AI-powered commerce and the ‘Protocols Battle’: Strategic moves for banks and businesses

By Yaacov Martin, CEO of Jifiti.

The agentic commerce wave is moving beyond simple discovery. It is fundamentally rewriting the rules of payment and platform architecture, with leading tech and finance firms already rolling out a new set of agentic AI technologies and tools. These tools are changing how money moves online and mark the beginning of the end of traditional checkouts.

The release of Google’s Universal Commerce Protocol in January 2026, following OpenAI’s.

Agentic Commerce Protocol, from late 2025, confirms that we are in an arms race. These protocols are designed to allow AI agents to explore and shop across a network of merchants, aggregators, and services. They enable bots to pass payment credentials and initiate checkouts while the merchant maintains nominal control.

The scale is already staggering. Salesforce’s 2025 holiday report noted that AI agents influenced $262 billion in U.S. sales alone. While Visa and other giants are already testing hundreds of agent-led transactions, the question remains for the rest of the market. How does this shift affect established banks and the broader lending ecosystem?

Banks, especially those that do not have the in-house resources to manage tech-heavy AI transformations, are not particularly at a disadvantage. As new agentic protocols emerge, the battle for control and ownership of data flows, transaction fees, and scalable infrastructure will lead to competition, resulting in more accessible and efficient technologies. Banks and lending organisations can benefit as they decide the best fit for their agentic commerce integration of portfolio and services.

How financial institutions can integrate to agentic commerce: A rough layout of the terrain ahead

There are several considerations at the board, executive, and departmental levels that are part of the equation of agentic commerce integrations. Fees and risks, and balancing out liability redistribution among banks, FinTechs, and AI platforms are important. These risks need to be detailed and laid out to understand the current offerings that agentic commerce third-party providers have presented to the market.

On the other hand, the tension between multi-protocol coexistence versus consolidation can be navigated through frameworks that outline short-term parallel protocol adaptation and continue with a gradual consolidation of the dominant standard in the long term. Lenders that are challenged with the complexity of agentic integrations can bridge these gaps by partnering with third-party providers. Beyond the portfolio of services that third-party agentic integration providers offer, the risk of overdependence needs to be considered. The biggest risk may not be choosing the wrong protocol but becoming operationally locked into someone else’s ecosystem.

The rise of protocols is also reshaping competitive dynamics. This is where the need to balance speed of innovation with long-term flexibility and independence comes into play. Relying on middleware platforms to simplify agentic commerce adoption may accelerate time to market, but potentially give away control over critical infrastructure layers.

Agentic commerce protocols are capable of breaking down data silos, moving customer data that sits on the AI platforms to the banks’ side. However, this feature is not available by default, but depends heavily on how the protocol is designed and how integrations are negotiated. Banks and FinTech organisations must determine how much transaction data is shared and who controls the interaction flows to remain visible in the customer journeys of agent-driven commerce.

Having considered economic control, technical interoperability, vendor dependency, governance and trust, visibility and customer ownership are next. In traditional commerce, customers engage directly with banks, merchants, or payment providers and gain data on behavior, preferences, spending patterns, and signals. They use all this data to build products, loyalty, and long-term relationships. However, when AI assistants or AI agents handle purchasing decisions, the AI becomes the primary interaction layer, and much of the customer data is at risk of being siloed out from financial institutions.

Final thoughts

Agentic commerce sounds a bit scarier than it actually is. But if you break it down with a trusted technical advisor – one who has expertise in lending as well as an understanding of how to orchestrate your requirements and systems with the external components required  – then you can untangle the yarn ball and create a clear and executable roadmap.