Instead of a copy of SAP, Sber is creating an ERP platform with AI-native architecture.

      Sber is developing a next-generation ERP platform with AI-native architecture and plans to launch it on the market in 2027. The company believes that the Russian corporate systems market is entering a phase of deep technological transformation following the exit of SAP.

      Sber is working on its own ERP platform with AI-native architecture, which it expects to bring to market in 2027. The project was discussed by Taras Skvortsov, Deputy Chairman of the Board and CFO of Sberbank, at the CIPR-2026 conference.

      According to the bank's estimates, the volume of the Russian ERP systems market will reach about 100 billion rubles by the end of 2025 and may increase to 120 billion rubles by 2030. The potential volume of the segment that Sber is targeting with its solution is estimated by the company to be more than 20 billion rubles.

      As Taras Skvortsov explained, Sber has abandoned the idea of creating a direct analogue of SAP. The company is focusing on developing a next-generation ERP system in which artificial intelligence is initially embedded in the platform's architecture. It is expected that such a system will not only perform accounting and automation functions but will also serve as a basis for autonomous management of business processes. Within this concept, applications will be able to independently analyze context, generate work scenarios, and initiate necessary actions.

      The platform has been under development for about three years. In terms of complexity and scale, the project is comparable to the implementation of SAP. The transition to the new ERP system within Sber is planned to be gradual to avoid disruptions in key processes. The launch of the first most mass operations into industrial operation is scheduled for January 2027, and the main volume of processes is expected to be transferred to the platform by January 2028.

      The system is being created based on Sber's own technological platform, Platform V, microservice architecture, and the GigaChat language model. Sber emphasizes that the use of its own LLM should reduce dependence on external technology suppliers and the associated risks. The platform is already being used internally by the bank for managing finances, procurement, warehouse operations, logistics, real estate, HR processes, and reporting.

      One of the central elements of the project is the GenAI process designer. It is designed so that employees without programming skills can independently create and modify processes for current tasks. In the future, the system is expected to automatically suggest optimal work scenarios using accumulated data and operational experience.

      Sber has focused on hyper-automation of business operations and the software development process. This involves digital assistants that can take on the execution of routine operations both in business processes and in software development. The company believes that this approach will accelerate the formation of necessary business competencies and reduce costs for performing standard tasks. The ultimate goal of the project at Sber is to create a model of a self-managing business, where a significant portion of operational activities will be performed automatically.

      A feature of the platform will be its self-learning architecture. It is expected that data structured and labeled in the logic of the AI layer will allow the system to learn independently and improve the efficiency of back-office functions without constant manual tuning.

      According to Taras Skvortsov, many companies today face the problem of disparate AI tools that are difficult to integrate with each other. Sber hopes to offer a comprehensive end-to-end solution covering the main back-office management processes. The company views the project not only as a standalone software product but also as an attempt to form a new model for organizing internal corporate processes.

      As the bank's top manager noted, the ERP systems market is currently undergoing a serious technological transformation. In his opinion, import substitution in this area is not only about abandoning Western software but also about the opportunity to rebuild the architecture of corporate systems with an AI-native approach in mind.