Better targeted and more personalized service offering to citizens has the potential to make state-citizen interactions more seamless, reduce inefficiencies in service provision, and lower barriers to service access for the less informed and disadvantaged social groups. What constitutes personalization and how the service offering can be customized to meet individual user demand is, however, much less clear and underdeveloped partially due to the technical and legal dependencies involved. The paper gives an overview of how personalization and customization of digital service offering have been discussed in the literature and systematizes the main strand emerging from this. It follows up with a case study of the Estonian X-road log data as one potential way to detect latent user demand emerging from an experienced life-event that could form a basis for letting users define their service needs as holistically as possible. The results show the existence of distinct service usage clusters, with specific user profiles behind them, a clear indication of latent demand that leads to a simultaneous consumption of otherwise independent digital services.

As interest in the digital transformation of public administration grows, the main challenge remains to improve government governance systems and integrate a wider range of evidence into decisionmaking processes. The successful digitalization and application of such approaches improves the quality, responsiveness and flexibility of public administrations. The digialtization of processes has made it possible to use micro-level data to assess the impact of a policy or program and apply the feedback to improve the design and delivery of public services. Evidence-based policy-making evaluates programs based on their visible impacts. Large-scale data collected through digitized governance, coupled with econometric impact assessment, provides an ideal working toolkit for this. However, the current situation of European governments is one of slow adoption, as they are often slow to respond to new challenges. This is due to the static one-off impact assessment approaches used, the results of which quickly become outdated. With further digitalization, improvement of systems, and a rapidly changing situation, there is a need to speed up institutions' ability to quickly draw working solutions to offset the effects of unexpected events in society and economy and react without delays if policy effects dissipate. This paper demonstrates how a high level of digitalization in government allows addressing such issues by automating causal impact assessment and making it a continuous part of the service delivery. The use case is an automated system for assessing active labour market policies in Estonia using individual-level data from government digital registers. Building on this, it shows how impact assessment automation depends on automatically generated data, only available due to the digitalization of other public services, and how versatile it is when it comes to proving casual evidence in a suddenly changing environment.