Implementación de Business Intelligence (BI) mediante test A/B y análisis comparativo de la eficiencia en la rendición de gastos en LATAM Airlines.
Loading...
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
El presente trabajo tiene como objetivo desarrollar y evaluar la implementación de un ecosistema de aprendizaje, basado en herramientas digitales y automatización mediante inteligencia artificial, orientado a mejorar la eficiencia en el uso de la plataforma SAP Concur (en adelante, CONCUR) por parte de los funcionarios de LATAM Airlines. La empresa enfrenta dificultades relacionadas con la existencia de múltiples errores en la rendición de gastos corporativos, tanto para comisiones de servicio como para fondos por rendir, producto del desconocimiento operativo de los funcionarios.
El diagnóstico inicial confirma una problemática que impacta directamente en el ámbito contable. Para abordar esta situación, se establece una propuesta de mejora inicial compuesta por un manual de COBUS y videos tutoriales. Posteriormente, se desarrolla una evaluación manual para determinar qué información resulta relevante y de qué manera es mejor comprendida por los funcionarios.
A partir de este proceso se construye una propuesta final compuesta por cinco herramientas: un manual de procedimientos detallado de COBUS, one-pager, resúmenes concisos en una sola página sobre procesos u otras acciones, videos tutoriales, un “Dashboard” interactivo en Looker Studio Pro denominado “Guía Rápida CONCUR”, un asistente virtual desarrollado en la plataforma Amelia y una presentación paso a paso que detalla el acceso a todas las herramientas. Estas soluciones se crean con un enfoque multilingüe (español, portugués e inglés), considerando la diversidad cultural y operativa de LATAM en sus distintas filiales.
El impacto de estas herramientas se mide mediante un experimento A/B. A través de tecnologías de Business Intelligence como Google Analytics y Looker Studio Pro, se lleva a cabo la recopilación y análisis de datos, evaluando indicadores clave como el nivel de satisfacción declarado, la tasa de revisión con rechazo de informes, el tiempo de uso y la frecuencia de utilización de las herramientas por los funcionarios.
Los resultados permiten demostrar que la implementación conjunta de materiales de apoyo y de un asistente virtual fortalece la autonomía de los funcionarios, reduce errores y contribuye a aumentar la eficiencia del proceso de rendición. En este sentido, la Guía Rápida CONCUR alcanzó más de 660 usuarios, con un total aproximado de 8.100 visualizaciones. Por su parte, el asistente virtual registró un total de 63 consultas, complementado con 304 visitas y 179 clics de acceso desde la guía rápida. Estos datos evidencian un nivel de adopción inicial significativo y confirman que las herramientas constituyen un apoyo concreto en la reducción de errores y en la eficiencia operativa.
The present work aims to develop and evaluate the implementation of a learning ecosystem, based on digital tools and automation through artificial intelligence, oriented to improve efficiency in the use of the SAP Concur platform (hereinafter, CONCUR) by LATAM Airlines employees. The company faces difficulties related to the recurrence of multiple errors in the reporting of corporate expenses, both for service commissions and for funds to be rendered, resulting from employees’ operational unfamiliarity. The initial diagnosis confirms a problem that directly impacts the accounting domain. To address this situation, an initial improvement proposal is established, consisting of a COBUS manual and tutorial videos. Subsequently, a manual evaluation is carried out to determine what information is most relevant and in which format it is better understood by the employees. From this process, a final proposal is constructed, consisting of five tools: a detailed COBUS procedures manual, one-pager, concise single-page summaries of processes or other actions, tutorial videos, an interactive dashboard in Looker Studio Pro called the “CONCUR Quick Guide”, a virtual assistant developed on the Amelia platform, and a step-by-step presentation detailing how to access all the tools. These solutions are created with a multilingual approach (Spanish, Portuguese, and English), considering the cultural and operational diversity of LATAM across its subsidiaries. The impact of these tools is measured through an A/B experiment. Using Business Intelligence technologies such as Google Analytics and Looker Studio Pro, data collection and analysis are conducted, evaluating key indicators such as declared satisfaction levels, report rejection rates, usage time, and the frequency of use of the tools by employees. The results demonstrate that the combined implementation of support materials and a virtual assistant strengthen employee autonomy, reduces errors, and contributes to increased efficiency in the expense reporting process. In this sense, the CONCUR Quick Guide reached more than 660 users, with an approximate total of 8,100 views. Meanwhile, the virtual assistant recorded a total of 63 queries, complemented by 304 visits and 179 clicks from the Quick Guide. These data highlight a significant level of initial adoption and confirm that the tools provide concrete support in reducing errors and enhancing operational efficiency.
The present work aims to develop and evaluate the implementation of a learning ecosystem, based on digital tools and automation through artificial intelligence, oriented to improve efficiency in the use of the SAP Concur platform (hereinafter, CONCUR) by LATAM Airlines employees. The company faces difficulties related to the recurrence of multiple errors in the reporting of corporate expenses, both for service commissions and for funds to be rendered, resulting from employees’ operational unfamiliarity. The initial diagnosis confirms a problem that directly impacts the accounting domain. To address this situation, an initial improvement proposal is established, consisting of a COBUS manual and tutorial videos. Subsequently, a manual evaluation is carried out to determine what information is most relevant and in which format it is better understood by the employees. From this process, a final proposal is constructed, consisting of five tools: a detailed COBUS procedures manual, one-pager, concise single-page summaries of processes or other actions, tutorial videos, an interactive dashboard in Looker Studio Pro called the “CONCUR Quick Guide”, a virtual assistant developed on the Amelia platform, and a step-by-step presentation detailing how to access all the tools. These solutions are created with a multilingual approach (Spanish, Portuguese, and English), considering the cultural and operational diversity of LATAM across its subsidiaries. The impact of these tools is measured through an A/B experiment. Using Business Intelligence technologies such as Google Analytics and Looker Studio Pro, data collection and analysis are conducted, evaluating key indicators such as declared satisfaction levels, report rejection rates, usage time, and the frequency of use of the tools by employees. The results demonstrate that the combined implementation of support materials and a virtual assistant strengthen employee autonomy, reduces errors, and contributes to increased efficiency in the expense reporting process. In this sense, the CONCUR Quick Guide reached more than 660 users, with an approximate total of 8,100 views. Meanwhile, the virtual assistant recorded a total of 63 queries, complemented by 304 visits and 179 clicks from the Quick Guide. These data highlight a significant level of initial adoption and confirm that the tools provide concrete support in reducing errors and enhancing operational efficiency.
Description
Tesis presentada para optar al título de Ingeniero/a Civil Industrial.
Keywords
Inteligencia artificial, Análisis de costos, Procesamiento de datos