Our client is a legal department of a financial company, which is engaged in the preparation and sending claim documents to the court to collect financial debts. With the growing customer base and increased volume of operations, the company faced the need to increase the number of lawyers and encountered enlarged number of errors while preparing documents for filing into court. The company management realized that the extensive development path would only lead to a further reduction in the efficiency of the department and a fundamentally new approach to solving problems was needed.

Initially, the task to prepare claim documents looked quite simple as there were many formal, similar and routine operations.

But after a thorough joint analysis of the basis of treaties, types of debt, the history of legislative changes and other significant factors, it became clear that the task was very complex. With significant volumes of transactions, the presence of "old" debts burdened by a complex legal history, a long-term contractual base, the automatic formation of a document package became a non-trivial task. Both the texts and the claim documents had to take into account all the nuances of the contracts made over a long period of time, as well as the requirements of the local judicial authorities to the submitted documents.

• To automate the process of dealing with problem debt, starting with the formation of a document package for the court and ending with a detailed recording of the interaction with the judiciary bodies and the debtor himself in each case.
• To increase the speed of document preparation as well as the speed of processing responses from the court.
• To minimize the impact of the human factor, leaving for lawyers the only function to input initial parameters and to conduct the final control, thereby reducing the number of errors.
• To organize recording of outgoing and incoming documents using digital paperwork.
• To improve the department efficiency with the help of the information system.
• To provide the automatic formation of court documents, namely:

1. Application for a court order;

2. Claim for recovery of funds;

3. Calculating the amount of debt after the term of the loan agreement;

4. Calculating the amount of debt for the duration of the loan agreement;

5. A court order.

• To ensure the management and analytic departments are able to quickly analyze and monitor the performance of each employee.
• To reduce training requirements for the staff working with the information system.
• To conduct staff training.

During the work on the project, all the business processes of the unit were analyzed, the specifics of the law changes was studied, the loan portfolio of contracts made during the last 5-6 years was analyzed. All existing contracts were typified and got flexible templates of claim documents, which took into account various features of the contract "life", namely:

• The existence of additional agreements to the loan agreement;
• The presence of partial loan clearance during the contract term and after it;
• The existence of contract extensions;
• Breakdown of debt by debt items;
• Facts of change in the name and form of ownership of the loan organization and in the name and form of ownership of the collection company itself;
• The history of contractual relations between the loan organization and the collection agency;
• The use of different correspondence addresses, depending on the geographical location of the debtor and the court reviewing the claim;
• Other additional factors.

As a result of the project implementation, the company's management received objective information about the state of court cases with any level of detailing, as well as up-to-date information about the effectiveness of each specialist in the unit.

Since the output was to be a set of claim documents requiring minimal control and manual editing and almost completely ready for the court, it became possible to reduce the qualification requirements to lawyers working with the system.

It became possible to flexibly take into account local courts requirements for drawing up court documents.

The system saved the drawn-up documents recording the date, time and data about the lawyer who performed the operation.

There was a track of correspondence with both the judiciary bodies and the defendants of court cases.

Python, Django, xslt, xml, apache, js, PostgreSQL, MongoDB, LibreOffice API.

• Number of docs: more than 1 million
• Changing in quantity of fails in court: became 70% less
• Project team: 9 developers
• Quantity of court cases: 200K+
• Hours spent for development: 700
• Complexity: 8 out of 10