Artificial intelligence “agents” are envisioned to surpass the capabilities of traditional chatbots.

Over recent months, the tech industry has enthusiastically introduced AI personal assistants that possess the intelligence to understand user needs and perform substantial tasks autonomously.

Yet, the reality so far has not quite met these high expectations.

In an effort to shift this paradigm, Visa is taking bold steps by integrating AI agents with its credit card services.

According to the Associated Press, users will soon be able to establish budgets and set preferences, enabling these advanced AI agents—anticipated successors to current conversational AI like ChatGPT—to shop for a variety of items on behalf of their users.

Whether it’s locating a stylish sweater, purchasing weekly groceries, or booking an airplane ticket, these agents are designed to streamline everyday shopping experiences.

Jack Forestell, Visa’s Chief Product and Strategy Officer, shared his vision during a recent interview, asserting that this initiative could be transformational, akin to the revolutionary impact ushered in by e-commerce itself.

This past Wednesday, Visa announced an exciting partnership with a coalition of leading AI development firms, including prominent U.S. companies like Anthropic, Microsoft, OpenAI, and Perplexity, alongside France’s Mistral.

Together, they aim to enable these AI systems to interface seamlessly with Visa’s vast payments network. In addition to these collaborations, Visa is also engaging with tech giants like IBM, online payment facilitator Stripe, and smartphone innovator Samsung to advance this initiative further.

Initial pilot projects have commenced, with broader implementation anticipated in the upcoming year. Visa is optimistic that what currently appears to be an avant-garde concept could soon evolve into a practical solution for handling the mundane aspects of shopping.

For emerging AI startups, Visa’s endorsement offers a significant opportunity to enhance their competitiveness against established tech behemoths such as Amazon and Google, which currently dominate the landscape of digital commerce and are also exploring their own AI agent capabilities.

Although the tech industry has already showcased various examples of what it terms “agentic AI,” the actual application of these concepts remains limited.

Many of these systems are, at their core, advanced iterations of large language models—the foundational technology that powers chatbots capable of crafting emails, summarizing lengthy documents, or assisting with coding tasks.

Trained on extensive datasets, these models can comb through the vast expanse of the internet to provide product recommendations, but they still face challenges when it comes to executing more complex tasks that require a higher level of autonomous decision-making.