Roadmapping tecnológico fundamentado em inteligência artificial: um estudo sobre gestão da inovação agrícola no setor público

Abstract

The disruptions caused by digital technologies bring both opportunities and threats, which manifest differently in the context of private companies and the public sector. As the focus shifts towards strengthening transformative innovation, public research organizations are increasingly called upon to consciously address the broader impacts and their responsibility for the various transformations to which they actively contribute through their Research and Development. The transition to Digital Agriculture, which incorporates robotic autonomy and artificial intelligence, involves not only technological development but also reflection on how various transition paths to sustainable agricultural and food systems are related to responsible innovation and mission-oriented innovation environments. Managing the activities of the innovation process is seen as the most recurrent organizational barrier in literature on public sector innovation. Technology Roadmapping (TRM), a process that utilizes structured systemic thinking to align strategic planning with innovation management, still has research gaps in its integration with data mining. Assessing the applicability and boundary conditions of innovation theories in light of Artificial Intelligence and the Data-Driven Innovation (DDI) is also in an early stage in the field of administration. It is suggested to propose an extended process that includes an explicit exploration phase before development, where the refinement of the idea and/or innovation concept occurs through data resource experimentation and exploration of social relationships. Therefore, the aim of this work was to propose a framework and an AI agent (chatbot) that together establish the basic precepts for an AI-augmented technology roadmapping process (Augmented TRM). The methodological approach followed the guidelines for projects in Design Science Research (DSR), related to the perspective that academic knowledge should be relevant to the practical field, bridging the gap between theory and practice. The DSR research project included the following phases: problem identification, definition of the solution's objective, development, demonstration/evaluation, and communication. As the dissertation is structured in article format, the first article addresses the issue of innovation and artificial intelligence, the second discusses a roadmapping solution for the context of agricultural innovation, and the third covers the developed framework process and its evaluation by experts (interviews) who volunteered to contribute to this research. As a result, there is a strong interest in using generative AI for activities requiring qualified cognition, such as technological prospecting. However, this should occur with some level of human supervision to be defined after experimentation cycles. In evaluating the applicability of a process enhanced by AI's self-learning, the organizational impact goes beyond automating discrete tasks and extends to innovatively altering existing processes and introducing entirely new tasks.