Apps & Software

Agentcard Raises $8M to Build Payment Infrastructure for AI Agents

By mohitdeep1122 • • 6 min read • 📷 3 High-Res Visuals
Share Dispatch
𝕏 Post in Share Y Hacker News
⚡ TL;DR — Executive Summary

AI agents can now research products, compare prices and prepare shopping carts, but completing a real transaction is still difficult. Agentcard wants to become the financial infrastructure that allows AI agents to spend money securely, with users retaining control over their cards, budgets and purchase approvals.

Agentcard raises $8 million in seed funding to build payment infrastructure that allows AI agents to make secure online purchases on behalf of users.
Agentcard raises $8 million in seed funding to build payment infrastructure that allows AI agents to make secure online purchases on behalf of users. 🔍 Click to zoom

AI agents are becoming increasingly capable of researching products, comparing prices, booking services and completing complicated online tasks. However, one important obstacle remains: making payments. An agent can find the right product and prepare a shopping cart, but completing a transaction requires access to payment credentials, financial authorisation and security controls. San Francisco-based Agentcard wants to solve this missing infrastructure problem. The startup has raised $8 million in seed funding from Peak XV, Fin Capital and Y Combinator, with additional backing from technology leaders including Paul Graham, Adam D’Angelo, Tobi Lütke, Guillermo Rauch and Dharmesh Shah. Founded by Karen Serfaty and Felipe Abello, the company is building payment technology that allows AI agents to make purchases using their users’ existing cards or dedicated virtual cards without exposing sensitive payment information directly to the AI system.

Agentcard Vault enables AI agents to use securely stored payment credentials with user-controlled approvals, spending limits and virtual payment capabilities.
Agentcard Vault enables AI agents to use securely stored payment credentials with user-controlled approvals, spending limits and virtual payment capabilities.🔍 Click image to enlarge

The payment problem could limit how useful AI agents become

Most AI shopping assistants can already help users discover products and decide what to purchase, but the final checkout process still requires traditional payment infrastructure. Giving an AI agent direct access to a credit card creates obvious risks: the agent could misunderstand instructions, purchase the wrong item or spend more than the user intended. Conventional payment systems were designed primarily for transactions initiated by people, not software agents independently interacting with online stores. Agentcard is building an intermediate layer that connects AI applications to existing payment networks while introducing controls around purchasing authority. Its goal is to let developers add payment capabilities to their AI agents without building card storage, checkout authentication and spending management infrastructure themselves.

Vault allows AI agents to use existing cards without seeing the card number

The company’s central product is Vault, which allows users to securely connect their existing payment cards to an AI-powered application. Once a card is saved, Agentcard’s system handles the sensitive payment information while the AI agent initiates purchases without directly accessing the actual card number. According to the company’s product documentation, card information is encrypted on the user’s device before being stored. When an agent attempts a purchase, the user can approve it through a passkey, Face ID or a master password. Customers can also configure automatic approval rules that allow smaller purchases within predefined merchant, amount and budget limits. This creates a practical balance between convenience and financial control: users do not need to manually approve every transaction, but larger or out-of-policy purchases can still require their permission.

Virtual cards give businesses another way to control agent spending

Agentcard is also developing an Issuing product that allows businesses to create dedicated one-time or reusable virtual cards for AI agents. Instead of connecting an agent to a company’s main corporate card, organisations can allocate a specific budget to an individual agent, task or purchasing requirement. A virtual card might carry a fixed spending amount for ordering equipment, purchasing software subscriptions or paying for online services. The platform supports both company-funded cards and individually funded user wallets, with identity verification and know-your-customer checks included in the issuing process. This is particularly relevant for enterprise applications where multiple AI agents may operate simultaneously across procurement, operations and administrative workflows. The objective is to make agent spending easier to separate, monitor and control rather than allowing every automated application to draw from the same unrestricted financial account.

