One of the leading mortgage brokers in Russia. The service employs about 400 people, and the volume of loans issued exceeds 5 trillion rubles a year.

To apply for a loan, customers must provide various documents. Including passports, INN, SNILS, several payslips, a certificate from work and much more. In general, the service receives more than 200 pages per minute for processing. Each document is manually processed by a specialist and transmitted to the system, which examines customer profiles in automatic mode.

Along with the growth in the volume of loan applications, the burden on specialists providing processing and approval of loans has increased. The client was faced with the problem of a “narrow neck”, without a technological leap it was impossible to maintain the planned speed of processing applications for mortgage approval. In the mortgage lending niche, the speed of customer approval is a key efficiency parameter, therefore it is very important to ensure seamless and instant data processing. The task was set to implement functionality that allows you to approve mortgage applications online as part of a session on the website.

Specialists manually entered heterogeneous data from the provided documents into a single database. Given the wide variety of documents with different layouts and structures, there could be no question of a standardized approach.

Before the introduction of IDP, processing of one document in manual mode took from 1.5 to 5 minutes.

We have implemented an intelligent document recognition system into the client's business process, which has reduced the required time resource for data recognition and structuring by 50 times.

The system successfully works with the following document formats:

• Passport of the Russian Federation - 3-5 sec / page
• SNILS - 3-5 sec / page
• Reconciliation Report - 10 seconds / page
• Acceptance Certificate - 10 seconds / page
• Invoice - 10 sec / page
• Contract - 10 sec / page
• Invoice - 10 sec / page

After the introduction of Visionist IDP, processing of one document takes no more than 10 seconds without loss of accuracy

Documents related to personal data, such as passports, TIN and SNILS are recognized by default. We have partially trained the system to recognize and read unique documents, for this it is enough to provide 100 documents of the same type for recognition. Immediately after implementation, the system achieved almost the same results in data accuracy as the reference values from people at a much higher speed.

1. The user uploads documents to the site
2. IDP displays all the data in 10 seconds
3. The user verifies them and sends them to the operator
4. The operator starts manual data verification

After implementation, the platform became available to all employees of the department through a convenient web application. Now they can initiate the reading process by uploading documents to the interface in manual or automatic mode.

Reading, classification, extraction, standardization are performed automatically. Using the graphical interface, employees can visually perform cross-checks and make the necessary adjustments. After that, structured information is automatically transmitted for credit analysis, and a personal dossier is created for each client based on the processed data.

• Increased throughput of the sales department;
• Reducing the processing time of the mortgage application;
• Digitalization and automation of the business process;
• Potential for business scaling;
• Standardization of the customer database.

After the introduction of Visionist IDP, the client received an application for automated processing of key data for this business — applications for a mortgage loan. The solution is flexible, which makes it easy to integrate document analysis into existing processes.

A federal food supplier company has faced a snowball of unstructured electronic documents. The wholesale company is in the middle of the supply chain and works with a large number of b2b counterparties. At the time of the beginning of cooperation, the company processed about 800 thousand documents per month.

The volume of unstructured documents is growing exponentially, which leads to a constant accumulation of raw information. There is too much data for a simple upload of information to SAP, the situation is complicated by obtaining documents in heterogeneous formats, manual entry of such documents can be very long and difficult.

The client creates a digital archive with all incoming and outgoing documents. All documents must be entered into 1C manually, processing of one document took from 3 minutes. Visionist was tasked with building a large-scale electronic archive without actual storage of source files. All data should be stored in a digital database followed by a simple catalog search.

Before the introduction of IDP, processing of one document in manual mode took from 2 minutes. The flow exceeds 800 thousand documents per month

We have implemented an intelligent document recognition system into the client's business process, which has reduced the required time resource for data recognition and structuring by 50 times. To start data processing, it is enough to upload documents to a digital archive.

After the introduction of Visionist IDP, processing of one document takes no more than 10 seconds without loss of accuracy

• Source of documents: SAP, mail, electronic document management with counterparties;
• Visionist starts data recognition and processing: all positions and total must converge;
• If the algorithm detects a discrepancy, the document is sent for reconciliation and verification;
• After processing, the data is transferred to SAP, the source documents themselves are stored in the Visionist cloud;
• You can work with the received data in SAP or the Visionist interface.

