Credit Cards Starting to Compete? Visa Acknowledges AI Agents Need New Payment Systems

Artificial Intelligence (AI) continues to develop very rapidly. If previously AI was better known as a chatbot that was able to answer questions or help create content, now the technology has entered a new phase through the presence of AI Agents (AI Agent) which are able to make decisions, carry out tasks, and even carry out transactions independently.

This change has begun to attract the attention of various global technology and financial companies. In a recent report released jointly with Artemis in July 2026, Visa revealed that traditional payment systems, including credit cards, are beginning to face challenges because they have not been designed to serve the economic activities carried out by AI automatically. In the midst of these changes, many investors are still using the bitcoin price as one of the indicators of the development of the blockchain industry, which is the foundation for the birth of a new generation of digital payment innovation.

According to Visa, these obstacles do not mean that credit cards will be abandoned immediately. However, the increasing use of AI Agents is expected to create a need for new payment infrastructure that is faster, cheaper, and able to process millions of microtransactions automatically.

So, what exactly is an AI Agent? Why is this technology beginning to be considered capable of changing the way digital transactions are carried out? And why is stablecoin touted as one of the most suitable solutions? Here's the complete explanation.

What is an AI Agent?

Before discussing why Visa believes the current payment system needs to evolve, it is important to first understand what is meant by an AI Agent.

Simply put, an AI Agent (AI Agent) is an artificial intelligence system that is not only able to provide answers like a chatbot, but can also observe conditions, make decisions, plan actions, and then execute them independently to achieve certain goals.

Unlike conventional AI that generally waits for user instructions at each step, AI Agents can run a series of processes automatically without having to be constantly directed by humans.

For example, when someone asks a regular chatbot to find the best laptop price, the chatbot will display a list of recommendations. But on AI Agents, the process can be much more complex. AI can search for various stores, compare specifications, calculate shipping costs, choose the best offer, and place an order if it has obtained permission from the user.

It is this ability that makes AI agents beginning to be considered the next step in the development of artificial intelligence.

How Does an AI Agent Work?

Although it looks complex, the way the AI Agent works actually follows a fairly systematic cycle. Each stage is interconnected so that AI can act like a digital assistant who is able to complete the task from start to finish.

In general, the process consists of five main stages.

1. Observe the Environment (Observe)

The first step is to collect information from various sources, such as websites, APIs, databases, sensors, and other applications.

For example, when asked to find the cheapest plane ticket, the AI Agent will access various booking services to obtain real-time data on prices, schedules, and seat availability.

2. Analyzing Information (Reason)

After the data is obtained, AI begins to evaluate all of this information using a large language model (Large Language Model/LLM), logic rules, and decision-making algorithms.

At this stage, AI not only understands the data, but also considers the best possible options based on the user's given goals.

3. Make a Plan (Plan)

Furthermore, the AI Agent makes a strategy to complete the task efficiently.

For example, if the price of a service is too expensive, AI can find other alternatives that have similar quality but at a lower cost.

This planning stage is what distinguishes AI agents from ordinary chatbots because AI is able to determine the next step independently.

4. Take Action (Action)

After making a decision, AI starts executing various actions, such as sending emails, ordering cloud services, buying data, making reservations, and making payments if it has the necessary authorization.

It is this ability to take real action that is the main focus of the Visa report.

5. Learning from the Results (Learning)

After the task is completed, AI can evaluate the results of its work to improve performance on the next task.

The more experience an AI agent has, the better it will be able to make decisions in the future.

Through this cycle, AI agents are no longer merely a tool, but begin to play a role as digital economic actors who are able to interact with various services automatically.

Why is AI Agents Becoming a Discussion in 2026?

The popularity of AI Agents is increasing rapidly not only because of their increasingly sophisticated capabilities, but also because their supporting ecosystem is developing very quickly.

Various technology companies are now competing to build platforms that allow AI to carry out cross-application tasks. AI is no longer only used to generate text or images, but is beginning to be used to manage business operations, conduct data analysis, assist customer service, and automate administrative processes.

This development is also driven by the increasing ability of AI to understand APIs, interact with various digital services, and manage complex workflows.

