Risk-oriented re-ingineering of digital government services using a modular test bench

DOI: 10.31673/2412-9070.2026.017415

Abstract

The article proposes an expanded step-by-step roadmap for the reengineering of digital public services, centered on a modular experimental stand that functions as a “digital testing ground” for testing transformation scenarios before their implementation in a productive environment. The approach combines organizational transformations, data structure changes, platform solution modernization, regulatory compliance, and risk-based change management. The stand architecture is structured into an e-services layer, event log collection and preprocessing modules, an analytical layer using process mining, machine learning models, hybrid CNN+LSTM and AE+LSTM autoencoder, as well as experiment orchestration and load generation modules. It is shown how the stand reproduces real cases of business registration, social assistance allocation, and obtaining permits with the ability to measure service provision time, failure intensity, staff workload, and risks of illegal decisions in AS-IS and TO-BE modes. An integral indicator of reengineering efficiency, a risk model and simplified queue models for assessing response time are proposed, which allows quantitatively comparing alter native scenarios and formulating an optimization problem for selecting a set of scenarios for a service portfolio taking into account budget and risk constraints. Fragments of the software implementation of microservices, the log collector module and data preprocessing are presented, as well as a scheme for training neural network models to detect anomalous application processing scenarios. A model for scaling the roadmap to different categories of e-services and an approach to organizing personnel training based on bench scenarios are substantiated, which together reduce the uncertainty of deep reengineering and increase the validity of management decisions in the field of digital transformation of the public sector.

Keywords: digital service reengineering; e-government; modular experimental stand; process mining; machine learning; AE+LSTM; CNN+LSTM; microservice architecture; risk management; digital transformation roadmap; government e-services.