Pavel Moroz

Defined and executed a long-term AI roadmap, integrating ML models for classification, sentiment analysis, and user satisfaction prediction. Led the first commercial LLM integration at MTS Group, increasing automation by 10% while maintaining user satisfaction. Transitioned development to a product-oriented model, leading a 50-member team. Established KPI frameworks (GCR, FCR, Automation), improving classification quality by 10% and automation by 5–7%.

Implemented user segmentation for assistant scenarios, optimizing automation across versions. Designed a new UI interaction format (buttons, tooltips, widgets) in JSON, overcoming platform limitations. Developed AI-driven tooltips, increasing automation by 3–5% and used in 10,000+ daily dialogues. Led the development of predictive scenarios, boosting automation by 20%. Conducted A/B testing to refine assistant features and improve performance.

Led JAICP platform migration, ensuring integration, backward compatibility, and ML model support. Optimized classification thresholds and streamlined scenario transitions. Established a dedicated release team, standardizing validation processes to improve content quality and deployment efficiency. Conducted A/B tests, optimizing model performance. Automated release validation, reducing errors. Developed ML-driven hint suggestions, increasing automation by 5–10% and improving classification accuracy by 5–20%

Optimized assistant response logic using client data to enhance user interaction and business metrics. Managed request classification, improving accuracy through A/B testing and training set optimization. Automated DialogFlow updates, accelerating releases and error detection. Led Voice of Client (VOC) analysis, identifying conversion patterns to refine UX. Analyzed scenario performance and optimized FCR, closed sessions, and interrogations to enhance chatbot efficiency.