ДОМ.РФ will develop AI projects in conjunction with Yandex Cloud

One of the areas is the automation of project financing processes. At the next stage - access to instant approval of loans for corporate clients. It is here that ДОМ.РФ see a large space for cooperation with Yandex Cloud.

Project financing is always a very large amount of issuance with a long life cycle of a credit product. The developer begins construction and receives money in tranches: for the first tranche, for example, he dug a pit, for the second - made a foundation, etc. Therefore, monitoring at all stages of the investment cycle is so important.

Nikolai Kozak, answering TAdviser questions, said that a tool was ДОМ.РФ created in the AI laboratory that analyzes construction in terms of dynamics in many respects in order to understand how the construction of the facility is going. Based on this, it is possible with a high degree of probability to predict the developer's default, see in advance possible problems with the company and take measures to improve the situation. And this is just one example of the possible application of AI in project financing.

Managing Director for IT and Digital Transformation ДОМ.РФ Nikolai Kozak (pictured right) and Commercial Director of Yandex Cloud Alexander Chernikov at the signing of the cooperation agreement "(photo: TAdviser)"

The signing of the agreement between ДОМ.РФ and Yandex Cloud was preceded by a joint pilot project, the results of which the companies summed up at the Finopolis forum. As part of this project, using machine learning technology, ДОМ.РФ analyzed the news background of investment and construction topics. She collects news according to given sources and automatically analyzes their tonality, which is divided into three categories: "positive," "negative" and "neutral."

If the company's information field is negative, automatic triggers are triggered. They highlight key information from publications, which is then shared with monitoring specialists in the project finance department.

The volume of information has now grown significantly, so using intelligent algorithms to analyze large amounts of text can be an effective way to organize and benefit from disparate information. This was shown by our pilot project. AI algorithms have been able to quickly and accurately analyze large amounts of information, automate the analysis process, ensure accuracy and relevance of results, be flexible and adaptive tools, and ultimately simplify the work of our departments and reduce business risks, "said Nikolai Kozak, Managing Director for IT and Digital Transformation of ДОМ.РФ.

As part of the pilot project, the technology was tested on several hundred news sources. The model is tested in the ДОМ.РФ bank circuit with access to the Telegram messenger via API and parsing messages from news channels. The accuracy of detecting negative public messages was 97%. In the future, the decision is planned to scale, gradually increasing the number of news sources. The ДОМ.РФ says that thanks to the data received, employees of the project financing department of the Bank will ДОМ.РФ be able to automate processes and avoid potential losses.

The range of applications of machine learning technologies in the tasks of banks and fintech companies is very wide and continues to grow. And partnerships between technology companies and financial sector leaders help to experiment faster with new technologies and more efficiently adapt the opportunities they provide to business processes and tasks, "commented Alexander Chernikov, Commercial Director of Yandex Cloud.

To implement the pilot project, the technological sandbox of the Fintech Association was used: a closed technological circuit created to test new IT projects. In the case of the pilot ДОМ.РФ and Yandex, it was important to be able to place a neural network outside the Bank's IT loop ДОМ.РФ to ensure regulatory standards and comply with the bank's information security requirements.