Real-Time Fraud Detection

Score API

A powerful fraud risk assessment API that delivers reputation scoring based on phone number intelligence, traffic patterns, machine learning and a global data consortium.

ASSESS FRAUD RISK WITH

PHONE NUMBER INTELLIGENCE & MACHINE LEARNING

At an alarming rate, bad actors are creating online and mobile application accounts that result in spam, phishing attacks, promo abuse and other costly fraud. Through the registration of fake accounts, fraudsters are able to attack legitimate users and damage a brand’s value, revenue and growth. Effectively identifying and blocking these harmful users at account registration, while streamlining the process for authentic and valuable users, has become critical.

0.1

USER ENTERS PHONE NUMBER

0.2

INVOKE MACHINE LEARNING

0.3

EVALUATE PHONE DATA ATTRIBUTES

0.4

REVIEW PHONE NUMBER VELOCITY

0.5

DETECT PHONE NUMBER TRAFFIC PATTERNS

0.6

CHECK FOR HISTORY OF FRAUD

0.7

SCORE THE NUMBER & EVALUATE RISK

PHONE NUMBER DATA & ANALYTICS

Phone number data attributes including phone type, telecom carrier, account and device ID and IP address are evaluated to identify potential fraud risk.

GLOBAL FRAUD DATA CONSORTIUM

Score leverages TeleBureau™, TeleSign’s global database of customer-contributed phone number reputation information, to help detect and identify known fraud faster. This also further calibrates scoring algorithms based on specific uses cases.

TRAFFIC PATTERN RECOGNITION

Anomalous traffic behavior patterns may raise a red flag. For example, if passcode requests are received in five different languages from the same number in the same week, it may be a sign that a phone is being shared, and the risk score will increase accordingly.

MACHINE LEARNING & RULES-BASED SYSTEM

Machine learning trains a customized model to uncover hidden insights in data to populate a range of fraud predictors. A rules-based system uses a historical fraud research knowledge base to evaluate risk. Customized machine learning models further increase the effectiveness and accuracy of Score’s fraud detection capabilities.

USAGE VELOCITY

If a particular number or range of numbers is seen repeatedly on one or more Web service within a relatively short timeframe, the number may be flagged as suspicious, increasing risk score.

ACTIONABLE RISK ASSESSMENT

A data-driven risk assessment score is delivered that helps determine the appropriate action of whether to allow, block, or flag a registration or transaction.

Please note, not all product features are available in all markets. Contact sales for specific country coverage.

Benefits

GROW USER BASE RESPONSIBLY

Streamline the account registration process, increase conversions and securely grow ecosystem of verified and valuable users.

IDENTIFY FAKE & SUSPICIOUS USERS

Mitigate more fraudulent activity and validate that end-users are who they say they are.

PROTECT BRAND REPUTATION AND VALUE

Reduce the negative impacts of fraud while creating a more authentic and valuable user base.

We needed more data on our customers. We match our data with TeleSign’s data and it gives us greater assurance that our customers are actually who they say they are, which reduces fraud.

JONATHAN EPPERS, CEO

How It Works

1
User enters phone number – During new account creation/registration, a user enters their phone number and it is passed to TeleSign using the Score API.
2
Invoke machine learning – Score uses machine learning to score phone numbers. Machine learning is an algorithm that uses historical indicators to uncover hidden insights and predict future events.
3
Evaluate phone data attributes – Information
associated with the phone number--including phone type (mobile, landline, VOIP, burner), telecom carrier, subscriber’s contact details, status of the phone, geographic location, and country code--helps identify any potential red flags.
4
Review phone number velocity – Knowledge about usage activity associated with the number, like whether it has been seen repeatedly on one or more Web services within a relatively short timeframe, influences risk assessment.
5
Detect phone number traffic patterns – Anomalous behavior patterns, like frequent account activity from various geographic locations, can be a sign that the phone number is being shared by multiple accounts for fraudulent purposes.
6
Check for history of fraud – Running the phone number through TeleBureau, TeleSign’s global phone number reputation data consortium, enables Score to identify any cases of the number having been associated with fraud.
7
Deliver a Score (risk level/recommendation) –
Once the above steps are complete (in a matter of milliseconds), Score returns a risk assessment back to the Web or mobile application. The score ranges from 0 to 1,000 and helps inform the decision to block, flag or allow a user to proceed in the account creation process.
8
Challenge user/prevent potential fraud – If the resulting risk assessment score determines that a verification/challenge should occur before proceeding, TeleSign can then verify the user via a simple SMS or voice-based one-time passcode.

 

Score & Machine Learning

At TeleSign, our Data Science team is using machine learning to power our Score API, which can now predict and prevent end-user account fraud with increased accuracy, in real time. See how!

Get Started With TeleSign

Sign UP Now

Need some help first? Ask an expert.