Semantic annotations of digital documents are used to explicitly describe the meaning of data. They are extensively used to automatising data exploitation, analysis, and sharing, especially in the Big Data and Artificial Intelligence (AI) era.The maintenance of these annotations is a complex problem that companies have to deal with. Domain knowledge evolves and new versions of models (ontologies) created to represent this knowledge are frequently published. These ontologies describes the meaning of terms (concepts) used in a domain. These concepts are also used to annotate documents. Thus, changes in the ontologies can impact the quality/validity of existing annotations leading to the paradox of “Improving the knowledge of a domain can causes loss of knowledge”. If the annotations do not follow the knowledge evolution, then the interoperability between systems can be compromised, misinterpretation of information become possible, and the accessibility to information is reduced.Our solution helps companies that produce huge quantity of data to deal with the dynamic nature of knowledge and its consequences on the annotations. We developed a method to automatically identify the ontologies’ changes and their impact on annotations. We propose to users to apply our annotation evolution method to update existing annotation, and we also propose a smart information retrieval method to ensure the accessibility to annotations that, for any reason, cannot be modified.In this project, we will evaluate in a specific market the scale up capacity of our solution and we will implement some new services, identified as necessary, during the market study.