For many Southeast Asian companies, selling across borders has become easier than getting paid across them.
A merchant in Singapore can source from Vietnam, sell to customers in Indonesia, pay a logistics partner in Thailand, and settle invoices with a platform in the US. The commercial opportunity is regional, even global. But the money still moves through a patchwork of banks, card networks, payment providers, foreign exchange desks, and local clearing systems that rarely speak to one another cleanly.
Also Read: Optimising cross-border payments for seamless APAC expansion
That gap has turned cross-border payments into one of the least glamorous but most consequential problems in the region’s digital economy. Fees can be hard to predict. Foreign exchange spreads vary. Settlement timelines differ by country and provider. Treasury teams often do not know exactly where cash sits at any given moment, or whether converting it today will be cheaper than waiting until tomorrow.
A new report by Sunrate and Mastercard, titled “Beyond Automation: Defining Agentic Global Payments”, argues that the next phase of payment technology will not simply automate existing workflows. Instead, it will use AI agents to make decisions across routing, liquidity, reconciliation, and foreign exchange in near real time.
The distinction matters. Traditional automation follows fixed rules: if an invoice is due, send a payment; if a transaction fails, try again through another channel. Agentic AI goes further. It refers to systems that can interpret changing conditions, weigh different options, and recommend or execute the best course of action within set guardrails.
In payments, that could mean choosing the lowest-cost route for a transaction, deciding when to convert currencies, identifying mismatches between invoices and receipts, or flagging only the exceptions that require human review.
The liquidity problem hiding in plain sight
The report identifies one pain point that will sound familiar to many finance teams: the “liquidity blind spot”. This refers to the lack of timely visibility into where cash is, what currency it is held in, and when it is needed.
For large corporations, this is a treasury headache. For startups and SMEs, it can be existential.
A regional e-commerce exporter may receive US dollars, pay suppliers in Chinese yuan, settle logistics bills in Thai baht, and cover payroll in Indonesian rupiah. If the company converts too early, it may lose out when rates move favourably. If it converts too late, it may face higher spreads or a cash crunch. If finance teams rely on end-of-day reports and manual spreadsheets, they are often reacting to yesterday’s position rather than managing today’s risk.
Agentic systems could reduce that lag. According to the report, AI agents can monitor FX trends and liquidity needs, then recommend or execute conversion timing based on real-time market data. In practice, this shifts treasury from periodic matching to continuous reconciliation. Instead of staff manually checking every payment and bank entry, the system reconciles routine flows and flags genuine discrepancies.
That may sound technical, but the business impact is straightforward: less trapped cash, fewer avoidable FX losses, and faster decisions about where to deploy working capital.
Why APAC is fertile ground
The Asia Pacific region is a particularly relevant test bed for this model because its trade and payment flows are both fast-growing and fragmented.
The report notes that commercial card transaction volume in the travel segment alone is growing at a 28 per cent compound annual growth rate from 2023 to 2025. Travel is one of the clearest examples of why cross-border payments are difficult in this region. Online travel agencies, hotel operators, airlines, destination management companies, and corporate travel platforms may operate across dozens of markets, currencies, and settlement arrangements.
Also Read: How fiat and crypto are redefining cross-border payments
A booking made in Malaysia for a hotel in Japan through a Singapore-based platform may involve several parties before the final merchant receives funds. Each leg can add cost, delay, or data loss.
The same pattern appears in other sectors central to Southeast Asia’s startup economy: B2B marketplaces, logistics, software-as-a-service, gaming, creator platforms, and cross-border e-commerce. These businesses scale by connecting demand and supply across markets. Their finance operations, however, often become more complex with every new country added.
Emerging markets across APAC, Latin America, and the Middle East and Africa are seeing commercial card transaction growth of more than 20 per cent, according to the report. That growth creates a larger data trail, but also more routing choices and more operational risk. Static payment setups are less suited to this environment because fees, failure rates, FX conditions, and local payment rails can change quickly.
From fixed rails to intelligent routing
One of the report’s central ideas is “Intelligent Payment Routing”. In simple terms, it means allowing AI agents to select the most efficient provider, payment rail, or route for a transaction based on the transaction type, market conditions, and historical performance.
Today, many companies still use static routing. A payment to one country goes through a preferred provider, while a payment in another currency follows a pre-set bank channel. That may be manageable at low volumes, but it becomes inefficient as a business expands across markets.
Intelligent routing could compare options dynamically. For example, it may decide that one provider is cheaper for a low-value supplier payout, while another is more reliable for high-value settlement. It may avoid a route with historically high failure rates during local banking cut-off times. It may select a card rail for speed in one case and a bank transfer for cost in another.
For Southeast Asian companies, this flexibility is especially useful because regional expansion rarely follows a neat path. A startup may begin in Singapore, add Indonesia, then serve the Philippines, Thailand, and Vietnam within a short period. Each market brings its own banking infrastructure, regulatory expectations, payment behaviours, and currency considerations.
Agentic payment systems will not remove that complexity entirely. But they can help businesses manage it without building large treasury teams before they have the scale to justify them.
The working capital argument
The strongest case for agentic payments may not be lower fees alone. It is working capital.
The report states that finance teams using intelligent automation in accounts receivable can reduce manual effort by up to 40 per cent. It also notes that leading platforms integrating AI agents can help companies reduce Days Sales Outstanding, or DSO, by up to 12 days and cut manual follow-ups by 50 per cent.
DSO measures how long it takes a company to collect payment after a sale. A reduction of 12 days can be meaningful for a growing business. It means cash arrives earlier, reducing the need for short-term borrowing or delaying supplier payments. In a funding environment where venture capital is more selective than it was during the 2021 boom, operational cash efficiency has become a competitive advantage.
This is particularly relevant in Southeast Asia, where many startups serve SMEs or operate in sectors with thin margins and uneven payment cycles. Faster collections and better reconciliation can give founders more room to invest in inventory, marketing, hiring, or market expansion.
Automation with guardrails
The promise of agentic AI in payments is significant, but it also raises practical questions. Finance teams will need clear controls over what AI agents can execute autonomously and what requires approval. Regulators will expect audit trails. Businesses will need explainability, especially when a system chooses one route, provider, or FX timing over another.
There is also the matter of trust. Companies may be willing to let AI recommend a conversion window or flag suspicious reconciliation items. They may be slower to allow autonomous execution of large transactions without human oversight.
Also Read: Singapore’s regulatory vision is shaping cross-border payments in Asia: Report
That suggests adoption will likely be gradual. AI agents may first handle low-risk tasks such as matching invoices, suggesting routes, detecting anomalies, and preparing payment recommendations. Over time, as accuracy improves and controls mature, they could take on more direct execution.
The direction, however, is becoming clearer. Cross-border payments are no longer just a back-office function. For companies expanding across Asia Pacific, they shape margins, customer experience, supplier relationships, and cash flow.
If agentic AI can make global payments less opaque and more responsive, it could help Southeast Asian businesses compete beyond their home markets without being slowed by the plumbing underneath.
The post AI agents could help Southeast Asian firms untangle cross-border payment costs appeared first on e27.