Narrativa y construcción de sentido: reflexión y práctica sobre el uso de la IA agéntica en la gestión del conocimiento personal (PKM)

Authors

  • Jesús López de Leyva

DOI:

https://doi.org/10.23824/ase.v0i44.1022

Downloads

Abstract

The study examines how Personal Knowledge Management (PKM) systems can be enhanced through agentic AI to support self-directed and sustainable learning. Building on a conceptual review (DIKW framework, classical knowledge typologies) and established PKM methodologies (Zettelkasten, PARA, GTD, and Bullet Journal), it proposes and documents an applied workflow using Obsidian that integrates bidirectional linking, YAML metadata, periodic reviews, and spaced repetition functions. Operational practices are detailed — including daily and weekly templates, filters and databases, and synchronization of cards with Anki — as well as the incorporation of RAG and AI agents to generate Socratic questions, suggest connections, and produce study materials. A replicable technical–practical framework is presented that combines organization, attention, and working memory with responsible automation, emphasizing data exportability, interoperability, and control over task context. Ethical and design implications (discipline, tool transparency, and the limits of lock-in) are discussed, and criteria are offered for adapting the ecosystem to different preferences for autonomy and usability. The study concludes that, when well integrated into PKM, agentic AI enhances sense-making and long-term retention, provided it is accompanied by regular reviews, quality metrics, and an open architecture.

How to Cite

López de Leyva, J. (2025). Narrativa y construcción de sentido: reflexión y práctica sobre el uso de la IA agéntica en la gestión del conocimiento personal (PKM). Avances En Supervisión Educativa, 44. https://doi.org/10.23824/ase.v0i44.1022

Published

31-12-2025

Keywords:

personal knowledge management (PKM), agentic artificial intelligence, self-directed learning, educational technology, Zettelkasten method, PARA method, Obsidian (software)