After implementation, the platform became available to all employees of the department through a convenient web application. Now they can initiate the reading process by uploading documents to the interface in manual or automatic mode.

Reading, classification, extraction, standardization are performed automatically. Using the graphical interface, employees can visually perform cross-checks and make the necessary adjustments. After that, structured information is automatically transmitted for credit analysis, and a personal dossier is created for each client based on the processed data.

After the introduction of Visionist UDP, the client received a tool for building a digital archive without limiting the maximum volume. The solution is flexible, which makes it easy to integrate document analysis into existing processes.

The regional telecom services operator works with a large number of B2B counterparties. Such a business model creates a large-scale document flow. The client's back office is forced to process about 1500 - 2000 heterogeneous documents every month, manually entering data into the 1C system.

Invoices and invoices for processing were received in heterogeneous formats through different tools: mail, cloud, WhatsApp and others. Each received document must be manually entered into a single database. Entering one document into the 1C database took up to 3 minutes. On average, up to 2000 paper documents had to be processed monthly. To do this, the documents were first scanned locally, then printed out at the main office and manually processed by the accounting department.

On average, a full intelligent scan of one contract takes no more than 10 seconds.

Before the introduction of IDP, the introduction of one document in 1C took up to 3 minutes. The total labor costs were approximately 120 working hours per month

The introduction of Visinionist idp allowed us to create a new business process based on intelligent data processing. Now everything looks different: all incoming and outgoing documents from contractors are scanned and sent to the Visinionist cloud.

Each new document is processed online. Data about the contractor, the service, the amount and the date of the service are extracted from the documents. A structured database of contractors and services rendered is formed from the collected data.

After implementation, the platform became available to all employees of the department through a convenient web application. Now they can initiate the reading process by uploading documents to the interface in manual or automatic mode.

After the introduction of Visionist IDP, processing of one document takes no more than 10 seconds without loss of accuracy. The platform is running in the background.

Reading, classification, extraction, standardization are performed automatically. Using the graphical interface, employees can visually perform cross-checks and make the necessary adjustments. After that, structured information is automatically transmitted for credit analysis, and a personal dossier is created for each client based on the processed data.

A large logistics operator operates in the Russian and CIS markets. Hundreds of contractors generate a constant stream of paper reports on orders taken and completed for the transportation and delivery of goods. The client's accounting department receives about 20,000 new invoices monthly.

The company works with a large number of contractors: delivery contractors, pick-up and delivery points. Such a business model creates a large-scale document flow. The client's back office is forced to process tens of thousands of heterogeneous documents every month, manually entering data into the 1C system.

Documents for processing were received in heterogeneous formats through different tools: mail, cloud, WhatsApp and others. Each received document must be manually entered into a single database. Entering one document into the 1C database took up to 3 minutes, in general, this task required about 500 work hours per month.

Given the wide variety of documents with different layouts and structures, there could be no question of a standardized approach.

Before the introduction of IDP, the introduction of one document in 1C took up to 3 minutes.

The introduction of Visinionist idp allowed us to create a new business process based on intelligent data processing. Now everything looks different: all incoming and outgoing documents from contractors are scanned and sent to the Visinionist cloud. Each new document is processed online. Data about the contractor, the service, the amount and the date of the service are extracted from the documents. A structured database of contractors and services rendered is formed from the collected data.

After the introduction of Visionist IDP, processing of one document takes no more than 3 seconds without loss of accuracy

Reading, classification, extraction, standardization are performed automatically. Using the graphical interface, employees can visually perform cross-checks and make the necessary adjustments. After that, structured information is automatically transmitted for credit analysis, and a personal dossier is created for each client based on the processed data.

• A qualitative leap in the extraction of documents and data
• Different layouts - Easily handles a wide range of document types
• Different languages - No barriers, the platform works seamlessly in different languages.
• Increases Values - Saving time and money can free up resources for you and your team
• Easy Accessibility - A machine learning system stores data in a well-structured and easily manageable way.

After implementation, the platform became available to all employees of the department through a convenient web application. Now they can initiate the reading process by uploading documents to the interface in manual or automatic mode.