Smart Engines clients have recognized more than 39.4 million documents using our technologies, only over the last 10 months.
How to automatically extract document data using AI? Which industries were the first to adopt the automated customer identification technologies? How do companies benefit from such technologies? And last but most importantly, do companies hold the personal data of their clients?
Artificial intelligence technologies can automate a large number of routine activities that we spend our time on, thus accelerating a lot of processes. What instantly comes into mind when you think about that, is how often we need to provide our personal data to access various services on the everyday basis: to book a hotel, check in on a flight, open a bank account, activate a sim card, etc etc. At some point, when you do that often enough, you might probably find yourself having memorized all the necessary personal data and can fill out any form without the need of having the document in front of you. Many people actually do so to avoid unnecessary fuss, however if they do make a typing mistake, this could get them in trouble.
In situations like that, where the data from the document is manually entered by a bank, customs, or police officer, as well as notary assistant or a security officer — there is a risk of not only human-related errors, but also the possibility of a data leak. With all those reports about data leaks happening around the world on a regular basis for the last decade, there is something to think about.
How is the automated document capture market developing?
Many have realized how lucrative the machine vision business is, which is the reason that recognition technologies are gradually becoming a common phenomenon. The market of identity documents recognition is growing rapidly: analysts predict that in 5 years it will increase up to $12.8 billion.
There are a number of large companies around the world that are developing such technologies, and there are also quite a few small ones that use ready-made recognition modules and pack them in a software shell. In the long run, those that are following a scientific path and have unique capabilities to create proprietary technologies will have a more advantageous position on the market.
The Russian-based company Smart Engines — developer of document recognition technology called Smart IDReader, has been working with personal data processing for over a decade now, and ever since the first release of the software the company has consciously chosen to put its primarily focus on the security of data it’s working with. The use of artificial intelligence and computer vision has allowed automating the data capture process in such a way that all personal information is being processed directly on the end device — a smartphone or a terminal. The data does not get transmitted to third-party organizations for processing during the recognition; only RAM of the device is involved, and the system neither creates copies of the data, nor does it transfer it to any cloud/server.
Over the last 10 months of 2019 Smart Engines has counted over 39.4 million identity recognitions carried out in the applications where the machine vision module was integrated. Today, companies use recognition technologies to capture mostly the following document types: passports, driver’s licenses, evidence of national insurance numbers, vehicle registration certificates, and birth certificates. The data extraction is carried out in the video stream with the use of smartphone’s camera or from scans and photos of documents that users of various online services upload into mobile apps.
Quick document identification from the photo instead of manual forms filling
Smart IDReader technology is used by most leading banks, insurance companies, notaries and telecoms in Russia, as well as a number of national airlines, retailers, tax refund companies around the world (Aeroflot, Croatia Airlines, atc).
In 2018, Russian insurance company Alfa Strahovanie managed to recognize over 1.6 million documents uploaded by their clients to the mobile app, with the help of Smart IDReader.
Where is the automatic recognition used and how do companies benefit from it (10 examples from 2019)
Artificial intelligence for documents capture serves two main purposes: on the client side it helps to provide a convenient service which saves a lot of time, and on the company’s side, it takes the service to a whole new level of technological effectiveness and security. This helps companies demonstrate their superiority over their competitors.
Russia’s Big Three Telecom Companies
1. Beeline – registration of new clients in the online store
The convenience of quick delivery services increases the use of technology for automatic registration of subscribers. It simplifies the sim-cards registration and allows to quickly become a new Beeline telecom client. The operator’s web interface provides a fully automated process that includes recognition and authentication of a relevant identity document.
2. MegaFon – registration of SIM cards from tables across the network branches
The MegaFon mobile application equipped with document recognition technology allows a branch assistant to help a new client quickly register a SIM card using a tablet — just like they do at the registration or sales desks equipped with fixed equipment.
3. MTS – equipping SIM issuing terminals with ID verification
There are two video capture cameras on MTS self-service terminals. One is located inside the terminal and is responsible for passport data capture, and the other one is placed on the front panel — it identifies the owner of the document by comparing their face with the picture from the passport.
Recognition of passport data in a video stream is performed by Smart Engines algorithms. The technology finds the photo on the document and cuts it out to compare with the client’s photo taken with the front camera.
Gazprombank – loan applications processing
The technical solution designed for Gazprombank digital channels allows automating the lending process by recognizing the data from the applicant’s passport, as well as the income tax certificate. Smart IDReader has allowed bringing the number of automatically retrieved attributes to several hundred per borrower.
Samsung Forward Leasing – Mobile device Leasing Program
The IT platform designed for the distribution of goods on leasing terms, called Forward Leasing, implements automatic extraction of passport data to build a contract online in the mobile application or on the website. After the documents get recognized, the extracted information is used to autocomplete the application fields of the leasing agreement.
SKB-Bank – Recognition of passports, bank cards and QR codes in mobile applications
AI technologies are used to recognize QR codes, bank cards and passports in the SKB-Bank mobile app, as well as in the SKB-agent mobile application, designed for issuing credit cards by bank agents.
Aeroexpress – Buying tickets quickly and securely
Entering personal data from the ID document into a mobile app or a web service accounts for the majority of time spent on buying tickets online. In the updated version of the Aeroexpress application, which uses Smart Engines algorithms, tickets can be purchased in a matter of seconds.
“Soglasie” Insurance– buying new insurance policy (5 minutes) and extending an old one (1 minute)
The introduction of document recognition solution in the updated mobile application of the insurance company Soglasie has allowed halving the time spent on issuing or prolonging the insurance policies online.
INMED Clinic – registration of patients in „1C: Medicine. Hospital“ platform and making it more accurate
To open the medical history of the new patient, the hospital’s medical officer only needs to scan their passport using a flatbed scanner. Smart IDReader extracts the data from the scan and automatically transfers it to the medical information system.
Automated data capture is used here to avoid human error associated with pronouncing the data out loud by patient to the medical officer, who would then manually enter that data to the system. Furthermore, it makes the communication process between the officer and the patient more comfortable.
Avtodor – Toll Roads – reducing time on sales of transponders
Since June 2019, Avtodor’s technical solution used for making transponders sales, has been equipped with of Smart IDReader software and working in all sales and service centers of the company. Due to the automatic extraction of data from passports, the sales process has been accelerated to a few seconds.