Abstract
The rapid expansion of digital recruitment infrastructure has generated unprecedented volumes of unstructured candidate data, imposing acute analytical burdens on Human Resource (HR) departments globally. Existing automated systems lack the contextual sensitivity required for prescriptive, equity-aware talent acquisition decisions. Objective: This systematic review synthesises the extant literature on contextual intelligence-driven HR analytics, with specific focus on Natural Language Processing (NLP), supervised Machine Learning (ML), and intelligent recommendation architectures applied to talent acquisition. Methods: A PRISMA 2020-compliant systematic search of five major academic databases Scopus, IEEE Xplore, Web of Science, ACM Digital Library, and Google Scholar identified 1,156 records, of which 72 studies met the eligibility criteria for qualitative synthesis and 58 for quantitative synthesis. Results: Logistic Regression consistently achieved the highest binary resume classification accuracy (97.63%, F1 = 0.99), while the proposed Decoder Attention with Pointer Network augmented with a coverage mechanism and Mixed Learning Objective (DA-PN + Cover + MLO) attained state-of-the-art abstractive summarisation performance (mean ROUGE = 27.78). A novel Contextual Intelligence-Driven Hiring Framework (CI-DHF) integrating an eight-layer pipeline from document ingestion through to prescriptive SaaS output is proposed and evaluated. Conclusion: The CI-DHF framework demonstrates strong potential for scalable, equitable, and prescriptively capable talent acquisition. Critical research gaps in algorithmic fairness, multilingual NLP, and deep learning integration are identified, and future research directions are delineated.
