Walking through the marketplace as a tourist, you see a lively scene: people bustling around, eyes fixed on goods, comparing items, tasting, bargaining with vendors, exchanging money. It looks like a one-time transaction—each interaction is a small negotiation, trust maintained through cash, or value exchanged via bank cards.
But this isn’t how most transactions at the marketplace operate. Look closely: most people are locals heading purposefully to their favorite merchants. restaurant owners visit friends, butchers, fishmongers, and farmers. Tailors go to repair shops, weavers, and artisans. They all use credit.
When we discuss how smart agents will pay, we often unconsciously think from a tourist’s perspective.
But smart agents will behave more like locals. The difference between smart agents and humans lies in their characteristics—limitless replication, flexible resource allocation, zero startup costs—which means a few smart agents can dominate niche markets. Even as creating smart agents becomes easier, human relationships, partnerships, and trust still help craft successful user experiences. Dominant smart agents don’t need tourist payment channels; they need vendor relationships, operational capital, and credit. Smart agents can guide tourists (that’s you).
What does this specifically mean? As smart agents integrate into business platforms, their payment methods must shift from retail channels to pre-negotiated B2B terms and credit, which current payment channels can’t fully support. If entrepreneurs can develop excellent solutions for next-generation payment scenarios (like smart agents, streaming payments, and high-frequency, low-value transactions for global businesses), then next-gen payment channels (like stablecoins) will have growth opportunities.
This article explores this idea from three angles: how the differences between smart agents and humans influence which payment strategies succeed; the shortcomings of current methods; and what elements the next-generation payment channels need to succeed.
Differences Between Smart Agents and Humans
To understand the relationship between smart agents and payments, we must consider two questions: Will smart agents behave more like humans or like enterprises? Will they focus on long-term or short-term interests?
Smart agents will resemble enterprises, establishing long-term relationships with vendors and partners. They are lightweight customized entities built on top of large organizational structures—like a perfect tour guide provided by a well-connected travel agency, or franchisees that can adapt services to local tastes without renegotiating supply chains.
Why Will Smart Agents Behave Like Enterprises?
First, the best experience comes from careful design. I don’t want a smart agent that’s still negotiating with suppliers, comparing prices, and bargaining at checkout. I want a smart agent that has already done that—knows which suppliers are reliable, has pre-negotiated prices, and can settle immediately. That’s true business relationship, not just travel transactions.
In fact, human agents have long existed: travel agents are one example, but literary agents, talent agents, watch dealers, real estate brokers—all are similar. Agents build key multi-party relationships—with publishers, production companies, watch distributors, or mortgage lenders—and each transaction is customized based on these relationships.
Second, while smart agents can be infinitely replicated, scalable business (and its advantages) cannot. Excellent smart agents leverage the benefits of scale: lower computational costs, better vendor pricing, deeper integrations, and more deterministic components. Scale breeds scale. A travel agency booking a million tickets a year can negotiate better terms with airlines than one booking only ten.
We see this trend already. Only ChatGPT has enough channels to negotiate partnerships with Shopify, Amazon, Expedia, and others. Small startups can only use automated browsers or reverse-engineered APIs, paying high retail fees.
That’s why smart agents tend to integrate or at least operate on larger platforms. Building agents is easy, but economic efficiency dictates that each vertical’s agent count should be limited—each should build deep vendor relationships and have enough profit margin to reinvest in improving user experience. Moreover, vertical-specific, vendor-strong agents can work alongside user agents, creating a win-win situation.
Two Types of Payment Relationships
If smart agents operate like enterprises, then two payment relationships must be designed: user → agent, and agent/agent platform/guide → vendor.
Users pay the agent—via subscription, task-based fees, credit lines, or authorized access to user accounts. The agent pays vendors through negotiated B2B terms, bulk pricing, 30-day invoicing, or via sub-agents. Currently, the agent occasionally pays vendors through retail channels, but this only accounts for a small part of total expenditure.
This reflects how credit cards work today: issuing banks establish retail relationships with consumers, bear risk, offer personalized rewards, and provide credit limits. Acquirers build business relationships with merchants, negotiate terms, handle bulk transfers, and manage complex working capital.
