Denis Vodeneyev, Director of the IBS

[tanjimajuha20]

reach tens of percent."

As Dmitry Svalov noted, AI in DevOps can be used for a wide range of tasks, including those beyond code testing, configuration management, and anomaly detection: “For example, AI can be used to quickly respond to incidents and make decisions on how to eliminate them by managing the infrastructure. However, in such cases, it is important to exercise caution and set limits to maintain a high level of security.”

Automated Testing Department, named areas where AI assistance in writing code, refactoring, and preparation of test data are successfully applied: "At the moment, the south korea whatsapp resource proven benefit within our projects where it is possible to use it (and for the most part, the tools are cloud-based) is aimed at middle+ level specialists. This is due to the fact that they can independently verify the result of AI work and maximize the useful effect of its application within routine tasks. For projects where it is impossible to use tools outside the circuit, we are preparing our own toolset using AI for testing, which can be deployed locally at the customer's and which does not require an Internet connection to work. According to our estimates, the efficiency of middle+ level specialists has increased by 30% without loss of quality. This has also allowed us to reduce the size of the work team and optimize the tasks of specialists on the ground. I think it is no secret that any specialist is much more interested in solving non-trivial tasks, instead of understandable but labor-intensive ones, thereby increasing the motivation of the guys, which, in my opinion, also has a positive effect on the timing and quality of the project."

Head of the Content AI Development Department Alexander Subbotin named the analysis of test results and script generation among the tasks that AI successfully copes with: "We try to use AI wherever possible. For example, we use ML/AI to generate test data, to analyze test results, in particular quality tests. We also use AI to write code, scripts, tests, test scenarios. To do this, we further train local LLM models on our data."

Alexey Lebedev, Head of Financial Sector at Reksoft Group, believes that AI tools can be effectively used at every stage of software development, provided that specialists are trained to work with this technology: "In the future, AI will be able to perform almost all software development tasks. Until then, AI will increasingly strengthen teams and specialists of all profiles: help analysts develop requirements, architects create and improve architectural solutions, developers write and debug code, testers create a strategy and test scenarios, create automated tests, emulate system users, information security specialists manage risks and identify vulnerabilities, Dev/Sec/Data/ML/LLM/Ops specialists deploy and operate systems. Over time, development teams will become smaller, as it will be easier for people to combine several roles with the support of AI. As a result, even one person will be able to create and support solutions, managing a team of highly effective AI agents. If today an AI agent operates at the level of a novice practitioner, then in a year it will be able to compete with experienced specialists. In 2024, according to the forecasts of the RUSSOFT association, about 20% of Russian IT companies will have a measurable effect from the use of generative AI in software development, the gain will be up to 20%.