Verification Checks
When individuals attempt to verify themselves, verification checksverification checks - A set of configurable criteria that are evaluated upon each verification attempt and determine whether it passes or fails. indicate something of note with the information that was submitted. They determine whether the verification attempt passes or fails and help you understand whether the submission is valid and meets your requirements, or if additional investigation is needed.
Configure and consume checks
There are a set of default checks that are required for the submission to be passed, while others are running and used for additional review but do not affect whether an attempt passes or fails. The defaults are determined based on the common needs across our customers, but we are happy to work with you to configure the defaults to meet your unique requirements. For more information about configuring requirements and the potential tradeoffs, please contact us.
Verification checks are returned in the Retrieve a Verification API response and are viewable in the DashboardDashboard - A hub for organizations to review inquiries and manage their integration with Persona. for each verification attempt.
Different checks are evaluated across different Verification Types, which are specified at the beginning of the check slug.
Check statuses
Check Status | Description |
---|---|
Passed | The check passed for the verification attempt. |
Failed | The check failed for the verification attempt. |
Not Applicable | The check was not evaluated for the verification attempt (e.g. |
Check types
There are two check types:
- User action required: The individual did not submit the at-minimum high-quality and unobstructed images needed for verification. For instance, if an individual submits a blurry image. Persona exposes these checks back to the individual in the inquiry flow as hints to help individuals submit better images.
- Fraud: The individual may be trying to verify with false or altered information. For instance, if the face portrait from the ID doesn't match the face from the selfie. There may be additional, noteworthy details about the submission that can be used to meet specific fraud requirements for different use cases. For instance, if the individual submitted an ID from a country that was not allowed on the templatetemplate - A configuration that defines the steps an individual must take to verify their identity..
Match results
For Database Verifications, each piece of information submitted like name or birthdate is matched against the record returned from the source database, and Persona returns three possible match results for each piece of information as separate checks:
- None: The information did not match.
- Partial: The property matched, but did not match exactly. Persona utilizes algorithmic processes (i.e. fuzzy matching) to determine the mathematical similarity between two pieces of information. This allows for the passing of near matches, which avoids unnecessary failures if the individual inputs an accidental typo or, in the case of name, a nickname (e.g. "Mike" instead of "Mikey"). For example, a date of birth of 12/15/1990 would partially against a birthdate of 12/14/1990.
- Full: The information matched exactly.
For more information about configuring the minimum match requirement for each piece of information, please contact us.
Check list
Slug | Name | Type | Description | Failure Reason(s) |
---|---|---|---|---|
| Account comparison | Fraud | Compare if ID details like name or birthdate are inconsistent. |
|
| Portrait age | Fraud | Detect if the age listed on the ID is different than the age estimated from the face portrait. |
|
| Barcode | User action required | Detect if there is a barcode on the ID. |
|
| Barcode inconsistency | Fraud | Detect inconsistencies in the barcode on the ID. |
|
| Blur | User action required | Detect if the submission is blurry. |
|
| Color | Fraud | Detect if the colors on the ID are different than the expected colors. |
|
| Compromised submission | Fraud | Detect if the submission can be found from a publicly available source (e.g. the internet). |
|
| Allowed country | User action required | Detect if the country of the ID is not allowed based on template requirements. |
|
| Allowed ID type | User action required | Detect if the ID type submitted is not allowed based on template requirements. |
|
| Double side | User action required | Detect if one side of the ID is submitted as both front and back. |
|
| Electronic replica | Fraud | Detect if the submission is a replica of an image presented on another electronic device or screen. |
|
| Government Entity ID | User action required | Detect if there is a government ID in the submission. |
|
| Expiration | User action required | Detect if the ID is expired. |
|
| Extracted properties | User action required | Detect if required ID details like name or birthdate are extracted. |
|
| Extraction inconsistency | Fraud | Detect if the details extracted from the front of the ID are different than the details extracted from the barcode. |
|
| Glare | User action required | Detect if the submission has glare. |
|
| Handwriting | Fraud | Detect if the handwriting is found on the ID. |
|
| Inconsistent repeat | Fraud | Detect if either the ID details or face portrait match that of a previously submitted ID. |
|
| Inquiry comparison | Fraud | Compare if ID details like name or birthdate are inconsistent between different submission attempts. |
|
| MRZ | User action required | Detect if there is encoded, machine-readable text on the ID. |
|
| PO box | User action required | Detect if the address is a PO box. | |
| Portrait clarity | User action required | Detect if the face portrait is clear and in focus. |
|
| Portrait | User action required | Detect if there is a face portrait in the submission. |
|
| Public figure | Fraud | Detect if the face portrait matches that of a known public figure. |
|
| Repeat | Fraud | Detect if the ID details and face portrait match that of a previously submitted ID on a different account. |
|
| Selfie-to-ID comparison | Fraud | Compare if the face portrait on the ID and the selfie are different faces. |
|
| Image tampering | Fraud | Detect if the submission is image tampered or manipulated by an editor. |
|
| Processable submission | User action required | Detect if the submission is unprocessable. |
|
| Valid dates | Fraud | Detect if the dates on the ID are in violation of issuing authority rules. |
|
| Account comparison | Fraud | Compare if the selfie is inconsistent with the selfie of the account. |
|
| Age comparison | Fraud | Detect if the perceived age from the selfie is different from the reference age. | |
| Glare | User action required | Detect if the submission has glare. |
|
| Glasses | User action required | Detect if glasses are present. |
|
| Selfie-to-ID comparison | Fraud | Compare if the selfie and the face portrait in the ID are different faces. |
|
| Multiple faces | Fraud | Detect if the selfies are different faces across poses. |
|
| Pose position | Fraud | Detect if the face in the selfies are correctly positioned across all poses. |
|
| Public figure | Fraud | Detect if the selfie submission matches a known public figure. | |
| Pose repeat | Fraud | Detect if the selfies are exact repeats across poses. |
|
| Suspicious entity | Fraud | Detect if the selfie submission is that of a government ID. |
|
| Address | Fraud | Detect if the address matches any databases. |
|
| Deliverable address | Fraud | Detect if the address is deliverable. | |
| Birthdate | Fraud | Detect if the birthdate matches any databases. |
|
| Alive | User action required | Detect if person is living. | |
| Name | Fraud | Detect if the name matches any databases. |
|
| Phone number | Fraud | Detect if the phone number matches any databases. |
|
| PO box | User action required | Detect if the address is a PO box. | |
| SSN | Fraud | Detect if the SSN matches any databases. |
|
| Compromised submission | Fraud | Detect if the submission can be found from a publicly available source (e.g. the internet). |
|
| Image tampering | Fraud | Detect if the submission is image tampered or manipulated by an editor. |
|
| Extracted properties | User action required | Detect if the details extracted from the document match patterns. |
|
| Recency | User action required | Detect if the document is recent. |
|
| Type | User action required | Detect if the document one of the expected types. |
Legacy Checks
If you are using our legacy verification checks, known as confidence reasons, please refer to the Confidence Reason(s) column in the table below to understand how they map to the new verification checks.
Check Slug | Confidence Reason Slug |
---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
|
|
|
|
| |
|
|
|
|
|
|
|
|
| |
|
|
|
|
|
|
| |
|
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
| |
| |
| |
| |
| |
| |
|
Updated 18 days ago