Smart Agents and Credit Cards: A McKinsey-Style Perfect Match
As many say, credit cards are actually a pretty reasonable payment product for smart agents. They are widely accepted; payments between $20 and $1,000 are considered reasonable; and they have built-in arbitration, cancellation, and digital features.
Credit cards also provide monthly statements—an important way for consumers to understand their spending—and as smart agents replace kids playing on iPads as the main source of accidental expenses, this concept will be further refined.
But there are two issues: first, credit cards are technically a poor fit for smart agents. Second, their fee models trap the industry in a typical innovator’s dilemma.
Credit Card Technology Is Difficult to Upgrade
Almost all credit card tech relies on human intervention: approval processes, user interfaces, and traditional payment methods (single payments, subscriptions). Stripe Link, Visa 3D Secure, and dozens of other virtual card products—allowing you to save cards on websites or register cards for recurring monthly subscriptions—are finally working well, but this tech has taken 15 years to develop.
The adoption of smart agents is happening so fast that thousands of payment service providers (PSPs), POS terminals, merchants, and client devices cannot slowly upgrade their interfaces, programmability, and fraud detection to accommodate this new payment flow.
Credit Cards Cannot Support High or Low-Value Transactions
Imagine a smart agent remitting to a compute provider or paying small API access fees. Neither can be done via credit card channels. Visa, for example, doesn’t support payments below one cent; its business model expects a fixed fee of 30 cents. While Visa might develop streaming or micro-payment tech, convincing stakeholders to accept lower revenue is even harder.
More troubling, credit cards are caught in the innovator’s dilemma. Although smart agent payments share similar user relationships and needs as credit card payments, their amounts often exceed $20–$1,000. Worse, many initial solutions involve paying for APIs that are hard to refund or resell (fraud risk). Credit cards are feasible but the innovator’s dilemma has long eroded existing businesses.
Even setting aside credit cards, traditional payment channels will still have a role in the future.
Existing Payment Methods Will Still Play a Role
As smart agents integrate into entities resembling business platforms, most large expenditures will shift to pre-negotiated B2B terms: invoices, net-30 payments, discounts, and credit lines. In that world, “payment channels” can be anything—often asynchronous settlements over traditional channels, somewhat dull. Fees are shared across larger transactions, and working capital can be negotiated between parties.
But the space for smart agents isn’t limited to this. They are already operating in areas where traditional payment methods struggle: first-time collaborations, cross-border payments, streamlining complex reconciliation, new agent-vendor models, instant payments to reduce borrowing costs, and microloans.
In these scenarios, stablecoins are a better payment option, and crucially, building the next generation of features on programmable money is much easier than on traditional infrastructure. New partnerships established with stablecoins will evolve into ongoing relationships. As stablecoin payment platforms go live, stablecoins (which are cheaper, faster, and more global) are likely to occupy an increasingly important position in payment mixes.
Opportunities in New Payment Technologies
To understand future trends, we should focus on technologies best suited for growing application scenarios.
Stablecoins—a faster, cheaper, globally accepted currency backed 1:1 by high-quality liquid assets—are a new platform capable of meeting unmet needs in commercial sectors like international payments and streaming payments. Importantly, stablecoins are programmable. Key functions like arbitration, monthly (or hourly) settlement, credit, escrow, and conditional payments can be flexibly extended to support many new use cases. Unlike bank or credit card payments, stablecoin payments can be easily integrated into APIs, databases, and agent checkout systems, greatly simplifying reconciliation, approval, and onboarding—crucial for entrepreneurs building agent businesses.
Practically, stablecoins solve the unit economics issues of credit cards in extreme cases. They have no $0.30 minimum fee, avoiding small-value payment problems. They don’t erode large transfer profits with interchange fees. Smart agents pay compute providers $0.001 per second, while manufacturers settle $50,000 vendor invoices—all on the same payment channel. This flexibility is vital for engineers and entrepreneurs considering their next platform.
Building More Stablecoin Infrastructure
The most common objection to stablecoins is high deposit and withdrawal costs. For tourists unfamiliar with stablecoins, this is true—but if a tour guide or smart agent assists, the problem is solved. Guides can help tourists exchange currency and facilitate necessary transactions precisely, saving on fees.
Adding billing and arbitration features to our stablecoin-supported guide services brings us closer to an ideal system.
