AI Agents Become the Next Payments Customer

The payments industry has been focused on ensuring bots cannot be mistaken for customers. Agentic commerce introduces a new challenge as some bots are intended to make purchases. SolvaPay, a FinTech company based in Stockholm, is developing payment infrastructure for artificial intelligence agents. Their work includes machine-readable payments, usage-based billing, and transactions between agents, as well as controls over what an agent can spend.

SolvaPay CEO and Co-Founder, Viggo Stenseth, views payments as a place where AI's growing autonomy collides with systems designed for humans. An agent can gather information, use software, and complete tasks, but money transfers still rely on assumptions about the opposite party in the transaction. SolvaPay has created robust systems to verify human transactions, including 3DS and CVV codes, to prevent fraudulent bots.

Merchants cannot simply remove these safeguards since an authorized agent may still spend money, while also determining whose funds the agent can use and what it can purchase. Businesses have not yet had the time to prepare for this scenario, according to Stenseth. He recalled selling companies on websites in the 1990s, when many doubted the necessity of an online presence.

It took years for eCommerce to become widespread. Smartphones arrived even earlier, before many businesses and payment systems were built around them. Agentic AI has developed more rapidly. Agents can now complete tasks independently, but the billing, payment, and flow processes have not been solved yet. SolvaPay aims to bridge the gap between financial infrastructure and agentic software.

Agent transactions can be complex, as one assignment can generate multiple payments. Several layer deep transactions may require reconciliation processes not typically encountered when payments are treated as single events. Agents can act both as buyers and sellers, necessitating usage-based billing for API calls, datasets, or services, rather than traditional monthly subscriptions.

Connecting agent transactions to existing financial systems requires addressing anti-money laundering requirements, ledgers, licensing, and security measures rather than treating autonomous software as a separate payments environment. For merchants, the implications extend beyond accepting payments initiated by software, as fraud systems must recognize legitimate automated activity, and billing may need to accommodate software purchases.