Sberbank is set to introduce the AI PSLC methodology in October, focusing on the operation of software products and AI systems. This framework is designed to complement the existing AI-Disrupt PDLC approach, integrating development and operational processes into a unified model.
According to Kirill Menshov, Senior Vice President and Head of Technology at Sberbank, the methodology addresses the need for businesses to simultaneously accelerate change and ensure operational reliability. The model emphasizes that while AI agents increasingly handle engineering tasks, human oversight remains critical for setting goals, defining risk parameters, and maintaining accountability.
The AI PSLC framework introduces new requirements for system reliability, specifically targeting the accuracy and behavioral boundaries of AI agents. It establishes mechanisms for monitoring, metrics, and intervention to ensure agents operate within defined limits, with human approval required for actions exceeding those boundaries.
The model treats operations as a continuous source of product development. Data gathered from service performance, customer experience, and incident reports are used to refine requirements and validation processes, creating a feedback loop that links the operational lifecycle back to the development phase.