Imagine shopping in a department store. You browse multiple vendors, add items, and finally settle a consolidated bill. The platform handles the complex process of distributing funds to each vendor. Smart agents need the same: a unified view showing purchase intents across multiple vendors, with one-click approval for bulk orders. Users see “Your smart agent wants to book flights, hotels, and car rentals,” rather than three separate checkouts. The agent platform manages vendor relationships, while users handle purchase intents. Users can approve, review, or dispute transactions.
While credit cards handle arbitration well, new payment channels need to extend this. When margins are high or returns are easy, arbitration is most convenient—for example, flights canceled within 24 hours, pre-activated subscriptions, or luxury goods with high margins that can afford refunds. Early agent use cases often involve low-margin digital goods like compute resources, API calls, or food delivery.
Summary
Smart agents won’t pay like tourists. They will pay like locals—through relationships, credit lines, and repeat business. This means real payment flows will be through pre-negotiated B2B terms, not card swipes. Frankly, pre-negotiated B2B terms don’t require new payment channels. Settlement layers can be anything—wire transfers, ACH, or batch transfers. Traditional payment methods are perfectly sufficient for established relationships.
But we are at a pivotal moment. Smart agents are emerging, entrepreneurs are building their systems, and they need payment methods that work immediately—not ones that require years of credit card tech upgrades. Credit cards aren’t ready: they’re too costly for small payments, hard to reconcile, burdened by technical debt, and human factors influence fraud decisions. Stablecoins are mature. They are programmable, global, easy to reconcile with digital services, and can be seamlessly integrated into APIs and agent checkout flows. Even without negotiated merchant agreements or complex B2B terms, they can work from day one.
This is a critical moment. Entrepreneurs building smart agents today will choose tools that can operate effectively right away. Payments are sticky. Ultimately, new relationships built on stablecoins will evolve into old relationships still based on stablecoins. Over the next few years, the ecosystem will mature, barriers to entry will lower, and infrastructure gaps—like billing, arbitration, credit, batch approval, and interoperability—will be filled by startups built on stronger foundations.
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
a16z Crypto: How to do business in the AI Agent economy?
Article author: a16z crypto, SamBroner
Article translation: Block unicorn
Preface
Walking through the marketplace as a tourist, you see a lively scene: people bustling around, eyes fixed on goods, comparing items, tasting, bargaining with vendors, exchanging money. It looks like a one-time transaction—each interaction is a small negotiation, trust maintained through cash, or value exchanged via bank cards.
But this isn’t how most transactions at the marketplace operate. Look closely: most people are locals heading purposefully to their favorite merchants. restaurant owners visit friends, butchers, fishmongers, and farmers. Tailors go to repair shops, weavers, and artisans. They all use credit.
When we discuss how smart agents will pay, we often unconsciously think from a tourist’s perspective.
But smart agents will behave more like locals. The difference between smart agents and humans lies in their characteristics—limitless replication, flexible resource allocation, zero startup costs—which means a few smart agents can dominate niche markets. Even as creating smart agents becomes easier, human relationships, partnerships, and trust still help craft successful user experiences. Dominant smart agents don’t need tourist payment channels; they need vendor relationships, operational capital, and credit. Smart agents can guide tourists (that’s you).
What does this specifically mean? As smart agents integrate into business platforms, their payment methods must shift from retail channels to pre-negotiated B2B terms and credit, which current payment channels can’t fully support. If entrepreneurs can develop excellent solutions for next-generation payment scenarios (like smart agents, streaming payments, and high-frequency, low-value transactions for global businesses), then next-gen payment channels (like stablecoins) will have growth opportunities.
This article explores this idea from three angles: how the differences between smart agents and humans influence which payment strategies succeed; the shortcomings of current methods; and what elements the next-generation payment channels need to succeed.
Differences Between Smart Agents and Humans
To understand the relationship between smart agents and payments, we must consider two questions: Will smart agents behave more like humans or like enterprises? Will they focus on long-term or short-term interests?
Smart agents will resemble enterprises, establishing long-term relationships with vendors and partners. They are lightweight customized entities built on top of large organizational structures—like a perfect tour guide provided by a well-connected travel agency, or franchisees that can adapt services to local tastes without renegotiating supply chains.
