Checkout.com and Agoda Partner on AI Payment Optimization: Acceptance Rates as the Transaction Layer Agentic Commerce Runs On
Reading the Checkout.com–Agoda AI payment partnership as agentic commerce groundwork: how per-transaction optimization lifts acceptance rates, why travel previews AI commerce conditions, and why payment investment is conversion strategy.
Key Takeaways
- Payments platform Checkout.com has announced a partnership with travel booking giant Agoda, deploying Intelligent Acceptance, its per-transaction AI optimization engine, along with network tokens and virtual card issuing for supplier payouts across Agoda's global booking infrastructure
- Rather than the conversational front end where most AI commerce attention goes, this applies AI to the back end of the transaction — and the measurable acceptance-rate gains make it a case study in building the rails that agent-mediated commerce will run on
- When AI agents transact on people's behalf, a failed payment converts instantly into a lost sale, which turns investment in acceptance rates, tokenization, and automatic failure recovery into a conversion strategy for retail and e-commerce operators
The Partnership Shoring Up Payments Behind the Discovery Boom
The partnership will see Agoda leverage AI-powered payment optimisation and digital payment technologies to improve transaction performance and reduce friction across its global travel platform.
www.itp.netOn June 30, 2026, London-based payments platform Checkout.com announced a partnership with digital travel platform Agoda. Most discussion of AI commerce — selling in which generative AI and AI agents handle the journey from discovery to purchase — gravitates toward the visible front end of chat-based search and booking. In this deal, the AI works somewhere shoppers never see: inside payment processing. This article reads the announcement not as travel-industry news but as a case study in building the transaction rails that agent-mediated commerce will depend on.
Part of Booking Holdings, Agoda offers more than six million hotels and holiday properties worldwide alongside flights and travel activities, serving customers in 39 languages. Trade publication ITP.net puts the scale at 130,000 flight routes and over 300,000 activities, and notes that Agoda has already consolidated hotels, flights, and experiences into a single checkout flow. Bundling multiple products into one payment means it confronts, ahead of most retailers, the same problem as a multi-item e-commerce cart.
The partnership rests on two pillars. The first is traveler-side payments: on top of Checkout.com's acquiring infrastructure (the processing that accepts card payments), Agoda runs Intelligent Acceptance, the company's AI optimization engine, together with Network Tokens, which replace card numbers with card-network tokens, and Real-Time Account Updater, which automatically refreshes expired or reissued card details. The second is supplier payments: payouts to hotels and other partners are processed through virtual card issuing. Money coming in from travelers and going out to suppliers moves through a single connected platform. Checkout.com itself is a processor that handled more than $300 billion in e-commerce payments volume in 2025 and supports over 145 currencies.
How Per-Transaction AI Lifts Acceptance Rates
The payment acceptance rate is the share of attempted card transactions that are actually approved. Among the declines sit false declines — rejections of transactions that are neither fraudulent nor short of funds. Checkout.com's announcement states that this combination of technologies reduces false declines across the payment flows typical of global travel: cross-border, multi-currency, and high-volume. Every falsely declined transaction recovered flows straight back into revenue, at zero acquisition cost.
At the core sits Intelligent Acceptance, the AI engine Checkout.com launched in June 2023. Trained on more than 20 billion data points flowing through the company's global network, it adjusts routing (which path processes each transaction), authentication, and the message formats sent to card networks in real time, per payment. According to the company's March 2025 announcement, the engine ran an average of 60 million optimizations per day in 2024, lifted merchant acceptance rates by an average of 3.8%, and has unlocked more than $10 billion in recovered revenue for merchants since launch.
Agoda's own results come with numbers attached. ITP.net reports, citing Checkout.com, that Agoda's acceptance rates have improved an average of 3% year on year through these optimization technologies, including a 9% increase between June 2024 and June 2025. Worth noting: these figures do not appear in the official release itself and reached the public through press coverage. Even so, the structure holds — this is a partnership expanding on top of measured results from optimization already in production, not a fresh projection.
| Technology | Role |
|---|---|
| Intelligent Acceptance | AI optimizes routing and processing parameters per transaction, reducing failed payments |
| Network Tokens | Replaces card numbers with card-network tokens, improving security and acceptance simultaneously |
| Real-Time Account Updater | Automatically updates expired or reissued card details, preventing failures from stale credentials |
| Virtual card issuing | Processes supplier payouts via virtual cards, ensuring reliable, controllable disbursements |
Travel's Payment Gauntlet Previews the Conditions of AI Commerce
From a payments standpoint, travel is about as hostile as it gets. Cross-border transactions, where the traveler's home country differs from the property's, are the norm, and currencies and payment methods vary market by market. Ticket values are high enough to put fraud models on alert, and holiday seasons multiply volumes. Precisely because borders, currencies, ticket size, and seasonality compound, this is the domain where per-transaction AI adjustment shows up most clearly in the numbers.