Agentcard's Opencommerce platform connects AI agents with thousands of online stores to discover products, complete purchases and track orders.
Agentcard's Opencommerce platform connects AI agents with thousands of online stores to discover products, complete purchases and track orders.🔍 Click image to enlarge

Opencommerce moves Agentcard beyond payments into AI-powered shopping

Another important component is opencommerce, which gives AI agents a unified purchasing tool for searching products, buying them and tracking orders across more than 5,000 online stores, according to Agentcard. The company lists merchants and services including Amazon, Walmart, Target, Best Buy, Home Depot, Lowe’s, Macy’s, Wayfair and DoorDash. This means Agentcard is not limiting itself to providing a secure payment connection after a product has already been selected. It is also building infrastructure that helps agents interact with retail environments and complete the broader shopping journey. A user could ask an AI assistant to find a product within a budget, compare available options, prepare the order and request payment authorisation. Bringing product discovery, checkout and order tracking into one workflow could make Agentcard more valuable to developers building autonomous shopping applications.

Voice approvals introduce another interface for agentic payments

Agentcard’s fourth product, Voice, explores how customers can approve purchases by speaking rather than navigating a screen. The company describes this as a way to let users authorise their agents’ purchases through voice interactions, although the feature remains in early access. This could become relevant as AI assistants move into messaging applications, connected devices and hands-free environments. Together, Vault, Issuing, Voice and opencommerce address different parts of the same problem: storing payment credentials, creating controlled spending instruments, obtaining approval and completing online purchases. Agentcard provides developer documentation and integrations for agent browsers, e-commerce APIs and messaging platforms, allowing businesses to connect its payment capabilities to existing AI applications instead of building an entirely new shopping product.

Investors are backing the financial infrastructure behind agentic commerce

The $8 million seed round brings together fintech-focused investors and influential technology operators. Peak XV, Fin Capital and Y Combinator participated in the institutional financing, while the company’s wider angel backing includes Paul Graham, OpenAI board member Adam D’Angelo, Shopify CEO Tobi Lütke, Vercel CEO Guillermo Rauch, WorkOS CEO Michael Grinich, HubSpot CTO Dharmesh Shah, Rappi CEO Simón Borrero and Increase founder Darragh Buckley. CEO Karen Serfaty previously led Atlas, an expense-management company acquired by Remote, while CTO Felipe Abello brings earlier software and product-building experience. Their backgrounds are relevant to a business that combines consumer payments, developer infrastructure and financial controls. The financing will support the company’s effort to establish Agentcard as a payment layer for the growing ecosystem of AI agents, although its valuation and transaction volumes have not been publicly disclosed.

The real test is whether consumers and businesses will trust AI with their money

Agentcard is entering a promising market, but giving AI systems access to real financial transactions introduces challenges that cannot be solved by payment technology alone. The company must demonstrate reliable purchase authorisation, strong fraud controls, accurate merchant identification and clear accountability when an agent makes a mistake. Automatic approvals also require carefully configured spending limits so that convenience does not become uncontrolled financial exposure. Developers must integrate the technology into useful applications, while consumers need a reason to trust those applications with their payment information. Competition from payment processors, card networks and other agentic commerce infrastructure providers will also increase as autonomous purchasing becomes more common. If Agentcard can make payments secure, predictable and easy to integrate, it could become an important part of the financial infrastructure supporting AI agents. Its $8 million seed round gives the company capital to pursue that opportunity; sustained transaction activity and trusted commercial integrations will determine whether the platform becomes an essential layer of agentic commerce.

Sources and further reading

📷 Complete Article Visual Gallery (3 Pictures) Click any photo for full-screen view
Agentcard raises $8 million in seed funding to build payment infrastructure that allows AI agents to make secure online purchases on behalf of users.
Agentcard raises $8 million in seed funding to build payment infrastructure that allows AI agents to make secure online purchases on behalf of users.
Agentcard Vault enables AI agents to use securely stored payment credentials with user-controlled approvals, spending limits and virtual payment capabilities.
Agentcard Vault enables AI agents to use securely stored payment credentials with user-controlled approvals, spending limits and virtual payment capabilities.
Agentcard's Opencommerce platform connects AI agents with thousands of online stores to discover products, complete purchases and track orders.
Agentcard's Opencommerce platform connects AI agents with thousands of online stores to discover products, complete purchases and track orders.
mohitdeep1122 • TechFront360 Editorial Desk

Published by TechFront360.com following our strict fact-checking, benchmark verification, and independent technology journalism standards.

More in Apps & Software