en

Language

02 Jun 20237 min readin ID verification & biometrics

The What and The How of Automated Document Verification

Jan Stepnov

Identity Verification Expert, Regula

Document verification has become a legal requirement for a variety of businesses, including highly regulated ones like Banking, Aviation, Gambling, and Fintech. However, time-consuming manual checks and data entry slow down the process and are prone to human error. This also negatively affects the user experience, since customers are used to accessing platforms and services in seconds. 

There are ways to reduce friction and streamline the document verification process by adding automation components to the journey while staying customer-centric and compliant with all regulations. 

This article breaks down the subject and provides you with a range of automation use cases in different industries.

What is document verification?

Document verification is the process of verifying the validity and authenticity of a passport, driver's license, ID card, and other ID documents. It involves examining the document to ensure that it is not counterfeit, tampered with, obsolete, or expired. 

The thoroughness of the process depends on the specific situation in which it is required. There are three stages of this process in terms of the depth of verification:

  • Checking data in the visual inspection zone of the document, including the main and secondary photo, personal data, ID number, the issue and expiry date, and the presence of an MRZ or barcode, hologram, and any other security elements, depending on the document type. Usually, it's part of a formal process, for example, when an insurance agent fills in the details for a policy, or a casino administrator asks a young-looking guest to show their ID card. 

  • Checking for any alterations in the document, for example, lexical analysis for identifying inconsistencies caused by data change in the MRZ or electronic RFID chip tampering. Such a check can be a part of a remote or on-site customer onboarding process, for example, when an individual asks for a loan at a bank. 

  • Checking for document authenticity, e.g., barcode format or embedded photo type. Is it genuine and issued by the authorized body, or forged? In case of doubts, questioned documents can be sent for a deeper examination by forensic experts.   

In the two latter scenarios, particular characteristics like stamps, watermarks, lenticular printing elements, fonts, and other security features are investigated with the use of hardware devices and dedicated software.

Where is document verification most applicable?

First and foremost, verifying documents is mandatory for governmental institutions, including border control services, immigration departments, unemployment agencies, and other social organizations responsible for paying benefits and allowances. The goal is to ensure that only authorized individuals are granted access to their services, facilities, or territories.

Second, document verification is a prerequisite for industries tied to KYC compliance requirements. They include: 

  • Banking, Fintech, and any other finance organizations

  • Telecom companies

  • Travel and Hospitality

  • iGaming & Gambling businesses

  • Insurance providers

  • Vehicle sharing companies

  • e-Learning platforms, etc.

Last but not least, document verification is essential for organizations handling sensitive personal identifiable information. A good example is healthcare institutions.

Manual vs. automated document verification: Making the right choice

There are two ways to verify documents: manually, or by employing automation solutions.

In the first scenario, it’s an employee's task to ensure the ID is legitimate. 

The second scenario leverages specialized software and purpose-built devices. Note that automation in this case doesn't eliminate a human from the loop, but does most of the heavy lifting. For instance, identity verification software built into a bank’s mobile application checks a customer's ID in seconds. However, when one or more checks fail, the software can notify a manager, who can provide additional manual verification of the document image.

Both scenarios are possible for on-site and online document verification. 

The decision on what path to take depends on specific use cases and what's at stake for an organization or business.

A graph shows the relationship between UX and the cost of human error to reveal the need for automated document verification.

This graph shows the need for automated document verification considering UX and the cost of human error

Passport control is usually a bottleneck at state borders, with many frustrated travelers waiting in lines. Moreover, an illegal border crossing with fake documents can probably have harmful consequences, from espionage to terrorism. That’s why automated document verification is a commonly used practice at land and air border crossing checkpoints. The personnel are equipped with specialized devices like passport readers for instant checks, and video spectral comparators for more detailed examination by forensic experts. 

Carriers also have other reasons for not using manual document verification. Automation helps them reduce queueing time, improve customer satisfaction, and cut general labor costs. These reasons nudged Wizz Air to incorporate self-check-in into the customer journey.

Some businesses from regulated industries like Banking still use manual checks such as video interviews when onboarding new customers. However, adopting this practice may be a costly and time-consuming process for compliance, especially when the level of competition is increasing. That prompted UBS, the world's largest private bank, to replace real-time interviews with automated document checks. 

Age-restricted goods and services is another area where document verification is a must. While manual checks may be commonplace for corner shops, big alcohol and tobacco retailers may consider age verification automation as an alternative. The potential benefit will depend on sales volume and expected ROI.

Manual verification of documents may seem like a more affordable alternative for some companies. However, the labor costs of employees, even if they are outsourced specialists, usually always turn out to be higher than the cost of licenses to run automated verification.

Additional risks are associated with the more sophisticated document forgery schemes fraudsters use now. Experienced specialists can still spot mismatches in physical documents by manually checking them, but even experts use specialized devices for such purposes; machines can detect the tiniest details on the document, which are beyond the capabilities of human vision.

Document scans are harder to detect with the naked eye. Deepfakes generated by AI tools can now easily simulate a range of photos and videos. And they are scarily plausible. 

That's why you should consider these risks when deciding on a solution that suits your case.

How does automated document verification work?

