Study goals
Propose an application developed in JavaScript, utilizing the ReactJS/NextJS framework, with a Redis database running locally in a Docker container, integrated with ChatGPT through LangChain, for the semantic search of lessons learned in project management.
Relevance / originality
The research integrates project management and artificial intelligence through emerging platforms and solutions, along with an original application, aiming to solve a practical problem in project management.
Methodology / approach
Instrumental case study in eight stages: data loading, embedding/vectorization processing, mathematical vector storage, question capture, embedding/vectorization processing, database search, prompt submission to ChatGPT, and response retrieval.
Main results
The study demonstrates the ability to provide accurate and contextualized answers, reducing the time needed to find relevant information, making knowledge accessible to all team members, regardless of their familiarity with technological tools or the complexity of the data
Theoretical / methodological contributions
The framework integrates two relevant areas: project management and artificial intelligence. Additionally, the study proposes an original application and a novel (authored) methodological process.
Social / management contributions
The application reduces search time, returns more accurate results, and is understandable even for individuals without technological expertise, minimizing barriers to the use of lessons learned.