Those conditions look a lot like what AI commerce is heading toward. Once AI agents become buyers, transactions can originate at any hour, from any country, at higher frequencies than human shopping. Agents compare sellers across borders, so cross-border and multi-currency combinations should multiply. How well payment and fraud systems built around human shoppers hold up against machine-speed transactions is an open question — and travel payments are serving as the dress rehearsal.
The supplier side deserves equal attention. An OTA (online travel agency) collects money from travelers and must then pay countless partners, from hotels to airlines. According to the announcement, Agoda's supplier payments are processed reliably at scale with a zero-downtime record. Handling money in and money out on one platform maps directly onto marketplace e-commerce, where operators carry payout obligations to their sellers.
Behind every booked trip is a digital moment that needs to work instantly and invisibly. As travel becomes increasingly digital, AI-driven optimisation and resilient payments systems are essential to connecting travellers, merchants, and partners worldwide.
When Agents Buy, Acceptance Rates Become Conversion Rates
Checkout.com itself frames payment optimization as groundwork for agentic commerce — commerce in which AI agents compare and purchase on the user's behalf. In a study published June 9, 2026, the company found a third of consumers expect at least 10% of their purchases to be AI-driven within a year, while AI agents currently touch just 3% of transactions. Seventy-two percent of merchants concede consumers will adopt agent-led shopping faster than the industry can prepare. The company is also among the 30-plus industry leaders first to support Mastercard's Agent Pay for Machines, the machine-to-machine payments platform announced in June 2026.
Agoda's trajectory points the same way. CEO Omri Morgenshtern said at the Skift Asia Forum in April 2026 that the company is rebuilding itself bottom-up with multiple AI agents deployed across the business, from security to the consumer experience, and connecting them into a unified experience. As AI takes over more of the journey from itinerary suggestions to booking execution, the success or failure of the final step — payment — decides how the whole experience is judged.
That is where the meaning of acceptance rates changes. A human shopper whose card is declined might try another card. In an agent-mediated transaction, a failed payment converts directly into a lost sale or a switch to a competing seller. Agents compare alternatives instantly, so merchants whose payments fail more often quietly drop out of the consideration set. This dynamic is already being discussed concretely, both as the argument that payment success rates become an agent's selection criterion and as the problem of false declines in agent-initiated transactions.
A line does need drawing, however. This partnership is not itself an implementation of agent payments, and there is no announcement that Agoda accepts transactions from AI agents or supports payment delegation. Acceptance-rate improvement and tokenization are the groundwork that comes before that step. Only when mechanisms for verifying who delegated payment, and within what limits, are layered on top of this foundation does agent-initiated payment become viable.
From Back-Office Cost to Transaction-Winning Infrastructure
Retail and e-commerce operators can take away three things. First, preparing for AI commerce is not just about adding a chat interface. Getting the payment back end in order — acceptance rates, tokenization, automatic failure recovery — feeds conversion more directly as agent-mediated traffic grows. Agoda's measured gains show this spending is a growth investment, not a defensive cost.
Second, get your payment data observable at the transaction level. If you cannot see how acceptance rates shift by market, card type, or time of day, you cannot verify whether AI optimization is working. When selecting a payment provider, it is worth weighing not just processing fees but how the provider improves acceptance rates and how it reports on them.
Third, stop treating money in and money out as separate problems. For marketplaces and businesses with many suppliers, the reliability of seller payouts underpins product supply and trust. Just as Agoda put buyer-side acceptance improvement and supplier payouts on the same platform, designing the full round trip of a transaction as one piece of infrastructure is the perspective that matters.
Conclusion
The Checkout.com–Agoda partnership is a reminder that the main battleground of AI commerce is not only the conversational front end. Payment rails that keep improving an unglamorous metric — the acceptance rate — with AI are what will carry transactions in an era when agents become buyers. Results demonstrated in travel, the hardest corner of payments with its cross-border, multi-currency, high-volume flows, amount to a leading indicator for retail and e-commerce businesses converging on the same conditions. The question the deal poses reaches well beyond booking businesses: before agent-mediated transactions arrive at scale, can you redesign payments from a back-office cost center into infrastructure that wins transactions?
