Understands Lab Results Better Than a Search Engine: Sber AI Interprets Medical Tests
In 2025, the KDL Medscan laboratory network began offering patients more than a printout of numbers. Alongside test results, patients now receive a structured, plain-language report with interpretation. The report is generated by AI.
Now It Makes Sense
Receiving a lab report filled with columns of numbers and unfamiliar abbreviations often triggers mild panic. What do those arrows mean? What are urates or phosphates? Why is one indicator marked “out of range”? The next step is predictable: people type “elevated lymphocytes” or “low hemoglobin” into a search engine, only to encounter a dozen conflicting explanations – from simple vitamin deficiency to alarming diagnoses. This ritual is familiar to almost anyone who has ever taken lab tests. Russia has now introduced a tool designed to put an end to it: an AI assistant based on the GigaChat neural model, developed by SberMedII.
The idea behind the rollout is straightforward. Anyone receiving lab results can now also get, free of charge, a detailed breakdown of indicators along with personalized recommendations. The system does not produce generic text. It factors in sex, age, and individual reference ranges, turning raw data into clinically meaningful information.
A Practical Medical Application
The project represents the practical deployment of advanced technology in one of the most conservative and high-responsibility areas of healthcare – laboratory diagnostics. At its core is a large language model that has already demonstrated medical competence by passing final medical examinations in general medicine and several specialties at a leading Russian medical center. Its knowledge is now being applied directly to patient support.
The system’s rollout illustrates how digital technologies are moving beyond abstraction to solve concrete, large-scale problems. Laboratory testing is among the most common medical services, used by tens of millions of Russians every year. For many patients, interpreting results is stressful and confusing, while searching online often leads to false conclusions and heightened anxiety.
Trust as a Key Factor
The Russian AI assistant addresses this challenge at a systemic level. It provides a baseline of reliable and safe information, reducing unnecessary strain on primary care physicians and saving patients’ time. This supports a shift toward preventive medicine, where the focus moves from treating disease to maintaining health. With clear explanations, patients can spot deviations earlier and make more informed decisions – whether adjusting lifestyle habits or consulting a specialist.
Developers emphasize a multi-layer validation framework. Each AI module is trained on verified medical data, and final moderation is carried out by experts from the KDL laboratory network. This creates a balance between innovation and oversight. Legal safeguards are equally important: handling of personal data, especially medical information, fully complies with Russian law, setting an important precedent for similar projects.
Offering an Alternative, Not Catching Up
Globally, the medical AI market is crowded with solutions focused on imaging analysis such as CT and MRI scans. Automated interpretation of routine laboratory data, however, remains far less developed. Preliminary testing shows the Russian system achieved 93% accuracy, which developers say is 8% higher than comparable foreign solutions. If confirmed in real-world use, this could translate into a significant competitive edge.
The export opportunity lies less in selling software outright and more in exporting a service model or methodology. This could include licensing the platform to foreign laboratory networks or private clinics, particularly in developing healthcare systems where shortages of laboratory physicians are more acute.
Setting a New Service Standard
Providing AI-based interpretation could become a strong competitive advantage for laboratories seeking to attract and retain patients. Success of the project may strengthen Russia’s position on the global HealthTech map as a source of practical, validated, and ethically designed solutions for mass healthcare.
The project’s value extends beyond individual convenience. With patient consent, the system processes anonymized datasets, becoming a powerful tool for medical analytics. It can help identify regional disease patterns, track population-level trends, and uncover previously unseen correlations between biomarkers.
For physicians, the software acts as a useful assistant, taking over routine preliminary analysis. Freed-up time allows therapists and specialists to focus on deeper clinical assessment and direct interaction with patients.
