In this edition’s newsletter, Vincent Chok from First Digital delves into the advent of ‘agentic finance,’ where AI agents are transitioning from providing advice to performing financial transactions. This evolution positions cryptocurrency as the critical infrastructure supporting this machine-driven economy.
During our “Ask an Expert” segment, we sought insights from three leading AI models—Grok, Gemini, and Claude—on topics related to AI payment applications and scalability strategies. Note that these responses are generated by AI systems, reflecting each model’s perspective, and should not be taken as financial or legal advice.
The rapid expansion of AI agents is a key trend from the past year. According to a recent PwC survey involving over 300 companies, 79% have adopted AI agents in some capacity, highlighting their shift from advisory to execution roles.
Initially used for chatbot services and copiloting tasks, these systems now autonomously plan, decide, and act within human-defined parameters, including executing financial transactions. This marks the beginning of ‘agentic finance,’ where AI agents operate within set constraints such as limits and permissions.
Agentic finance comprises three layers. The agentic commerce layer handles discovery and decision-making, like finding optimal hotel deals for travel. The agentic payments layer manages transaction execution upon approval. Lastly, the asset management layer encompasses portfolio management, payment handling, and financial strategy optimization based on real-time data. This conditional delegation retains user control while delegating execution.
While AI agents are theoretically viable in finance, they don’t align with traditional infrastructures. They lack direct access to global banking systems and function continuously. Cryptocurrency addresses this gap by offering programmable money (stablecoins), instant settlement via blockchains, and permissionless fund access through crypto wallets—making it ideal for autonomous systems.
Early implementations are evident in machine-to-machine payments and autonomous commerce, enhancing inter-merchant efficiency and optimizing consumer transactions like travel deals. In crypto-native settings, trading agents manage portfolios and optimize yields. On the enterprise level, AI-driven automation is prevalent in supply chain management and vendor payments, reducing errors and resource use.
AI agents also drive new investment categories and demand for cryptocurrencies. As they can’t operate on existing infrastructure, there’s a growing need for agent-native wallets and stablecoin payment rails. Coinbase has introduced x402, an open protocol catering to agent transactions, crucial for micropayments where traditional systems falter due to high volume and low value.
Despite progress, challenges remain, particularly regarding security against rogue agents and regulatory clarity around authorisation and liability. Building trust through stakeholder-defined regulations is essential for widespread adoption.
Over the next year, key indicators will include growth in agent-driven transactions, development of agent-native wallets and payment protocols, and deeper stablecoin integration with AI systems. Regulatory developments will significantly influence adoption across sectors.
AI agents are no longer hypothetical; they’re active in limited settings. As this trend evolves, cryptocurrency is increasingly seen as the infrastructure for machine-driven economies. While currently an infrastructural theme, its role as a crypto utility driver is expanding with rising adoption rates. Advisors should monitor it closely.
-Vincent Chok, CEO and co-founder, First Digital
This week, we explore AI payment developments with insights from three leading AI models—Grok, Gemini, and Claude. Despite common themes on future growth requirements, distinct perspectives emerged. We hope you find this exploration as engaging and insightful as we do.
Q1: What current AI payment use cases have you identified?
Q2: What is necessary for the scaling of AI payments?