Agentic AI is becoming the next big step in the development of AI technology. Multi-agent AI systems are revolutionising complex problem-solving through collaboration, allowing the introduction of a new level of automation in multiple industries, from software development to finance. In early 2025, Amazon Web Services formed an entirely new business unit dedicated to AI agents, while Microsoft announced at its Build conference that it is developing new protocols for agent to agent communication. Soon, Google introduced its Agent Communication Protocol to support agent interoperability. This synchronized industry movement signals that autonomous agents are no longer experimental—they’re becoming core infrastructure. However, rapid growth brings a new set of challenges with it: to operate reliably and efficiently, independent AI agents need a way to coordinate their actions, avoid contradictions, repetitions, and other mistakes that can become costly at a bigger scale.
Resolving these challenges plays a key role in scaling autonomous intelligence further. Brandon Vrooman, founder of Innobit Inc., has spent the past four years building the infrastructure that answers this question. At Fuel Labs, he developed the high-performance blockchain verification layer that makes trustless coordination practical at enterprise scale. At ChainML, he contributed to the Theoriq protocol—the communication framework that enables AI agents to discover capabilities, coordinate multi-step workflows, and operate reliably in decentralized environments. His work addresses the fundamental challenge facing companies that want to deploy agentic AI as part of their operation: when you have autonomous agents handling customer service, data analysis or marketing optimization, they need robust collaboration protocols to collaborate efficiently. Multiple projects, developed and implemented by Brandon Vrooman, clearly demonstrate: the key to solving the agent coordination problem lies in building infrastructure in the intersection of blockchain and AI, and the companies able to master this convergence are the once which will be able reap the benefits of the oncoming agentic economy.
Why Making Agents Smarter Isn’t Enough
In essence, multi-agent systems suggest the next step in developing AI technology: instead of relying on a single AI to handle complex tasks, an environment with multiple specialized agents is created, with them being able to collaborate and thus solve problems more efficiently. Imagine a research task where one agent searches academic databases, another analyzes patent filings, and a third one monitors market trends, with the final one synthesizing the findings of other AI agents into a detailed, coherent report.
However, as the system scales, the fundamental challenge also intensifies. To interact, agents need to coordinate their actions, avoid duplicated work, and establish trust in each other’s outputs, performing all this without human intervention. Traditional approaches lack proper architecture and clear communication protocols, which often create critical bottlenecks. For instance, if multiple agents work on the task at the same time, they often have to wait idly for other components of the system to complete tasks, wasting time and creating inefficiencies. Operating in a decentralized system adds another layer of complications, as agents may perform identical work or generate contradictory outputs, undermining their reliability.
“You can have the most sophisticated individual AI agents, but without a robust protocol for how they discover, communicate with, and verify each other’s work, the system quickly becomes unable to operate,” explains Brandon Vrooman. “The real challenge is not making the agents smarter but establishing architecture for them to collaborate efficiently.”
Coordination problems may seem manageable when two or three agents operate in a controlled environment, but complex tasks may require coordinating hundreds of them, and as the scale grows, the necessity of a new approach becomes clear. In a decentralized agent economy, which may emerge in the near future, AI assistants might need to interact with each other, verify credentials, coordinate multi-step tasks, and handle payments — the lack of such an approach will become a barrier preventing multi-agent systems from reaching their full potential.
Brandon Vrooman confronted these challenges directly as Lead Blockchain Engineer at ChainML, where he developed critical components of the Theoriq Protocol — a system designed specifically to enable AI agents to communicate and collaborate in trustless environments.The protocol addresses a problem that becomes critical at scale: when thousands of AI agents need to work together on complex tasks like market analysis, data synthesis, or multi-step customer workflows, they require standardized ways to discover each other’s capabilities, coordinate actions, verify outputs, and maintain consistency, without relying on centralized coordination that creates bottlenecks.
Closely collaborating with the engineering team, Brandon significantly contributed to designing and implementing the communication protocol itself, defining how agents would interact through standardized data flows, APIs, and authentication mechanisms. Then, he built the distributed server infrastructure to support these interactions at scale, developed the database architecture and created the API layers that allow agents to discover and invoke each other’s capabilities.
Currently, the Theoriq platform supports over 100,000 unique users while processing more than 2 million requests per day, each of them potentially involving coordination of multiple AI agents. The infrastructure Brandon helped design was specifically built to handle the unpredictable load patterns of AI agent systems, maintaining performance during massive spikes in activity that occurred during testnet campaigns and product launches. The incentivized testnet itself became a large-scale validation: hundreds of thousands of participants interacted with AI agents, completed feedback quests, and participated in referral programs, demonstrating that the protocol could handle both the technical demands and the market interest in decentralized agent coordination.
