Moscow, 1 June 2026 – X5 (the “Company”, MOEX ticker: X5) a leading Russian food retailer, has summarised the results of scaling artificial intelligence (AI) across its business processes. Throughout 2025, the Company launched AI solutions across multiple areas – from commerce and logistics to customer experience and corporate functions. The cumulative economic impact of implementing AI solutions amounted to approximately RUB 5 billion in additional operating profit for 2025.
Technology-wise, the Company scaled AI using X5’s AI Core framework, which integrates computing infrastructure, model management tools, mechanisms for testing and evaluating solutions, AI agent operations, access to various language models, and business-facing application services. This centralised approach enables X5 to move beyond isolated pilots and transition successful AI models and agent-based scenarios into business use under unified standards for quality, security and impact assessment.
The primary contribution to EBITDA came from AI models embedded into core retail processes: demand forecasting and restocking, pricing, assortment management, and recommendation engines. Additional benefits were achieved by automating operational processes, developing digital products, and accelerating the deployment of solutions into production.
At the application level, X5 leverages AI in commerce, logistics, customer experience, and software development. Computer vision automatically monitors product availability on shelves and compliance with merchandising standards; recommendation systems help create personalised offers; and the AI assistant GalyaGPT helps shoppers select products, recipes, and ready-for-use shopping scenarios.
A key outcome of 2025 was the widespread adoption of AI by Company employees. The corporate X5 CoPilot service, which provides access to various AI models through a unified interface, was used by over 32,000 employees by year-end, with 21,000 using it regularly. One the basis of ACE X5 (AgentOps), employees independently created more than 2,300 AI agents that processed over 150,000 user requests.
“X5 has been using machine learning and artificial intelligence in daily operations for many years. In 2025, we moved to the scaling phase and integrated AI Core into all key processes. This delivered a significant economic impact from AI adoption, measured against transparent business metrics and validated through financial models, A/B testing, and joint assessment with business teams. In 2026–2027, we will focus on systematically embedding AI into everyday business processes to make it an integral part of X5’s operating model.” – said Alexander Kostin, Managing Director of Х5 Tech.
To validate the business impact of digital and AI solutions, X5 uses its proprietary online A/B testing platform, ABsalute, which accelerates experiment launch and analysis and helps to assess the effect of adopted changes on key business metrics. Meanwhile, Machine Learning Operations (MLOps) standardises model training, production deployment and monitoring, shortening the path from development to implementation. The AI-Run service provides unified access to various AI models and helps to optimise the cost of solution deployment.
The financial impact of AI initiatives is evaluated through their contribution to EBITDA, operational cost savings, and total cost of ownership of solutions. For key initiatives, impact is validated using A/B tests and financial models. This approach enables the Company to scale not the number of AI pilots, but solutions with proven economic returns.