More and more areas of life are being reshaped by artificial intelligence, including places we might least expect, such as shopping. What still sounds like the future today is already beginning to take concrete form. AI agents are increasingly taking over tasks that were previously reserved for humans, from searching and comparing to deciding and purchasing.

Imagine simply describing your needs and preferences while an AI works in the background, evaluating and comparing all options and ultimately paying on your behalf, without any further input from you. This is exactly where agent-based commerce, also known as “agentic commerce,” comes in. While traditional generative AI advises and supports, an AI agent in agentic commerce handles every step of the purchasing process. It searches, evaluates, makes decisions, and completes the payment entirely autonomously.

Agentic Commerce Brings Many Benefits

The potential of agentic commerce is huge. Consumers enjoy an unprecedented level of convenience, saving time and effort as AI handles the research. While the AI finds the best deals, people can focus on other things. Price comparisons and purchases are managed quickly and accurately, tailored to individual needs. This also changes the entire customer journey. The classic “moments of truth,” when a customer makes a conscious decision, are increasingly replaced by the AI’s invisible micro-decisions. This makes brand loyalty harder to maintain, as agents make choices based primarily on price, performance, and other rational criteria rather than emotional attachment to a brand.

Businesses stand to gain as well. AI agents can manage supply chains, monitor inventory, and proactively optimize purchasing. Prices can be refined through algorithmic negotiation, boosting efficiency and improving terms for everyone. Entirely new business models are emerging. For instance, subscription agents could handle regular repeat purchases, or specialized “deal-hunter” agents could continuously scan for the best prices, all without human intervention.

The New Buying Experience: Invisible Payments and Ongoing Transactions

AI agents are fundamentally changing how we shop. In the future, payments will happen in the background, guided by user preferences but requiring no active participation. Shopping becomes an automated, ongoing process. The AI checks prices, reorders items as needed, and responds in real time, often before the consumer even realizes an offer is available or supplies are running low.

For retailers, this means a shift from classic marketing to making products technically discoverable and machine-readable. Only products that are well-structured and accessible in machine-friendly formats will be visible to AI agents. Companies will need open interfaces and high-quality data. Performance marketing and SEO will give way to “Agent Experience Optimization” (AXO), which focuses on optimizing data for AI agents rather than human shoppers.

Payment service providers will need to adapt as well. The traditional shopping cart may disappear, replaced by efficient API infrastructures, tokenization, and fast, secure real-time transactions.

Challenges: Legal, Technical, and Regulatory Uncertainties

Agentic commerce also brings significant challenges. Until now, payment and commerce have been built around human decision-makers. Innovations like credit cards, e-wallets, and mobile payments were designed to make transactions convenient and secure for people. AI agents change this dynamic. Humans step back from the payment process.

This raises a host of new questions. How can we ensure that AI agents act in the best interests of their users? How are chargebacks handled if no human directly initiated a purchase? How much control do users have, or want to give up? How are security tokens for agents generated and managed? Who is liable for mistakes or fraudulent transactions, especially when commerce operates globally and continuously without human approval?

There are many technical and regulatory issues to resolve. Security tokens for AI agents are not yet standardized, and traditional models for fraud prevention and risk assessment may not be sufficient. Constantly active transaction patterns require new, advanced security and monitoring approaches. At the same time, legal and regulatory frameworks lag behind, leaving open questions about liability, responsibility, and compliance. Legislators will need to create an environment in which agentic commerce can develop in a way that is both safe and beneficial to consumers.

Massive Market Potential: Billions of Transactions Driven by AI Agents

Where do things stand today? The potential for agent-based commerce is enormous. Analysts expect that by 2029, AI agents could handle between 1% and 4% of all digital payment transactions. With a projected total transaction volume of over $36 trillion a year, even a small share translates into a market worth up to $1.47 trillion. AI could also boost conversion rates by removing human hesitation and reducing the likelihood that customers abandon their cart during checkout.

Even if their current market share is small, the impact could be significant. Transactions that are regular, predictable, and standardized, such as B2B payments, recurring subscriptions, international trade, and micropayments, are especially suitable for agent-driven automation. AI agents can handle these efficiently and autonomously, with minimal human involvement. For more complex, high-value, or unique purchases, people will likely remain part of the process.

Practical Examples and Market Developments

Agentic commerce is still in its early stages, but progress is rapid. For example, the US company Perplexity AI, a generative AI provider focused on commerce, launched an AI agent in September 2024 that could independently handle orders. It is not yet fully reliable. The agent reportedly required multiple attempts and several hours to buy a tube of toothpaste. These limitations show that widespread adoption will take time, but the direction is clear.

Major payment providers like Visa and Mastercard are also investing in this field. In early 2025, both announced new AI commerce services. Mastercard is focusing on agent-based tokenization, while Visa is developing a scalable platform for data protection and risk management. Other payment providers and platform operators are moving forward as well, enabling businesses and developers to automate payments, billing, shipping, and subscription management through AI-ready interfaces. Integration barriers are being lowered, allowing even non-technical users to adopt AI-powered workflows.

Conclusion: The Future Is Closer Than You Think

The rapid progress of artificial intelligence means agentic commerce is no longer just a distant vision. It is already becoming a reality. To revisit the Perplexity example, just six months later, US customers could book trips, order products, or buy tickets while payments were handled automatically in the background via services like PayPal or Venmo. The path to autonomous commerce is well underway.

How widely this technology will be adopted depends on people’s willingness to let machines handle their purchases and the level of trust they place in these systems. One thing is clear: agentic commerce has the potential to fundamentally transform the shopping experience for consumers, merchants, and the financial sector as a whole. Companies that engage early stand to gain a significant advantage.

‍