AIOps (Artificial Intelligence for IT Operations)

AIOps (Artificial Intelligence for IT Operations) is the application of artificial intelligence (AI) and machine learning (ML) technologies to automate and improve the management of IT infrastructure and services. This is not one specific product, but rather an approach or platform that analyzes huge amounts of data from different sources in real time to help IT teams work smarter and faster. In fact, AIOps is a smart assistant for IT professionals who takes over the routine, analyzes data and prompts solutions.

2025: How AIOps is developing in Russia and what is hindering it. TA opinions

In a world where the complexity of IT infrastructure is growing avalanche-like and incidents are measured in millions of lost revenue, old manual control methods are no longer working. Under these conditions, AIOps is gaining popularity - artificial intelligence for autonomous management of IT systems. TAdviser spoke with experts about the current development and prospects of this method in Russia in December 2025.

Vladimir Bobrov, head of research and development at Mons (part of Korus Consulting Group of Companies), said that the Russian AIOps market is showing steady growth: many companies are at the stage of active implementation. The focus has shifted from experiments and hypotheses to solving practical problems - this is especially true for large corporations and state IT systems. The development is supported by a course towards digital transformation and the emergence of mature domestic solutions that are trying to solve key problems to increase fault tolerance and automate incident analysis, the TAdviser interlocutor noted.

𝓲

Фото: Freepik

Rostelecom notes that AIOps in Russia is developing fragmentally: it is already working somewhere - in first-line assistants, in simple scenarios for classifying incidents. The general director of RTK Soft Labs (part of Rostelecom) Evgeny Semenov listed two main barriers that interfere with the development of this area. The first of them is the integration of AI into the production circuit.

AIOps assumes that information security, operation and development work in the same cycle. We don't have that yet. Information security does not understand how to control the work of models, does not see transparent control mechanisms and therefore acts through prohibitions, - said Semenov.

The second problem is data. Without the normalization of logs, camouflage and uniform storage rules, no ML agents will work, he said.

Igor Dyachkov, head of the DevOps service at Nobilis.Team, says that by the end of 2025, AIOps in the Russian Federation is more under discussion than real implementation in most companies. AI is still used in IT Operations pointwise - mainly in large banks, telecom operators and individual IT holdings that are trying to automate the monitoring and support of complex infrastructure. However, there is no need to talk about widespread use, the vast majority of enterprises still rely on traditional monitoring systems and manual response to incidents, the expert added. In his opinion, there are several "obvious" barriers to the development of AIOps in Russia:

• lack of qualified personnel and expertise;
• Skepticism and cultural resistance (many IT professionals do not trust the AI "black box," preferring proven manually configured tools)
• Financial and economic (implementation of AIOps requires significant investments in software, computing resources, and employee training, while the return on investment is difficult to calculate in advance)
• sanctions and withdrawal of major Western vendors;
• purely technological problems (many organizations do not have the data collection in the amount and quality that is needed for the efficient operation of algorithms).

According to Pavel Chushmarov, Deputy General Director for Sales of Software Solutions at Mobius Technologies, AIOps penetration restrains the difficulties of import substitution.

Many customers still have or are just planning projects to replace Western monitoring platforms, ITSM, CMDB, observability, unified DWH for IT data. Without this "base," the launch of AIOps analysis on disparate or planned IT decommissioning systems will be ineffective, he said.

Outlander also draws attention to such a factor as high risks and distrust of AI's automatic intervention in operational IT processes. This is especially evident in conservative industries and CIIs, where a long downtime may be less of a problem than an erroneous AI decision, the expert emphasized.

Georgy Kashintsev, head of the SRE department at Rambler & co, adds to this list a lack of tools and standards. According to him, most DevOps/SRE processes were already well automated long before the advent of AIOps. These processes are described directly as code. At the same time, the very idea of ​ ​ automated infrastructure is not new and has long been implemented, for example, in the form of cloud platforms, the expert continues. According to him, it cannot be said that the development of AIOps does not occur at all.: It is concentrated in the form of R&D within large IT companies and continues to this day.

And about. Ksenia Polyantseva, head of the Data Intelligent Analysis Department at the Moscow Technical University of Communications and Informatics (MTUSI), says that in practice AIOps is more often used to correlate events, intelligently monitor, predict failures and reduce the burden on operation services. Complete autonomous automation is still rare, AI mainly helps specialists make decisions, she added.