Designing a Trustworthy AI-Enabled Knowledge Management Architecture to Improve Sales Readiness in B2B Technology Services

Authors

  • Gerry Gusman Wibowo Swiss German University
  • Heru Purnomo Ipung Swiss German University
  • Eka Budiarto Swiss German University

DOI:

https://doi.org/10.46799/ajesh.v5i11.843

Keywords:

Knowledge Management, Design Science Research, Trustworthy AI, Retrieval-Augmented Generation, Sales Readiness

Abstract

Early-stage sales readiness in business-to-business (B2B) technology services remains inconsistent due to fragmented organizational knowledge, varying levels of sales discovery capability, and the absence of standardized governance mechanisms for validating and controlling knowledge reuse. Although artificial intelligence (AI) offers opportunities to enhance knowledge retrieval, concerns regarding output reliability, hallucinations, and data integrity present significant challenges to trustworthiness. This study aimed to design a trustworthy AI-enabled knowledge management architecture to support early-stage sales readiness in B2B technology services. Adopting a Design Science Research (DSR) methodology, the proposed architecture was developed by integrating the Socialization, Externalization, Combination, and Internalization (SECI) model; the Knowledge Management Lifecycle; The Open Group Architecture Framework (TOGAF); Control Objectives for Information and Related Technologies (COBIT); the National Institute of Standards and Technology AI Risk Management Framework (NIST AI RMF); the European Union's Ethics Guidelines for Trustworthy AI; ISO/IEC 27001; and Indonesia's Personal Data Protection Law. The architecture was demonstrated through a retrieval-augmented generation (RAG) proof of concept at PT XYZ, an Indonesian information and communications technology (ICT) service provider. Data were gathered from structured and unstructured information sources, including documented organizational knowledge and tacit knowledge elicited from personnel. The proposed architecture was evaluated by six experts through focus group discussions (FGDs), using experience-weighted mean scores. The evaluation produced weighted mean scores of 4.54 for Knowledge Structure, 4.61 for Architecture, 4.61 for Governance, and 4.59 overall, all classified as Highly Achieved according to the study's evaluation criteria. These results provided expert-based support for the three proposed hypotheses concerning the architecture's assessed dimensions. The findings suggested that early-stage sales readiness depended not only on AI capabilities but also on effective knowledge management practices, structured knowledge resources, enterprise architecture alignment, governance mechanisms, information validation, and human oversight. The study concluded that organizations should prioritize centralized, well-governed knowledge repositories and adopt AI capabilities incrementally. Future research should evaluate the proposed architecture across multiple organizations and assess its long-term effects on sales readiness and business performance.

Downloads

Download data is not yet available.

Downloads

Published

2026-11-18