Reserving is one of the core businesses of the actuarial profession. For insurance companies, it represents major strategic and economic challenges and its practice is subject to ever more advanced developments. Standard claims reserves valuation techniques - including the Chain Ladder method - have historically been based on aggregated claims data. However, the recent expansion of data and analytics now allows a new framework. Innovative estimation methodologies based on information at the individual grid level are being designed. These methods, broadly spread by the actuarial community, are innovative theoretical developments. At this stage, and given the strong operational constraints (deadlines, data availability, etc.), these have not yet reached a sufficient level of maturity within the companies. In practice, the Chain Ladder method thus remains the reference for the valuation of claims reserves. There are therefore still many challenges to be addressed, such as obtaining individual estimates, detecting large claims, or, more generally, streamlining processes and in-depth knowledge of risks. To support the actuary in this context, and beyond the estimation method, we would like to reposition the data at the center of developments. We propose some best practices and a pragmatic exploitation of the possibilities offered by weak Artificial Intelligence: innovative visualizations, automation of repetitive tasks, machine learning modeling of complex processes.