RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2

T-Bank: Financial Health Comprehensive Customer Financial Assessment Service

Product
The name of the base system (platform): Artificial intelligence (AI, Artificial intelligence, AI)
Developers: T-Bank (Tinkoff Bank)
Last Release Date: 2024/08/15
Branches: Financial Services, Investments and Auditing
Technology: BI,  Big Data

The main articles are:

2024: Service Launch

On August 15, 2024, T-Bank announced the launch of a comprehensive assessment service for the financial condition of customers - Finzdorovye. With the help of Big Data and ML models, it analyzes the data of millions of T-Bank customers according to different parameters in order to display individual recommendations for each of them on changing the financial condition. The mission of the service is to help Russians manage their money, change the approach to savings and spending, as well as increase financial literacy. Finheald was developed as part of an AI banking strategy based on the technologies of the T-Bank Center for Artificial Intelligence (AI-Center).

source = T-Bank
Financial health

As reported, the service is based on those algorithms AI created by AI the T-Bank Center, which analyze tens of millions of impersonal data customers (spending, savings, investments, etc.). Next, the service displays indicators and personal recommendations on expenses and income.

Finhealth shows, for example:

  • the total amount of user money, including in other banks and cash, and recommendations for changing the financial condition;
  • expenses on paid subscriptions (including a function with a reminder of the debit of money three days before the payment date is available);
  • what is the probability of obtaining favorable conditions under; to the credit
  • credit history and user load and tips for improving them;
  • the share of loans in monthly income;
  • tips for improving financial literacy and prudent money management.

source = T-Bank
Financial health

Based on current analytics, the service analyzes the financial profile of the user according to five main parameters: all the client's money, credit history, income balance, subscription costs and investments.

In the near future, AI technologies and functions will appear in Finhealth:

  • automatic updating of information about their accounts from different banks in each other's mobile applications and obtaining a complete picture of their financial condition;
  • Financial Assistant based on Gen-T - T-Bank's own large language models (LLM). A personalized assistant will be able to answer questions about personal finance;
  • planning for the future financial condition using predictions from ETNA, a predictive analytics and forecasting service. Technology will help predict future spending, create a realistic budget for the next month, and plan financial goals and a plan to achieve them;
  • educational content for the development of financial literacy.

Finzdorovye displays how much money the client keeps in T-Bank (on the card, in investments, in savings accounts and deposits). In order for the service to take into account more cash and those stored in other banks, the user can deposit the amount manually. The overall financial picture will help the user track all his savings and understand whether there is an opportunity to go on vacation, buy a phone or take another action that requires expenses. Sberbank and T-Bank have already begun to introduce open banking in test mode in order to provide customers with the opportunity to combine information about their accounts in each other's mobile applications.

The service will tell you how long the accumulated money will last with ordinary spending - for example, for less than a month, for one to two months, or for six months or more. Depending on this, the assessment of financial health will change: from "already not bad" to "ideal." In addition, you can track changes in your financial condition and identify patterns (for example, in which month savings/expenses are growing faster), as well as get recommendations on how to optimize these indicators.

In Finzdorovye, all client loans from different banks are visible: cash loans, credit cards, installments, mortgages. The feature is free and available to all users who consent to a request for their credit history. Based on data from the credit history bureau (BKI), the service is trained to predict banks in which the user supposedly has loans - you can manually correct the name of the bank. In Finzdorovye, you can see your credit rating: the higher it is, the higher the likelihood of getting favorable loan conditions. To calculate the rating, ML scoring models created by the T-Bank Center for Artificial Intelligence are used.

source = T-Bank
Financial health

In addition, the service shows the credit load, that is, what percentage of the income goes to repay loans. This is one of the important indicators that banks look at when deciding to issue loans: if the percentage of debt burden is below 50%, the chances of approval are high.

In Finheald, the user can see the difference between their income and expenses in the form of a balance of income - how much money was left in the accounts after monthly spending. Artificial intelligence will show ideas on how to optimize the balance of income with increased cashback, transfers to savings and other tools in the bank. Also in this section you can see graphs that reflect the balance in accounts in dynamics - for example, in which months a person spent more, and when they managed to save money.

Based on past regular transactions, artificial intelligence is trained to determine what the user is subscribed to: all subscriptions paid for by the T-Bank card are reflected in Finheald. Users can track how much money they spend on subscriptions to external services - online cinemas, music streaming, etc. With a large number of subscriptions to them, a significant amount can go, which also needs to be taken into account when planning a budget.

You can also set up a function with reminders three days before the money is debited, which allows the client to cancel an already irrelevant subscription on time. The feature is available in a mobile app on iOS version 6.22 and higher.

In the service, the user can find out how much money is stored in his investment portfolio, how much he replenishes it and how much of the income is spent on investments.

For those who do not yet invest, the service will calculate how much can be transferred to investments, depending on what percentage of the income the user is ready to invest. Also in Finzdorovye there are useful materials and collections that will help the user immerse himself in the world of investments - for example, to find out what types of securities exist, how to collect your investment portfolio, exchange rules.

File:Aquote1.png
In the T-Bank mobile application, you can see a couple of clicks, for example, taxi spending over the past month. As part of the development of the AI banking strategy, we decided to go further and create a revolutionary service for Russia - a check of the financial condition of each client. As of August 2024, 43 million Russians - T-Bank clients, right in the application, can view detailed analytics of their financial condition, form financial habits and learn how to competently manage their money.

told Konstantin Markelov, Vice President, Director for Business Technologies of T-Bank
File:Aquote2.png

File:Aquote1.png
The combination of recommendation systems and ML scoring allowed us to create a product for assessing and controlling personal finances of the next level. We want to give users a tool for fully automatic management of personal finances, but under the control of a person. This will require the use of other technologies - our large Gen-T language models and ETNA predictive analytics.

told Viktor Tarnavsky, Director of Artificial Intelligence at T-Bank
File:Aquote2.png

The service is available to all T-Bank customers in the mobile application on Android version 6.14 and on iOS version 6.22 and higher. To use it, you need to open the action ribbon on the main page of the application and click on the blue icon.