Pierre Demarche, VP Product and Marketing, TeleSign
Session is divided into 3 parts, based on use cases:
Registration fraud. Identifying fake users, preventing bulk account creation, etc.
ATO (Account TakeOver). Preventing ATO fraud through phone hijacking attacks like SIM Swap, Porting attacks, call forwarding, etc.
IRSF (International Revenue Sharing Fraud) attacks on enterprises. Registration to SIP trunking.
5. STOP
FRAUDSTERS
FROM
CREATING
ACCOUNTS
5
Websites/apps have huge problems with
fake accounts
Using bots, fraudsters have created millions of social
media, email, and other accounts
Fraudsters create fake accounts to:
• Spam
• Phish
• Post fake reviews
• Increase number of followers
• Etc.
Phone Number based verification adds
a layer of protection
• However, fraudsters continue to find new ways to
creat bulk fake accounts
18. COMMUNICATION DATA CLASSIFICATION
Strong co-relation with online frauds
MOBILE
NUMBER
New Dimensions
Behavior
No of outgoing/incoming calls/messages
Velocity
Diversity of called numbers, countries, operators
Diversity of receiving numbers, countries, operators
Duration of calls
Tenure of numbers
Length of messages
No of countries roamed
Mobile data usage
Failed call attempts
Failed messages
Failed signaling
Type of carriers
- Tier 1,2,3,4 VOIP
Risky destinations
Premium calls
Spam content
Calling Number
Called Number Roaming Number
Sender SMS Receiver SMS
Customer Supplier
Origination Country/Operator Destination Country/Operator
Trillions of Global
communication
transactions
Tokenized Privacy
Compliant Platform
19. CLASSIFICATION OF PHONE NUMBERS BASED ON BEHAVIOR
Velocity histogram (bar = day) Velocity histogram (bar =
hour)
Call duration histogram (bar = sec)
Application
Traffic
Robocalls
23. DATA PLATFORM
ENABLING COMMUNICATIONS &
PROTECTING ONLINE EXPERIENCES
Lead
Scoring
Registration
Risk
Score
Identity
Verification
Score
Credit Risk
Score
Account Take
Over Score
Payment
Transaction Risk
Score
Communication
Context Score
GLOBAL COMMUNICATION PLATFORM
As we all know website and apps DO HAVE a huge problem with the large number of fake accounts already created
Highligh t how trust is important for brans engaging with their end users, outside of compliance considerations, brnds need to have trust with their end user transacting on their latforms
Balanve to strike between providing less friction to good users for growth and enough security to avoid fraudulent account cretaion
Intelligence, security & people
Add notes from Sean
KE team
Voip detection
Thickness is different, indicator of quality of calls
Call duration indicate that ebhviour od end user is deifferet
Legit is p2p calling. Middle is otp and last is robotcalling
It is not a one time event authentication is history of behaviuor
Good users
Call behavior gives a level of trust
Data used to solve different outcomes
Lead scoring: capcble to understand behavior
Registration: most of customer cobined with SMS 2fa
Identity verification: account creation or payment to validate end user data provided identiy
Reidt risk scoring: not a customer score – dg with developing country lending market for micro loans