“The goal was to establish not just a working system but infrastructure other developers could rely on,” comments Brandon Vrooman.”When hundreds of thousands of testnet participants were using our protocol, that validated the architectural decisions. If the coordination layer is solid, developers can focus on making their agents smarter rather than solving trust and communication problems themselves”
His contributions supported the public release of Theoriq’s agentic base layer in late 2024, positioning the platform as a modular and composable protocol for decentralized AI agents – infrastructure that other developers can build upon to create their own agent-powered applications. By providing open, composable infrastructure, Theoriq enables developers to focus on building specialized agents for specific domains while relying on proven infrastructure for the complex coordination layer, operating in a fashion similar to how APIs and communication protocols enabled mobile and web ecosystems to scale, providing a shared infrastructure that allows specialized innovation on top.
Theory Means Little Without Speed
This example illustrates that the coordination challenges facing multi-AI systems can be resolved through a new type of infrastructure that allows establishing trust, ensuring transparency, and verifying actions without centralized control. This is exactly the set of tasks that can be addressed with blockchain. It enhances the capabilities of AI agents through enabling what researchers now call “agentic blockchains” — systems that coordinate AI agent activity while allowing those agents to act independently in other environments.
Blockchain solves the coordination problem through three key mechanisms. First, smart contracts can automatically verify completion conditions and execute agreed-upon actions without requiring a trusted intermediary. Second, on-chain registries create a shared source of truth where agents can discover each other’s capabilities and verify credentials, again, without the need of centralized directories. Third, cryptographic verification allows agents to prove the authenticity of their outputs, validating data from other agents without revealing sensitive information. Together, these mechanisms create an infrastructure where coordination happens through mathematics and transparent rules rather than through trusting a central authority to enforce fair behavior. In the future, such a decentralized system where the infrastructure itself becomes a source of trust has the potential to become a foundation for the agentic economy.
However, the foundational infrastructure only matters if it can perform at the speed and scale that real-world applications demand.
“There’s a significant gap between blockchain as a concept and blockchain as production infrastructure for AI agents, implementing it requires making blockchain fast for the specific workloads that AI agents generate,” comments Brandon Vrooman. “Agent coordination creates constant state verification and frequent micro-transactions. Building algorithms optimized for these patterns was what transformed blockchain from a theoretical solution into practical infrastructure.”
This is exactly the practical engineering challenge Brandon Vrooman tackled during his tenure at Fuel Labs, where he served as Senior Software Developer from 2021 to 2024. There, he designed and implemented one of the protocol’s most critical components, namely the Merkle Tree library, a cryptographic component that provides verification for transactions, smart contracts, and other essential data within the network. Brandon’s work enabled the Fuel blockchain to sustain exceptional throughput, processing over 21,000 transactions per second while lowering transaction fees significantly. These metrics matter as they represent the infrastructure capable of handling the demands of millions of AI agents operating autonomously as the ecosystem develops. The library developed by Brandon provides critical support for Fuel’s scalable, secure infrastructure, creating a high-performance foundation for building agent coordination systems that can work at a large scale.
Shaping the Future of Agentic Economy
Brandon’s work exists within a rapidly evolving ecosystem where the blockchain and AI communities are actively converging to solve the coordination challenges that limit autonomous systems, with several major initiatives launched in 2025, like the projects by Amazon, Microsoft and Google mentioned above, focusing specifically on the agent AI collaboration.
Brandon Vrooman’s progression from working as a 3D Programmer at Ubisoft, where he contributed to developing graphics engine technology for AAA titles like Assassin’s Creed Origins and Odyssey, to blockchain protocol development at Fuel Labs and AI agent infrastructure at ChainML exemplifies the cross-domain expertise required to build at this intersection. His cross-domain expertise has earned recognition beyond the blockchain industry, including winning Platform of the Year in the blockchain category at the American Business Expo and serving as a judge for Cases & Faces International Business Award, where he evaluated innovations across multiple sectors. His contributions span both sides of the convergence: cryptographic verification systems enabling 21,000+ transactions per second at Fuel; distributed communication protocols supporting 100,000+ users and 2 million daily requests at ChainML. This combination—understanding how to build high-performance infrastructure and knowing what coordination mechanisms autonomous agents actually need—represents the specialized knowledge that transforms theoretical concepts about multi-agent coordination into production systems handling real-world scale.
For the broader industry, this infrastructure convergence addresses a critical bottleneck in AI agent deployment. Companies launching autonomous agents for customer workflows, financial operations, or data analysis currently face a choice: build centralized coordination that creates bottlenecks and single points of failure, or invest heavily in developing decentralized infrastructure themselves. Platforms like Theoriq and blockchains like Fuel are resolving this dilemma by providing proven coordination mechanisms and verification systems that developers can build upon.