In the report, Visa and Artemis even mentioned that since mid-2025, AI Agents have passed a key point in their evolution. AI is now able to find APIs it has never known, compare prices from various service providers, and then decide on payments independently if necessary.

This ability is the foundation of the birth of a new concept that is beginning to be known as the Autonomous AI Economy, an economic ecosystem that not only involves humans, but also machines that are capable of carrying out economic activities independently.

However, the growing capabilities of AI Agents have created new challenges. When millions of AIs start making small transactions every day, a big question arises: are the payment systems that humans use today really ready to serve them?

Why Visa Is Assessing Credit Cards Is No Longer Enough?

After understanding how the AI Agent works, the next question is why Visa concluded that traditional payment systems are beginning to face limitations.

The answer is not because credit cards have decreased in use or will soon be replaced. Instead, Visa sees a major change in who is making transactions.

For decades, payment systems were designed with the assumption that transactions were made by humans. From shopping in stores, paying bills, to subscribing to digital services, almost everything involves human decisions and a relatively limited number of transactions.

However, the pattern began to change when the AI Agent was able to act independently.

In a joint report with Artemis, Visa explained that AI Agents can now find digital services themselves, compare prices, choose the best provider, and make payments without human intervention at every step. This kind of activity has the potential to generate millions of small transactions every day, something that has never been the main focus of conventional payment systems.

In other words, the biggest challenge lies not in the value of transactions, but in their frequency and speed.

This is the reason why Visa believes that payment infrastructure needs to evolve to be able to support a digital economy that is starting to be driven by machines.

From Human Commerce to Machine Commerce

To understand the change, imagine how humans and AI Agents transact.

A user usually only makes a few payments a day, for example buying food, paying for transportation, or subscribing to streaming services.

In contrast, one AI agent can perform dozens to thousands of small transactions at the same time.

For example, an AI Agent tasked with managing an online store automatically can perform the following activities in minutes:

Buy API access to check inventory. Pay for automatic translation services. Rent cloud computing for a few seconds. Buy the latest market data. Pay for cyber security services. Access premium AI models to solve specific tasks.

Each transaction may only be worth a few cents or even less than one cent of a US dollar. But if it is done millions of times by millions of AI Agents around the world, the volume will be very large.

This is what many analysts are starting to call machine commerce, that is, economic activities carried out between digital systems without direct human interaction.

The change is one of the reasons why many technology companies are beginning to develop payment infrastructure that is designed for the needs of machines.

What is Machine-to-Machine Payment?

One of the terms that is increasingly appearing in the discussion of AI Agents is machine-to-machine payment (M2M payment).

Simply put, machine-to-machine payment is a transaction that is automatically carried out between two systems or devices without the need for a human to press the payment button.

For example, imagine you have an AI Agent tasked with managing a digital marketing campaign.

When it needs the latest search trend data, the AI Agent can automatically purchase access to a certain data provider. After the data is received, the AI then pays for computing services to analyze the information. If you need promotional images, the AI can again buy credits from an AI-based image-making service.

The entire process can take place in a matter of seconds.

Humans only determine the final goal, while the entire series of transactions is carried out by AI.

This concept is the basis for the development of the digital economy based on AI Agents.

Why is Micropayment a Challenge?

The more automated transactions are carried out, the more important it is to have a system that is capable of processing very small payments or micropayments.

Micropayment is a transaction with a very low nominal, for example only a few cents or even a fraction of a cent of a dollar.

For humans, this type of payment may not be too common. However, for AI Agents, micropayments can actually be the main activity.

As an illustration, an AI Agent can incur the following costs in a series of works:

Activity

Example of Cost

Calling the weather API

US$0.001

Access premium AI models

US$0.003

Using translation services

US$0.002

Storing temporary data in the cloud

US$0.001

Taking market data

US$0.004

The value is indeed very small.

However, if an AI Agent performs thousands of transactions like this every day, the transaction fee can actually be larger than the value of the service purchased if you still use the traditional payment system.

This is where Visa sees an obstacle that needs to be overcome.

Why Credit Cards are Less Suitable for Micropayments?

It doesn't mean credit cards have no role in the future. However, according to Visa, the mechanism currently used is designed for human transaction patterns, not for automated transactions in very large amounts.

Some of the challenges highlighted include the following.

