Inside Airwallex Optimize 360: How AI Powers Full‑Stack Payment Optimization

What is Optimize 360

Optimize 360 is an AI‑driven intelligent acquisition engine embedded in Airwallex’s global payment stack. It applies real‑time intelligence to fraud protection, network acceptance, and transaction recovery, aiming to balance payment success rate, cost, fraud risk, and long‑term continuity.

The engine achieves global optimality across the payment chain, improving overall acquisition performance and merchant revenue efficiency.

Four Key Design Principles

1) Embedded Integration – Instead of layering an external tool, the engine is built directly into the payment flow, allowing routing, retries, risk control, 3DS, tokenization, and result handling to cooperate within a single process.

2) Real‑time Adaptation – Static rules lag behind dynamic payment environments. The model continuously ingests real‑time transaction outcomes, issuer behavior, consumer trends, and risk signals to adjust decisions on the fly.

3) Global Insight with Local Tuning – The model learns from data in over 35 markets, delivering deeper judgment than a single‑market model because it sees more examples (“the more you see, the better you predict”).

4) Balancing Conversion, Cost, and Risk – By placing success rate, cost, and fraud risk into a single real‑time decision framework, Optimize 360 finds the optimal trade‑off for each transaction.

Setup: Configuration Optimization

The Setup module tailors the payment entry to local user expectations, reducing decision friction at checkout. Data shows 43% of cart abandonments stem from unavailable payment methods.

Match market and device to the most acceptable payment method order.

Automatically adapt desktop and mobile checkout flows.

Optimize system configuration to avoid failures caused by missing fields, currency settings, or incomplete parameters.

Support automatic currency conversion so users see local prices.

Incorporate local payment methods to reduce user distrust.

The underlying logic is to lower user decision difficulty, making payment choices more intuitive.

Maximize: Success Rate and Revenue Recovery

This layer focuses on routing transactions through the most successful path. Different issuers prefer different routes; selecting the right path can turn a potentially successful transaction into a success.

Payment failures are classified as hard rejects (e.g., stolen card, frozen account) or soft rejects (temporary issuer issues, network glitches). Maximize identifies soft rejects and resubmits them using smarter routing and retry strategies.

Intelligent Routing – Uses issuer history, real‑time success rates, and cost to pick the optimal path. Example: If a Visa issuer historically authorizes 5% higher on route A than route B, the engine prefers A.

Intelligent Retry – First determines if a reject is soft, then decides whether, when, and how to retry, avoiding blind repeated attempts that could lower success rates or trigger issuer anti‑fraud rules.

The model’s advantage lies in its scale: data from 35+ markets yields deeper understanding of reject codes, enabling more effective recovery.

Secure: Risk Prevention

Secure aims to detect fraud while minimizing false positives. Over‑strict rules increase fraud but hurt genuine users; overly lax rules increase chargebacks.

AI‑driven real‑time risk scoring combines payment behavior, device, network, and transaction history. The system adapts risk rules continuously and applies smart 3DS challenges only when risk warrants it.

AI‑driven fraud detection and real‑time risk scoring.

Signal aggregation from behavior, device, network, and history.

Self‑adapting risk rules for evolving threat landscapes.

Smart 3DS: reduce friction for low‑risk flows, enforce stronger verification for high‑risk or regulated cases.

Compliance support for SCA, PSD2, etc., with market‑specific applicability.

The core principle: protect high‑risk transactions while avoiding harm to legitimate users.

Sustain: Continuous Stability

Sustain addresses long‑term operation, especially for subscriptions, memberships, recurring billing, and saved‑card scenarios. Issues like card expiry, replacement, or re‑issuance can cause silent churn.

Network Tokenization : Replace part of the card number with issuer‑issued tokens to improve stability and possibly reduce fees.

Credential Lifecycle Management : Handle card replacements and expirations.

Account Information Update : Auto‑update saved card data where supported.

Anomaly Detection & Early Warning : Monitor payment performance fluctuations and help pinpoint causes.

These capabilities are combined into a single module that many peers can reference.

Market Performance Data

Public cases show Shein’s payment success rate improved up to 3% in key European and APAC markets, while Cider saw a 1% lift over industry averages. These figures represent best‑case scenarios; actual impact varies by business type, market coverage, and transaction mix.

Different Business Scenarios

Cross‑border e‑commerce, DTC, global retail : Focus on Setup, Maximize, Secure to address local payment method gaps, currency mismatches, issuer rejections, and checkout abandonment.

Subscriptions, memberships, SaaS : Prioritize Sustain with intelligent retry and network tokenization to rescue recurring payments.

Games and digital content : Emphasize Secure and smart 3DS to balance high‑frequency fraud risk with conversion.

Global platforms with multiple markets : Leverage intelligent routing, message optimization, and anomaly monitoring for complex, multi‑path operations.

Final Thoughts

Adding more payment channels alone hits a ceiling; complexity grows while global efficiency may not improve. Optimize 360 adds an AI layer on top of existing flows, enabling each transaction to be processed optimally in real time.

The enduring logic of payment success optimization is multi‑objective balancing, real‑time adaptation, and global collaboration.