Document verification using automation tools includes four main steps:

A diagram showing the stages of the automated document verification process

Key steps of the automated document verification process

Let’s have a look at each of them:

 

Step 1. Document capture

At this stage, a device or software application should recognize the submitted document and assess its quality in terms of text and image clarity, size, and form. 

This step also impacts overall UX from the very beginning. The solution must guide the user through the image capture process to get the best shot. The better the image, the more accurate the verification results. 

When choosing a solution, see how many users are asked to retake an image. The higher the number, the worse the UX is.

 

Step 2. Document assessment

During this step, the solution accesses the document template database for data validation. The bigger it is, the wider the possible use cases, and the more document types and countries it can read and verify. 

Document verification tools assess the presence of security elements such as MRZ, barcode, and RFID chip. They can also automatically identify the type of the submitted document, for example, a Swiss passport or a Utah driver’s license.

Regula Document Reader SDK relies on Regula’s extensive document template database, the largest in the world, compiled and regularly updated by the company's experts. It’s not just composed of driver’s licenses and passports—the solution verifies many other types of identity documents, as you may also want to verify ID cards or work permits, for instance.   

 

Step 3. Data extraction

Once the software gets the data, it’s time to read and cross-validate it. Employing OCR technology, the solution extracts and reads the data from the Visual Inspection Zone (VIZ), MRZ, barcode, and electronic chip. It is important to consider how many languages a tool supports, especially if you need to process identity documents issued in countries with non-Latin alphabets.

 

Step 4. Document verification

Once the data is extracted from the various elements, it also should be matched against the information from the VIZ, MRZ, barcodes, and RFID to ensure data consistency and authenticity. To streamline the verification process, Regula Document Reader SDK automatically conducts a set of necessary thorough checks, depending on the type of identity document being examined: for example, document number format analysis, correctness and validity of dates, static text validation, deprecated blank detection, barcode and MRZ format checks, and server-side verification of the embedded RFID chip. It also checks data presence or absence in the mandatory fields on the data page. Moreover, in a hardware scenario, authentication can be additionally performed in different lighting modes: IR, UV, coaxial light, etc. 

Note that different vendors incorporate a varying set of checks into the document verification flow. As a customer, you should clarify what parameters are verified in each case. 

Biometric checks can be an optional part of the verification process. Usually, a customer needs to take a selfie via a mobile app to confirm that the live picture and the photo in the document depict the same person. When a biometric check is included in the process, it turns document verification into complete identity verification.

Can you trust automated document verification?

In a nutshell, yes, you can. There are at least four reasons:

 

1. Reducing human error to zero

The manual verification process may involve personnel, such as casino administrators, who are not forensic experts. As a result, the quality of checks may be impaired due to a lack of knowledge of what to do and how to do it. 

Automated solutions are able to fill this gap. They investigate each tiny detail and detect the slightest discrepancy, such as a relocated or resized element. For example, Regula Document Reader SDK can identify inappropriate font size and placement on a passport, driver’s license, or other ID document.

 

2. Proven verification process

The process behind automated document verification consists of a precise sequence of actions backed by best practices in information security. 

It is based on guidelines that describe each step of the process, including, for example, the logging format (the way machine-readable data should be structured, encoded, and stored). 

There are two core standards: Technical Guideline BSI TR-03135 from the German Federal Office for Information Security (BSI) and ICAO Doc 9303 about machine-readable travel documents. 

Both guides are widely used across different industries, including non-regulated ones, as the basis of document verification procedures. Compliance with these standards is a must for identity verification (IDV) solution vendors as well. 

 

3. Extensive knowledge base

Hardware and software solutions refer to extensive document template databases as a data source and “know” more about identity documents, including foreign ones, than your employees. 

A regular bank clerk can hardly imagine what an Albanian ID card or an Uruguayan driver’s license should look like. But advanced software such as Regula Document Reader SDK knows for sure. Thanks to the database, which includes over 12,500 document templates, the solution can verify ID cards and passports from 247 countries and territories.  

 

4. In-depth analysis and fraud detection

Automated document verification usually includes more steps than a manual check. Using specialized devices and software allows you to conduct a deeper analysis of the captured data, like MRZ reading and document liveness detection. Also, an automated verification session may include extended authenticity checks to tick all the security boxes.

The fact that the implementation of automated document verification is now a legal requirement for some industries, such as Aviation and Border Control, is another confirmation of its reliability.

At the end of the day, a document flagged as questioned by an automated solution can be submitted for additional manual verification. So, you can always perform double checks involving in-house or external experts if necessary.

Make faster decisions with reliable document verification software

Automated document verification is a game-changer for businesses looking to streamline their processes and stay compliant with industry regulations. By automating this task, companies can save time and money, while ensuring the accuracy and authenticity of document checking. Whether you're in Finance, Aviation, Insurance, or any other industry that requires document verification, automation is the way to go. 

Regula's solutions for automated document verification can help you meet all requirements, while making the process simpler and smoother for your customers. Embrace automation and enjoy the benefits it brings by checking out a free trial or requesting a demo.

Regula Document Reader SDK

Automate identity document reading
and authentication.

Learn more

On our website, we use cookies to collect technical information. In particular, we process the IP address of your location to personalize the content of the site

Cookie Policy rules