Why Will Smart Agents Behave Like Enterprises?
First, the best experience comes from careful design. I don’t want a smart agent that’s still negotiating with suppliers, comparing prices, and bargaining at checkout. I want a smart agent that has already done that—knows which suppliers are reliable, has pre-negotiated prices, and can settle immediately. That’s true business relationship, not just travel transactions.
In fact, human agents have long existed: travel agents are one example, but literary agents, talent agents, watch dealers, real estate brokers—all are similar. Agents build key multi-party relationships—with publishers, production companies, watch distributors, or mortgage lenders—and each transaction is customized based on these relationships.
Second, while smart agents can be infinitely replicated, scalable business (and its advantages) cannot. Excellent smart agents leverage the benefits of scale: lower computational costs, better vendor pricing, deeper integrations, and more deterministic components. Scale breeds scale. A travel agency booking a million tickets a year can negotiate better terms with airlines than one booking only ten.
We see this trend already. Only ChatGPT has enough channels to negotiate partnerships with Shopify, Amazon, Expedia, and others. Small startups can only use automated browsers or reverse-engineered APIs, paying high retail fees.
That’s why smart agents tend to integrate or at least operate on larger platforms. Building agents is easy, but economic efficiency dictates that each vertical’s agent count should be limited—each should build deep vendor relationships and have enough profit margin to reinvest in improving user experience. Moreover, vertical-specific, vendor-strong agents can work alongside user agents, creating a win-win situation.
Two Types of Payment Relationships
If smart agents operate like enterprises, then two payment relationships must be designed: user → agent, and agent/agent platform/guide → vendor.
Users pay the agent—via subscription, task-based fees, credit lines, or authorized access to user accounts. The agent pays vendors through negotiated B2B terms, bulk pricing, 30-day invoicing, or via sub-agents. Currently, the agent occasionally pays vendors through retail channels, but this only accounts for a small part of total expenditure.
This reflects how credit cards work today: issuing banks establish retail relationships with consumers, bear risk, offer personalized rewards, and provide credit limits. Acquirers build business relationships with merchants, negotiate terms, handle bulk transfers, and manage complex working capital.
Smart Agents and Credit Cards: A McKinsey-Style Perfect Match
As many say, credit cards are actually a pretty reasonable payment product for smart agents. They are widely accepted; payments between $20 and $1,000 are considered reasonable; and they have built-in arbitration, cancellation, and digital features.
Credit cards also provide monthly statements—an important way for consumers to understand their spending—and as smart agents replace kids playing on iPads as the main source of accidental expenses, this concept will be further refined.
But there are two issues: first, credit cards are technically a poor fit for smart agents. Second, their fee models trap the industry in a typical innovator’s dilemma.
Credit Card Technology Is Difficult to Upgrade
Almost all credit card tech relies on human intervention: approval processes, user interfaces, and traditional payment methods (single payments, subscriptions). Stripe Link, Visa 3D Secure, and dozens of other virtual card products—allowing you to save cards on websites or register cards for recurring monthly subscriptions—are finally working well, but this tech has taken 15 years to develop.
The adoption of smart agents is happening so fast that thousands of payment service providers (PSPs), POS terminals, merchants, and client devices cannot slowly upgrade their interfaces, programmability, and fraud detection to accommodate this new payment flow.
Credit Cards Cannot Support High or Low-Value Transactions
Imagine a smart agent remitting to a compute provider or paying small API access fees. Neither can be done via credit card channels. Visa, for example, doesn’t support payments below one cent; its business model expects a fixed fee of 30 cents. While Visa might develop streaming or micro-payment tech, convincing stakeholders to accept lower revenue is even harder.
More troubling, credit cards are caught in the innovator’s dilemma. Although smart agent payments share similar user relationships and needs as credit card payments, their amounts often exceed $20–$1,000. Worse, many initial solutions involve paying for APIs that are hard to refund or resell (fraud risk). Credit cards are feasible but the innovator’s dilemma has long eroded existing businesses.
Even setting aside credit cards, traditional payment channels will still have a role in the future.
Existing Payment Methods Will Still Play a Role
As smart agents integrate into entities resembling business platforms, most large expenditures will shift to pre-negotiated B2B terms: invoices, net-30 payments, discounts, and credit lines. In that world, “payment channels” can be anything—often asynchronous settlements over traditional channels, somewhat dull. Fees are shared across larger transactions, and working capital can be negotiated between parties.
