Good practices of project management recommend that companies systematize the lessons learned in projects, that is, the key experiences that reveal the learning obtained in the process of executing a project and that may be relevant for future projects. But information needs vary according to the users' profession or social group, their demographic backgrounds, and the specific requirements of the task being performed. Thus, in order to make it possible to learn from the experiences and best practices of other people, each and every action of recording information needs to be carried out considering elements of context that take into account professional, organizational, social, geographic and temporal aspects. Based on a theoretical framework based on Information Science, Archivology and Computer Science, this research seeks to understand what it is and how to identify the context for structuring information systems to propose a conceptual model of the context for producing lessons learned to record learning in projects. Being an exploratory descriptive research, it uses the hypothetical deductive method and analyses the structure of five repositories of lessons learned available online in light of the basic contextual elements to record the context of lessons learned defined from an adaptation of the 5W+1H formula. As a result, it delivers a set of metadata for systematizing lessons learned and their respective association with the context, and the Conceptual Model for Lessons Learned in Projects (MCLAP), which has its centrality in the LESSON class (which describes the learning achieved and its applicability), and is directly related to other classes that define the objects necessary to record the context of production of this learning. Finally, the applicability of high-level ontologies to represent lessons learned in projects is discussed. By carrying out the mapping of the MCLAP in the BFO (Basic Formal Ontology) and in the IAO (Information Artifact Ontology), the applicability of these ontologies is corroborated as a basis for a domain ontology for lessons learned in projects, establishing the possibility of reuse of 8 (eight) entities already existing in the IAO and the proposition of 15 (fifteen) new entities with their respective properties for more specific elements. The concepts and relationships of the proposed conceptual model open opportunities for restructuring lessons learned repositories that, using artificial intelligence, can reveal useful content to users of a company in an intuitive and personalized way, contributing to the performance of companies and their projects.