Desarrollo de una habilidad de inteligencia artificial (skill) para la automatización de los reportes de cartera: propuesta aplicada en MCASSAB COLOMBIA SAS

Abstract

The manual preparation of accounts-receivable reports consumes a considerable share of the accounting and credit-and-collection team’s time, delays the availability of information and increases the risk of error. In response, this report proposes developing a custom artificial intelligence (AI) skill on Anthropic’s Claude assistant to automate the generation of accounts-receivable reports at MCASSAB COLOMBIA SAS. An applied case-study approach is followed in which, based on several company databases receivables, clients, sales representatives and allowance, the process rules are coded into a reusable skill, so the user only needs to upload one report for the system to generate the deliverables automatically. The system does not output a single isolated report on its own; it is built from the integration of those databases, which are the process input. Once the report is uploaded, the skill produces the aging classification, per-salesperson reports, collection letters, the consolidated receivables report with US-dollar conversion, and the period-close report. A substantial reduction in preparation time is thus expected, moving from several hours of manual work to a single instruction, without replacing the accountant’s professional judgment.