All Products
Search
Document Center

Fraud Detection:What is Fraud Detection?

Last Updated:Oct 14, 2026

Fraud Detection is a smart, lightweight, and mature risk control service from Alibaba Cloud, built on more than 10 years of risk control experience from Alibaba Cloud, Taobao, and Alipay.

Fraud Detection launched in 2018 and offers three types of risk control capabilities: a model as a service (MaaS) option for fast API-based integration, a software as a service (SaaS) decision engine for teams that want to build and manage custom risk control workflows, and on-device security components (SDK and JavaScript) for client-side attack detection.

Business risks it addresses

As internet services expand into more industries, fraud attacks have become common across four scenarios:

Scenario Common risks
Account registration Bulk account creation using illegal or false information
Logon Account takeover, fund theft, unauthorized profile changes
Marketing Bots and fake accounts exploiting reward points redemption, coupons, and flash sales
Financial services Fraud in consumer loans, business loans, and auto finance

Solutions

Fraud Detection provides three types of risk control capabilities. The table below summarizes when to use each:

Solution Best for Integration effort
MaaS-based risk identification model Teams that want immediate fraud detection via API, no custom model training Low — call the API directly
SaaS-based decision engine Teams that need custom risk control workflows and full operational control Medium — configure rules and modes in a visual platform
On-device security components Applications that need client-side attack and device anomaly detection Medium — integrate SDK or JavaScript component

MaaS-based risk identification model

A pre-trained decision-making AI model with pre-defined rich data features, ready to use via API. Covers common scenarios including account registration, logon, marketing, and credit applications. It provides ease of use, stable performance, and all-in-one processing — call the Fraud Detection API to get risk decisions immediately.

SaaS-based decision engine

A visual risk management platform where your team builds and operates custom risk control workflows. Use it alongside Alibaba Cloud's pre-defined risk control modes, or configure your own rules. For details, see Overview of the decision engine.

On-device security components

SDKs for Android and iOS, plus a JavaScript component, that run on the client side to detect attacks, device anomalies, and emulators. Helps identify malware and block malicious devices before they reach your backend. Building on these components, the AI/LLM Edition of Device Fraud Detection adds semantic analysis based on large language models (LLMs). It detects device group control, device farms, automation scripts, device spoofing, emulators, and cloud phones. It can also identify previously unknown attack patterns. For details, see Service event parameters and response parameters for Device Fraud Detection.

What's next