1. Transaction costs are relatively high

Every payment using a card generally involves various parties, ranging from merchants, payment processors, card networks, to issuing banks.

The process produces costs that are still reasonable for large-value transactions, but become less efficient if the transaction is only worth a few cents.

2. Transaction completion takes time

Although card payments feel instant to users, the settlement process behind the scenes still takes a certain amount of time.

Meanwhile, AI agents often require payments that can be completed almost immediately in order to continue the next process without obstacles.

3. Not designed for machine transactions

The card system is developed on the assumption that each transaction involves a user who authenticates, verifies identity, and approves payment.

On the other hand, AI Agents need a mechanism that allows transactions to take place automatically but still be safe according to the rules set by the owner.

4. Very high transaction volume

If millions of AI Agents make small payments every second, the payment infrastructure must be able to handle a much larger volume of transactions than the current human payment pattern.

That is why Visa believes that the evolution of payment infrastructure is an inevitable need as the AI-based economy continues to grow.

Stablecoin is Considered More Suitable for AI Agents

In the report, Visa and Artemis do not conclude that stablecoins will completely replace credit cards. Instead, both see opportunities where various payment methods can complement each other as needed.

For transactions between machines that require low fees, fast settlement, and can be programmed automatically, stablecoins are considered to have a number of advantages.

First, stablecoins can be sent through the blockchain network in a relatively short time without having to follow a long transaction settlement process like traditional payment systems.

Second, transaction costs on some blockchain networks tend to be lower so that they are more suitable for micropayment needs.

Third, stablecoins are programmable, meaning payments can be automatically executed based on certain rules through a predetermined smart contract or system.

Fourth, blockchain-based payments also offer transparency because each transaction can be verified on the public network according to the characteristics of the blockchain used.

Visa even assesses that the future of payments is likely not to be dominated by one system alone. Instead, payment cards, stablecoins, and various machine payment protocols are expected to run side by side according to their respective needs.

This view suggests that the development of AI Agents not only drives innovation in the field of artificial intelligence, but also opens a new chapter in the evolution of digital payment systems.

Examples of Technologies that Start to Support AI Agent Payments

Although the concept of automatic payment between machines is still relatively new, a number of companies have actually begun to develop technology that can support this need.

One example highlighted in the Visa and Artemis report is x402, a payment protocol developed by Coinbase to facilitate automated transactions on the internet.

This protocol allows an application or AI Agent to make a payment directly when accessing certain digital services. For example, an AI Agent who needs weather data, access to a premium AI model, or market analysis services can directly make a payment to the service provider without having to go through a checkout process as humans usually do.

According to data in the report, since its launch in May 2025, x402 has processed more than 109 million transactions with an adjusted transaction value (adjusted volume) reaching around US$15 million. This growth shows that the need for a special payment system for machines is beginning to grow, although it is still in its early stages.

In addition to x402, the report also highlights the Machine Payment Protocol (MPP) developed by Tempo.

Unlike conventional payment systems, MPP is designed to be able to connect blockchain-based payments and fiat currencies through the same framework. Visa has even developed a Card Specification SDK to support the integration of card-based payments into the ecosystem.

The presence of these various innovations shows that the payment industry is no longer only focused on transactions between people, but is also beginning to prepare infrastructure that can be used by AI Agents.

Will AI Agents Replace Credit Cards?

Seeing this development, some people may ask if credit cards will lose their role in the future.

The answer is no.

In the same report, Visa actually describes the future of payments as a hybrid ecosystem, where various payment methods will complement each other as needed.

For example, credit cards are still expected to be the main choice for human transactions on the existing merchant network, such as shopping at stores, booking hotels, or buying travel tickets.

Meanwhile, stablecoins are seen as more suitable for automated payments between machines that require low fees, fast settlement, and can be programmed automatically.

This means that there is no competition between credit cards and stablecoins, but rather a division of roles based on their respective characteristics.

With this approach, AI Agents can still utilize various payment methods according to the context of the transaction being carried out.

What is the Impact on the Blockchain and Crypto Industry?

The development of AI agents not only impacts the artificial intelligence industry, but also has the potential to accelerate the adoption of blockchain technology in everyday life.

