Introduction. Differential diagnosis of malignant and benign skin neoplasms remains a key task in dermatovenereology and oncology. Primary care specialists are often overburdened due to excessive referrals of benign neoplasms to dermatovenereologists and oncologists. Traditional diagnostic methods and algorithms rely on physician experience and can be error-prone. Artificial intelligence (AI) based software (smartphone apps) using convolutional neural networks demonstrate high diagnostic accuracy comparable to that of experienced specialists. Integrating AI into diagnostic algorithms can improve differential diagnosis, optimize patient routing, and reduce the workload of specialists. Objective. To develop an algorithm for using AI-based software "Derma Onko Check" to improve the diagnosis of skin lesions, optimize patient routing, and reduce the burden on the healthcare system. Materials and Methods. A patient routing algorithm was developed using the findings of the Derma Onko Check program for the differential diagnosis of malignant neoplasms and dermatologic neoplasms based on photographic images. The results of clinical and instrumental studies and photographic images of 230 melanocytic and 151 keratinocyte (epidermal) tumors from 381 patients at the Burdenko Main Military Clinical Hospital were used in the development. The analysis was performed using a Python program with the pandas, numpy, scikit-learn, and matplotlib libraries. The optimal threshold for the routing algorithm was determined using ROC curves and the Youden Index to balance sensitivity and specificity. Routing and diagnostic duration were calculated based on the standard timeframes of the State Guarantees Program for Free Medical Care to Citizens for 2025 and for the planning period of 2026 and 2027, as well as orders of the Russian Ministry of Health on procedures for providing medical care to the population of the Russian Federation. Results. An optimal routing algorithm threshold of 62% was established, resulting in a sensitivity of 0.9123, a specificity of 0.9126, and a Youden index of 0.8249. Algorithms were developed for eight diagnostic report options. Without the AI program, diagnostic duration was 15–37 days; with the AI program, it was 1–37 days, reducing the overall duration to 1–23 days. Discussion. The routing algorithm minimizes false negative and false positive results, accelerates diagnosis, and reduces the workload of oncologists and dermatovenerologists. A routing algorithm threshold of 62% ensures the detection of malignant neoplasms. Conclusion. Using a patient routing algorithm based on the Derma Onko Check AI program for differential diagnosis of skin cancer and non-cancerous lesions improves accuracy and reduces diagnostic time, optimizes routing, and reduces the burden on the healthcare system.