Paper Type
Complete
Paper Number
PACIS2026-2060
Description
Organizations increasingly use Generative AI (GenAI) to support knowledge-intensive work, yet many deployments remain centered on ad hoc chat interactions rather than governed organizational processes. This creates problems for accountability, provenance, human review, implementation control, and continuous improvement. This paper proposes Process-Based Knowledge Management for AI (PBKM-AI), a BPMN-centered method for transforming GenAI-enabled knowledge work into explicit, implementable, and measurable knowledge processes embedded in organizational processes. PBKM-AI structures each use case through three connected layers: a business layer that clarifies the organizational process and central knowledge/information object, an implementation layer that maps process steps to GenAI capabilities, and an evaluation layer that links process performance, technical quality, knowledge-output quality, business value, and human experience. The method is formatively evaluated through Purchase Order Processing and Incident Reporting use cases, illustrating how GenAI-enabled knowledge work can be redesigned as auditable and improvable organizational routines.
Recommended Citation
Maslov, Ilia; Kudryavtsev, Dmitry; Khan, Umair Ali; and Simonofski, Anthony, "A BPMN-Centered Method for Process-Based Knowledge Management with Generative AI" (2026). PACIS 2026 Proceedings. 18.
https://aisel.aisnet.org/pacis2026/ai_ml/ai_ml/18
A BPMN-Centered Method for Process-Based Knowledge Management with Generative AI
Organizations increasingly use Generative AI (GenAI) to support knowledge-intensive work, yet many deployments remain centered on ad hoc chat interactions rather than governed organizational processes. This creates problems for accountability, provenance, human review, implementation control, and continuous improvement. This paper proposes Process-Based Knowledge Management for AI (PBKM-AI), a BPMN-centered method for transforming GenAI-enabled knowledge work into explicit, implementable, and measurable knowledge processes embedded in organizational processes. PBKM-AI structures each use case through three connected layers: a business layer that clarifies the organizational process and central knowledge/information object, an implementation layer that maps process steps to GenAI capabilities, and an evaluation layer that links process performance, technical quality, knowledge-output quality, business value, and human experience. The method is formatively evaluated through Purchase Order Processing and Incident Reporting use cases, illustrating how GenAI-enabled knowledge work can be redesigned as auditable and improvable organizational routines.
Comments
01-AIML