Until now, blockchain has been more often associated with crypto assets or digital investments. However, in recent years, the technology has begun to be used as the foundation for a variety of new services, including asset tokenization, digital identities, and cross-border payment systems.

If AI Agents really become part of global economic activity, the need for fast, transparent, and programmable payments is likely to increase.

This is where blockchain has a great opportunity to play a role as an infrastructure that supports automated transactions between machines.

It is not surprising that stablecoins are one of the digital assets that are most often mentioned in various discussions regarding the future of AI payments. Its relatively stable characteristics compared to other crypto assets make stablecoins more suitable as a means of transaction than an investment instrument.

However, this development is still in its early stages and will be greatly influenced by various factors, ranging from regulations, technology readiness, to the level of adoption by industry players.

Why Should the Topic of AI Agents Begin to Be Understood?

Technological developments often start from something that looks simple before finally becoming part of everyday life.

The Internet, cloud computing, and smartphones have all experienced the same phase.

Now, many observers see AI agents as one innovation that has the potential to follow that pattern.

Not because AI will replace humans completely, but because AI can take over various repetitive, administrative, and require interaction with many digital services simultaneously.

If the prediction is true, then understanding of AI Agents is no longer only relevant for technology developers, but also for investors, business actors, and the general public who want to follow the development of the digital economy.

Therefore, understanding the relationship between AI, blockchain, stablecoins, and modern payment systems is becoming an increasingly important resource. If you want to deepen your understanding of blockchain technology, digital assets, and other basic concepts, you can start learning crypto through trusted educational resources to have a stronger knowledge foundation.

Conclusion

The emergence of AI Agents marks a major change in the way technology interacts with the digital world. If previously AI only functioned as an assistant that waited for human instructions, now AI Agents are beginning to be able to observe the environment, make decisions, and take action independently.

The latest report from Visa and Artemis shows that this development brings new challenges for traditional payment systems. The infrastructure that has been designed for human transactions is not fully capable of supporting the needs of large-scale automated payments made by machines.

However, this does not mean that credit cards will be replaced in the near future. Instead, Visa sees the future of payments as a combination of various technologies, where credit cards, stablecoins, and inter-machine payment protocols can run side by side according to their respective needs.

For the blockchain and digital asset industries, this development is a signal that the role of blockchain technology is no longer limited as the foundation of cryptocurrencies, but also has the potential to become the backbone of an increasingly automated digital economy.

FAQ1. What is an AI Agent?

AI agents are artificial intelligence systems that are able to observe conditions, make decisions, plan, and carry out tasks independently to achieve certain goals without having to receive instructions at every step.

2. What is the difference between an AI agent and a chatbot?

Chatbots generally only provide responses based on user questions. Meanwhile, AI Agents can run a series of actions automatically, such as searching for information, comparing services, to making transactions if they have obtained permission.

3. Why does Visa say the old payment system needs to evolve?

According to a report by Visa and Artemis, traditional payment systems are designed for human transactions with a relatively low frequency. Meanwhile, AI Agents are expected to carry out millions of microtransactions automatically, requiring a more efficient infrastructure.

4. What is micropayment?

Micropayment is a transaction with a very small amount, even less than one US dollar. This type of transaction is expected to be increasingly commonly used by AI Agents when purchasing digital services or accessing APIs.

5. Why are stablecoins considered suitable for AI Agents?

Stablecoins offer relatively low transaction fees, faster payment settlement on certain networks, and can be programmed through smart contracts. These characteristics make them more suitable for supporting automatic payments between machines.

6. Will AI agents replace credit cards?

No. Visa actually predicts that credit cards and stablecoins will complement each other. Credit cards are expected to continue to be used for human transactions, while stablecoins are more suitable for automated payments made by AI Agents.

7. What is machine-to-machine payment?

Machine-to-machine payment is an automatic payment system that is carried out between devices or applications without requiring direct human interaction. This concept is one of the foundations in the development of AI-based Agent economy.

8. How does the AI Agent relate to the blockchain?

Blockchain provides a transparent and programmable infrastructure to support digital transactions. When combined with stablecoins and smart contracts, this technology is considered capable of supporting the needs of automatic payments made by AI Agents in the future.

Tag: ekonomi kartu kredit artificial intelligence