A reference architecture for governed cognitive enterprises combining knowledge memory and decision intelligence with agentic AI under continuous assurance

Iranian Journal of Science and Technology, Transactions of Electrical Engineering · Springer Nature · 2026-11-19 · DOI: 10.1007/s44163-026-02719-6

Аннотация

Enterprises increasingly combine large language models, retrieval-augmented generation, and autonomous software agents with existing enterprise architecture, knowledge management, and analytics investments, yet these capabilities are commonly adopted as isolated point solutions without a shared architectural foundation for organizational knowledge, decision reasoning, human accountability, or continuous assurance. This fragmentation leaves organizations unable to explain how an automated recommendation was produced, to bound what an autonomous agent may do, or to demonstrate governance to auditors, regulators, and boards. This paper asks how enterprise knowledge, decision support, agent execution, and governance can be structured as one coherent architecture. Using a literature-grounded conceptual synthesis across enterprise architecture, knowledge management, retrieval-augmented generation, decision intelligence, agentic artificial intelligence, and artificial intelligence governance, the paper derives a consolidated research gap and proposes the Governed Cognitive Enterprise Architecture: six integrated layers, a human accountability boundary, a structural formal representation, an operational lifecycle, and a maturity model, the Cognitive Enterprise Readiness Index, for organizational self-assessment. The architecture is evaluated conceptually against explicit criteria for completeness, capability integration, and governance alignment, and is compared with the reviewed literature through a source-referenced framework rather than an experimental benchmark. The principal contribution is an integrated architectural vocabulary and structure that treats human accountability and continuous assurance as first-class concerns alongside knowledge and decision capability. As a conceptual reference architecture it has not been empirically validated in an operational enterprise; expert evaluation, case-study, and quantitative validation are specified as future work.

Читать у издателя →

Похожие статьи