The financial world is on the cusp of a radical shift, and it’s not about new currencies or faster human transactions. It’s about machines paying machines. Natural, a new startup, has just announced a $9.8 million seed round, backed by Abstract, Human Capital, Forerunner, and a roster of prominent founders, to build the foundational payments infrastructure for AI agents. This isn't just an incremental upgrade; it's a complete reimagining of how money moves, designed from the ground up for an agentic future where AI systems handle a significant portion of global transaction volume.
For decades, payment systems have been meticulously crafted around human interaction. Account numbers, card portals, and dispute resolution processes all assume a person is at the helm. But as AI agents become increasingly sophisticated, capable of negotiating logistics, managing properties, or automating complex back-office workflows, the current financial rails simply break down. When an agent can hire a contractor, book a flight, or procure supplies in seconds, the payment process can’t lag by days.
I’m excited to finally announce Natural (@naturalpay), our $9.8m seed round, and commitment to building the best payments ecosystem for agents. pic.twitter.com/WV15njgZph
— Kahlil Lalji (@bykahlil) October 23, 2025
Kahlil Lalji, CEO of Natural, highlights this disconnect, stating that while agents can manage procurement or sales, "the second money needs to move, the workflow breaks. A human has to step in to execute the payment." Natural aims to eliminate this bottleneck by creating a natural language layer that allows agents to transact directly with businesses and other agents, initially leveraging existing rails like ACH, but with an eye toward consolidating fragmented payment workflows into machine-accessible protocols.
The implications are vast. Industries currently missing billions in value due to payment friction — logistics agents unable to settle freight charges automatically, property management agents unable to pay contractors in real-time — could see end-to-end autonomous workflows. This isn't just about efficiency; it's about unlocking entirely new paradigms of automated commerce and labor.
The Unseen Challenges of Autonomous Transactions
The transition to an agent-led economy introduces a host of complex problems that current payment systems are ill-equipped to handle. Traditional financial rails, designed for human-speed transactions, are too slow and costly for the rapid, high-frequency interactions of AI agents. Imagine a master agent orchestrating a dozen specialized sub-agents to build a website; each sub-agent might need to be compensated in near real-time for tasks completed in minutes. The existing infrastructure, with its 1-3 day clearing times for ACH or multi-second card authorization loops, simply can't keep up.
Dispute arbitration also becomes a minefield. When an agent autonomously books a non-refundable flight with a 12-hour layover, who is liable if the human principal is dissatisfied? The current system, which largely focuses on whether an authorized individual made a payment, doesn't account for "satisfaction" with an agent's decision. Similarly, managing identity and trust in an unbounded network of interacting agents presents a novel challenge. How does an agent verify the legitimacy of an unknown counterparty programmatically?
Furthermore, as agents optimize for efficiency and cost, payments will quickly become global. An agent tasked with "writing code" or "completing translations" might find the best vendor anywhere in the world. This necessitates low-cost, low-latency international transaction capabilities, a far cry from today's expensive and slow cross-border payments. Even existing risk and fraud controls, built on predictable human patterns, will need to be re-engineered for the unique behaviors and potential fraud vectors of autonomous agents.
Natural's approach seeks to address these issues head-on. They envision "controllable wallets" that allow businesses to set granular spending limits and liquidity for agents, along with robust rules and controls that can be applied at the organization, wallet, or transaction level. Real-time authorization tools, mirroring or exceeding the speed of credit card networks, are crucial for synchronous agent-to-agent transactions. Observability and tracing tools will provide the necessary transparency for auditing and improving agent payment flows. The goal is to support transactions across any preferred rail, from ACH to stablecoins, ensuring flexibility for diverse counterparty preferences.
While the market for agent interoperability is still maturing, and existing players like Stripe are beginning to offer agent toolkits, Natural believes the core problem of building generalizable payments infrastructure for agents remains largely unsolved.
