Large companies invest hundreds of millions of dollars in systems with artificial intelligence in Russia. What do they hope to gain from machine learning?
The introduction of AI by 2030 will give an increase in world GDP by 14%, that is, about $ 15.7 trillion, according to PwC. The consultants note that this is more than the current total volume of industrial production in China and India. According to the Teradata study, the vast majority of large companies (80%) invest in artificial intelligence technologies, and according to Gartner forecasts by 2020 they will be present in almost all new software products and services.
In Russia, the volume of markets for the use of artificial intelligence and machine training in 2017 amounted to about 700 million rubles, but by 2020 to 28 billion rubles, according to a study of TAdviser and the system integrator of Jet Infosystem. Director for the development and implementation of Jet Infosystems, Vladimir Molodykh, believes that the reasons for the use of AI vary from the considerations of following the fashion to economic justification: "It became possible to use AI in a cost-effective manner in broad areas." True, while the examples presented on the RAIF are not impressive. The future lies in everyday trivia.
Drivers of applying AI were the financial area, insurance and retail. The researchers reported that the leader in the number of projects remains the financial industry. Here, technologies allow to reduce costs, minimize risks, prevent fraud, check borrowers, assess their solvency, carry out forecasting, etc. In particular, PJSC "Bank URALSIB" applies AI for analyzing customer data, which allows more personalized approach to its proposals. Also, the bank recently transferred the product catalog to a non-relational database managed by AI, which, when the tariff is changed, structures the data itself without requiring the software code to be changed. "The financial sector has always been at the forefront of technology use, and 15 years ago banks began using AI to predict defaults using self-learning neural networks. Now AI allows us to solve a wider range of tasks: minimize risks, prevent fraud, check borrowers, assess their solvency, carry out forecasting, and so on, "explains Alexander Sakharov, Head of Digital Banking at PJSC Bank URALSIB.
In retail, the main direction of artificial intelligence applications is related to customer service, logistics optimization, stock inventory inventory, cost reduction and demand forecasting. Among examples from the retail was the company M.Video, which introduced artificial intelligence technologies to improve customer service through an online store. It analyzes the client's behavior on the site, transitions between pages, views of the sections, after which the system prepares personal product recommendations that are sent to the registered user by mail.
To increase the conversion due to the return of departed visitors and increase the response from marketing mailings, the company implemented a solution based on machine learning, which determines the optimal time for interaction with the customer to make a purchase. In particular, a mechanism is used that analyzes the preferences of other buyers. According to Sergei Sergeev, director of information technology M.Video, as a result of these and a number of other measures to optimize the website, sales for the first 9 months increased by 30%, and the conversion of visitors to customers - by 10%. In insurance, artificial intelligence is used to manage document circulation, process customer data, select risks, fight fraud, identify personalized insurance services and distribute insurance payments. In Russia, among the leaders of this segment can be identified company "Alpha Insurance", which was one of the first to introduce elements of AI to address the main problem of the insurance market - fraud in settling losses. The system, analyzing the data of insurance cases of OSAGO, including client data, allows to identify among the total number of doubtful cases those in which there is really a possibility of fraud on the part of the client, for the subsequent investigation. This helps to identify fraud and fence off other customers from additional spending.
"We can observe how the revolution takes place in the insurance market - from the object approach, when earlier we tried to aggregate factors by groups of clients, that is, to all people with the same age, length of service, machine, etc. to offer a common tariff, to an individual approach, when each client now has the opportunity to get a personal fare based on their skills, "explains Sergei Shishkin, deputy financial director of Alfa Insurance. He also notes an important feature of AI, which probably slows down its implementation: "The machine is not trusted to make money, because from it you will not ask the result. And with a manager's man, you will ask. "
Director of sales of industry solutions Yandex Data Factor Konstantin Gorbach introduced cases in the industry, primarily, the introduction of artificial intelligence at the Magnitogorsk metallurgical plant. The launched solution helps to decide on the range of the chemical composition of raw materials for steel to optimize the quality and cost of production. The claimed effectiveness was about 5%. Further, according to a similar model, the company introduced artificial intelligence into the production of chocolate and extraction of gold. The payback of such projects, according to the developer, is 6-9 months.
As the experts note, despite the high potential for market growth, in Russia the practical introduction of technologies is hampered by rather high investments in projects, given the existing doubts in financial returns. In addition, for the development of a business-oriented AI barrier is a lack of computing power, the availability of high-performance infrastructure will allow to correct the situation. Despite this, there are already examples of AI implementation in the domestic market, which prove the effectiveness of their application and the benefits for business. Many companies realize the need for active investment in AI, this is confirmed by the desire of the leaders of the venture market represented by Sberbank and Larnabel Enterprises investment funds, intending to invest more than $ 100 million in projects based on AI. Proceeding from this, experts say that already today the use of AI / MO technologies in business is an "arms race". Who will be the first to introduce, optimize business processes, make the best offer to the client, he will win.
Author: Taufik Khisamov