AI-Assisted Strategic Intelligence: A Hybrid Approach Using LLMs and Deterministic Methods

Abstract

Strategic and scientific intelligence increasingly draw on artificial intelligence (AI), yet its precise role across different tasks remains unclear. This paper contributes a three-fold perspective. A review of recent research shows how NLP and machine learning are applied to patents, publications, and other textual data. A market overview highlights strong automation capabilities but limited integration of large language models (LLMs), with only exploratory use in current tools. Finally, a case study on the Online Mendelian Inheritance in Animals (OMIA) database compares deterministic and LLM-based methods across data acquisition, extraction, change detection, coherence evaluation, and rigor assessment. Deterministic techniques remain optimal for structured tasks, while LLMs add value in summarization and contextual analysis. Together, the findings underscore the potential of hybrid architectures that combine deterministic precision with LLM flexibility to strengthen future AI-assisted intelligence systems.

Ähnliche Arbeiten