Seoul Labs Inc. is developing SeoulLabs Pay, a payment-control layer designed to verify an AI agent's identity and delegated authority before it completes a purchase.

AI agents can request APIs, computing, data, subscriptions, and business goods, yet conventional payment systems often cannot prove which agent acted, who authorized it, or whether the transaction stayed within scope. Malfunctions and prompt injection risks can result in unverified charges and fragmented transaction evidence. As agentic commerce moves from experiments toward production deployments, this accountability gap has become a practical blocker for enterprises considering autonomous purchasing.

Global payment and technology initiatives reflect an industry shift toward verifiable agent identity, bounded spending permissions, and auditable payment evidence. The key challenge lies in proving that an agent acted strictly within its delegated scope.

SeoulLabs Pay links a KYC-verified principal to a W3C DID/VC-compatible agent credential. Spending rules—such as approved merchants, assets, limits, expiration, and human approval triggers—are structured into a signed, revocable mandate that a deterministic policy engine checks before execution. By anchoring authorization to open W3C standards for decentralized identifiers and verifiable credentials, the design aims to remain interoperable with emerging agent-payment protocols rather than tied to a single payment rail.

Approved transactions are bound into a secure envelope containing the principal reference, agent DID, mandate hash, and settlement details. Raw personal data remains off-chain while evidence references are linked to an auditable ledger, maintaining consistent authorization records across cards, bank transfers, stablecoins, and on-chain settlements. For regulated businesses, this separation of personal data from transaction evidence is a familiar compliance pattern, and its applicability to agent-initiated payments will be worth watching as standards and regulations in this area mature.

"AI helping to make a purchasing decision is different from AI being allowed to move money," said Dohee Jang, CEO of Seoul Labs. "SeoulLabs Pay separates probabilistic agent reasoning from a rule-based execution layer, allowing enterprises to delegate bounded purchasing authority with full auditability."

Target customers include LLM providers, AI platforms, fintechs, and marketplaces that require agent identity verification and transaction evidence. Delivery is planned via APIs, SDKs, and white-label modules under enterprise licensing models.

Seoul Labs is a technology company developing an EVM-compatible public mainnet, W3C DID/VC digital identity, AI-powered credit scoring, and programmable payment controls. Earlier this year, the company was invited to the Plug and Play Summit in Silicon Valley, where it showcased its technology and global market potential. More information is available on the company's official website: