The AI assistant in the Moscow region has checked more than a million medical prescriptions.
An AI-powered digital assistant, integrated into the Moscow region's healthcare system in February 2026, has processed over one million physician prescriptions in six months of operation.
The tool operates in two modes: during in-person patient visits and during telemedicine consultations when an existing prescription needs to be extended or adjusted. This is the first case of such large-scale neural network use in Russian clinical practice, and the results are impressive.
How prescription verification works
The algorithm automatically evaluates each prescription against 15 key criteria. These include drug compatibility, compliance with the patient's age and gender, as well as risks of overdose or exceeding recommended treatment durations. The system activates at the moment a prescription is issued: the neural network cross-references patient data with the prescribed medications and flags potential threats before the person receives the drugs.
The regional department's press service emphasizes that the technology adds an extra layer of accuracy when selecting a treatment plan, but the final decision always remains with the physician. Moscow Region Deputy Governor Lyudmila Bolataeva noted that the tool helps select the optimal therapy option, working equally effectively both during in-person visits and remote consultations.
It is important to understand: the assistant does not replace the specialist but merely complements their work, relieving the routine burden of checking drug compatibility. This is a sensible approach — AI takes on the analytics, but responsibility for the patient's health remains with the human.
Artificial intelligence in Russian medicine
The Moscow region is not the only area integrating neural networks into healthcare. Similar projects are being launched at the federal level and in individual regions, where AI handles routine checks and data processing. In Moscow, 200 medical AI services were tested over six years, 67 of which are already operating in hospitals. Over 2,000 clinics from 75 regions have connected to the city's MosMedAI platform, which provides access to radiology diagnostic algorithms based on imaging.
Yandex is also eyeing this market. In late July, the company registered LLC "Yandex Med" to develop medical software. The digital assistant transcribes conversations between doctors and patients, searches for relevant data in the knowledge base, and prepares a brief summary from the medical history.
The use of neural networks extends beyond clinics. They were used for the first time to assemble complete bacteriophage genomes — viruses that infect bacteria. Of the 16 such samples created using AI, all proved viable under laboratory conditions.
My conclusion: integrating AI into medicine is not just a trend but an inevitable stage in the industry's evolution. The Moscow region case demonstrates that neural networks can significantly reduce the risk of medical errors and improve the quality of therapy. The only question is how quickly other regions and private clinics will adopt these technologies. The market is clearly moving toward a hybrid model where humans and algorithms work in tandem.