Category of Non-Human Labor

Mari describes the future of AI not in monolithic models, but in swarms of specialized agents—autonomous AI entities trained to perform repetitive, cognitive, and operational tasks across industries. These agents, he argues, are poised to become the non-human workforce that powers everything from HR departments and finance to email management and customer support.

“This isn’t hypothetical,” Mari stressed. “It’s already happening before our eyes. Enterprises will soon rely on agentic AI the same way they rely on human staff today.”

However, with this shift comes a critical challenge: how do we measure, evaluate, and trust these agents? Are they secure, effective, and legitimate—or just untested chatbots with potential for error?


AI Dojo: Building the Standard for Trust in Agentic AI

To address this gap, Mari founded AI Dojo, a platform dedicated to:

  • Testing agent performance
  • Evaluating safety and error rates
  • Verifying legitimacy and usefulness
  • Filtering out low-quality or fraudulent agents

Much like cybersecurity firms vet code and platforms, AI Dojo aims to be the trust layer between real-world users and AI-driven agents. “We’re trying to fix the trust issue between AI and the real world,” Mari explained. “We can’t have agents sending money to the wrong accounts or sharing sensitive data due to poor testing.”


Advice for Enterprises: Get Smart, Get Safe, Get Started

When asked what advice he would give companies looking to enter the AI space, Mari provided pragmatic guidance:

  1. Understand You’re Already Using AI
    Many businesses already leverage AI—often without realizing it. From email filters to video cameras, AI is quietly embedded in modern workflows.
  2. Ensure Safety and Compliance
    Mari warned about the risks of mishandled data, especially when employees unknowingly feed sensitive information into public LLMs like ChatGPT.
  3. Start with Integrations
    For small and medium enterprises, integrating LLMs for tasks like content creation, marketing, or financial analysis is a great entry point.
  4. Prepare for Agents
    The next evolution—agentic AI—is approaching fast. These agents will soon require no coding and will be deployed via simple prompt-based systems.
  5. Vet Before You Trust
    Mari emphasized the importance of trust and verification, citing examples of agents making critical transactional errors due to lack of proper evaluation.

His ultimate recommendation: hire a consultant if budget allows. If not, invest in research and vetted tools—and proceed with cautious optimism.


Why Dubai? A Nexus of Tech Innovation and Global Convergence

Mari also highlighted the significance of attending events like HODL 2025 in Dubai, which he calls one of the emerging tech and crypto capitals of the world.

“Real innovation happens when people are in the same room,” he noted. “Events like this foster genuine collaboration, grassroots movements, and conversations that lead to world-changing projects.”


Marcello Mari’s vision is clear and urgent: as agentic AI becomes the new backbone of business and society, we must build transparent systems of trust, evaluation, and accountability. Through AI Dojo, he’s laying the foundation for a safer, smarter future—one where humans and machines collaborate, not collide. And at HODL 2025, that future felt closer than ever.