The Future of AI-Enabled Commerce: A Dive into Agentic Payment Protocols

When Walmart and OpenAI announced their collaboration to integrate ChatGPT, curiosity surged about how rapidly OpenAI could fulfill the vision of AI agents autonomously purchasing items on behalf of users. As the landscape of AI-driven commerce evolves, a critical challenge remains: instilling the confidence necessary for AI agents to securely handle financial transactions.

The Emergence of AI in Shopping

An increasing number of users are turning to chat platforms like ChatGPT instead of traditional web browsers for shopping needs. This trend is evident when a user inquires about the best humidifiers on the market. ChatGPT delivers optimized results, and the user is often led directly to clickable purchase links. However, the current capabilities of AI agents face a significant hurdle—they lack the trust framework required by both consumers and banking institutions to handle financial transactions safely.

For AI agents to progress toward facilitating purchases, there must be a consensus on a universal understanding of trust between AI model providers, banking entities, retailers, and consumers. This complexity in relationships has prompted the emergence of several competing standards, each aiming to establish a secure environment for agentic commerce.

Competing Standards: The Rise of Protocols

Over recent weeks, three major protocols have arisen, each with distinct approaches to enabling secure transactions through AI agents:

1. Agent Pay Protocol (AP2): Launched by Google in partnership with financial giants such as PayPal, Mastercard, and Salesforce, this protocol aims to create a secure payment environment for chat AI.

2. Agentic Commerce Protocol (ACP): Developed by OpenAI and Stripe, the ACP is designed to facilitate transactions with minimal disruption to existing infrastructures. It acts essentially as a courier, delivering information from the AI agent to the retailer.

3. Trusted Agent Protocol (TAP): Introduced by Visa, TAP requires agents to be added to an approved list, receiving a digital key for identity verification, ensuring only authorized agents perform transactions.

While these protocols aim to bolster the trustworthiness of AI agents, their proliferation could lead to complications, particularly with the potential to create isolated "walled gardens." This scenario makes it harder for enterprises to select a chat platform and corresponding payment standard that will effectively interoperate across the space.

More Read

Unlocking the Power of Google Home’s Gemini AI: Tackling Complex Requests with Ease

Samsung Achieves Record Quarterly Profit with Nearly 50-Fold Surge in Chip Revenue

OpenAI Halts MLK Deepfake Videos on Sora Due to Disrespectful Content

Gmail Introduces Slider Feature on Android Tablets with AI Integration

Anthropic Unveils Claude Gov: Advanced AI for Military and Intelligence Applications

Understanding the Differences

The rise of multiple protocols is not entirely new; industries often take years to rally around a single standard. However, with the rapid pace of innovation in enterprise AI, the situation has evolved more quickly than expected. Although tools like the Managed Control Protocol (MCP) have emerged as a sort of baseline for tool identification, the presence of competing payment standards may slow the industry’s progression towards a unified standard.

Protocols like AP2 and TAP focus on cryptographic proofs for authorization, while ACP seeks to minimize infrastructure alteration by serving mainly as a message relay between the agent and the seller. With such varied approaches to establishing security, many organizations might find it challenging to navigate the developing landscape.

The Pitfalls of Walled Gardens

The existence of multiple agentic payment protocols could inadvertently lead companies into walled gardens, severely limiting their ability to transition between different systems or leverage multiple chat platforms. This isolation is troubling for enterprises that want to interact seamlessly with agents across various platforms while also ensuring customer trust.

As highlighted by Louis Amira, CEO of agent commerce startup Circuit, the increasing number of competing protocols creates opportunities for companies to step in at the interoperability layer. Nevertheless, it also introduces confusion, as companies may find themselves tethered to a specific payment standard that doesn’t align with their broader operational needs.

Opportunities for Enterprises

For the time being, the best strategy for enterprises may be to experiment with different protocols and monitor their evolution closely. Wayne Liu, chief growth officer at Perfect Corp., notes that having multiple standards could provide valuable learning experiences, emphasizing the role of open source as a catalyst for collaboration among protocols.

Looking ahead, it’s uncertain whether other major retailers or chat platforms will join the fray with additional competing protocols, but the innovative tension in the space suggests that organizations need to be agile and ready to adapt. As AI-driven shopping grows increasingly complex, transparency, trust, and interoperability will be the cornerstones of successful agentic commerce.

Inspired by: Source