SentiLink's models are trained on fraud insights, not performance.
As a result, our synthetic scores are renowned for precision and best-in-class
false positive rates.
Detect synthetic identities with
Start with multiple data sources
There are over 300M consumers in the U.S. Data on when they applied for credit, what credentials they used when applying, married with a large collection of data elements associated with those consumers is the foundation. This, combined with the unique consortium data shared across our partners is the starting point for evaluating identities.
Fuse the data with human intelligence
SentiLink brings intelligence, people and data together to empower one another. Intelligence comes from our Risk Operations team who applies deep knowledge of known synthetic identities. Nuances and edge cases are identified that technology can’t catch.
Define model targets based on fraud insights
The clean, consistent, and clearly defined fraud labels applied by our Risk Operations team to subtle differences in synthetic identities differentiates our solutions. These insights based on what fraud looks like is what trains our model to catch significantly more synthetic fraud with a much lower false positive rate.
Three Synthetic Scores
Composite SyntheticComposite synthetic fraud score that indicates whether an identity is likely synthetic.
First Party SyntheticDetects applicants using their true name and date of birth, but an SSN that doesn’t belong to them. They are manipulating their true identity.
Third Party SyntheticIdentifies applicants using a name, date of birth and SSN combination that is completely fabricated.
Synthetic Score Delivery
Integrate easily to our single endpoint API and get scores in milliseconds.
Batch can occur with an SSH File Transfer Protocol (SFTP) or upload data manually to the SentiLink dashboard.
Intelligent UI that surfaces the most important attributes for investigation.