But the space for smart agents isn’t limited to this. They are already operating in areas where traditional payment methods struggle: first-time collaborations, cross-border payments, streamlining complex reconciliation, new agent-vendor models, instant payments to reduce borrowing costs, and microloans.
In these scenarios, stablecoins are a better payment option, and crucially, building the next generation of features on programmable money is much easier than on traditional infrastructure. New partnerships established with stablecoins will evolve into ongoing relationships. As stablecoin payment platforms go live, stablecoins (which are cheaper, faster, and more global) are likely to occupy an increasingly important position in payment mixes.
Opportunities in New Payment Technologies
To understand future trends, we should focus on technologies best suited for growing application scenarios.
Stablecoins—a faster, cheaper, globally accepted currency backed 1:1 by high-quality liquid assets—are a new platform capable of meeting unmet needs in commercial sectors like international payments and streaming payments. Importantly, stablecoins are programmable. Key functions like arbitration, monthly (or hourly) settlement, credit, escrow, and conditional payments can be flexibly extended to support many new use cases. Unlike bank or credit card payments, stablecoin payments can be easily integrated into APIs, databases, and agent checkout systems, greatly simplifying reconciliation, approval, and onboarding—crucial for entrepreneurs building agent businesses.
Practically, stablecoins solve the unit economics issues of credit cards in extreme cases. They have no $0.30 minimum fee, avoiding small-value payment problems. They don’t erode large transfer profits with interchange fees. Smart agents pay compute providers $0.001 per second, while manufacturers settle $50,000 vendor invoices—all on the same payment channel. This flexibility is vital for engineers and entrepreneurs considering their next platform.
Building More Stablecoin Infrastructure
The most common objection to stablecoins is high deposit and withdrawal costs. For tourists unfamiliar with stablecoins, this is true—but if a tour guide or smart agent assists, the problem is solved. Guides can help tourists exchange currency and facilitate necessary transactions precisely, saving on fees.
Adding billing and arbitration features to our stablecoin-supported guide services brings us closer to an ideal system.
Imagine shopping in a department store. You browse multiple vendors, add items, and finally settle a consolidated bill. The platform handles the complex process of distributing funds to each vendor. Smart agents need the same: a unified view showing purchase intents across multiple vendors, with one-click approval for bulk orders. Users see “Your smart agent wants to book flights, hotels, and car rentals,” rather than three separate checkouts. The agent platform manages vendor relationships, while users handle purchase intents. Users can approve, review, or dispute transactions.
While credit cards handle arbitration well, new payment channels need to extend this. When margins are high or returns are easy, arbitration is most convenient—for example, flights canceled within 24 hours, pre-activated subscriptions, or luxury goods with high margins that can afford refunds. Early agent use cases often involve low-margin digital goods like compute resources, API calls, or food delivery.
Summary
Smart agents won’t pay like tourists. They will pay like locals—through relationships, credit lines, and repeat business. This means real payment flows will be through pre-negotiated B2B terms, not card swipes. Frankly, pre-negotiated B2B terms don’t require new payment channels. Settlement layers can be anything—wire transfers, ACH, or batch transfers. Traditional payment methods are perfectly sufficient for established relationships.
But we are at a pivotal moment. Smart agents are emerging, entrepreneurs are building their systems, and they need payment methods that work immediately—not ones that require years of credit card tech upgrades. Credit cards aren’t ready: they’re too costly for small payments, hard to reconcile, burdened by technical debt, and human factors influence fraud decisions. Stablecoins are mature. They are programmable, global, easy to reconcile with digital services, and can be seamlessly integrated into APIs and agent checkout flows. Even without negotiated merchant agreements or complex B2B terms, they can work from day one.
This is a critical moment. Entrepreneurs building smart agents today will choose tools that can operate effectively right away. Payments are sticky. Ultimately, new relationships built on stablecoins will evolve into old relationships still based on stablecoins. Over the next few years, the ecosystem will mature, barriers to entry will lower, and infrastructure gaps—like billing, arbitration, credit, batch approval, and interoperability—will be filled by startups built on stronger foundations.