Letter of Credit (LC) transactions depend on the accurate examination of documentary presentations against the credit terms and applicable documentary-credit rules. Traditionally, this activity has been performed by skilled document checkers, making the process labour-intensive and sensitive to document volume, wording variation, and cross-document inconsistencies. This review examines the evolution of LC document management and verification from manual examination and basic digitisation to intelligent document processing, artificial intelligence, and human-in-the-loop approaches. A structured literature review was conducted using recent peer-reviewed research, international trade rules, and authoritative guidance, with emphasis on developments reported during 2025–2026 while retaining foundational sources required to explain the regulatory and technical context. The literature indicates that optical character recognition, document classification, machine learning, natural language processing, semantic matching, and optimisation can support extraction, consistency checking, discrepancy identification, and workflow prioritisation. Recent research also shifts attention from isolated automation toward hybrid human–AI processes, where AI handles repetitive or high-volume analysis and experienced reviewers handle exceptions and consequential judgments. Digital presentation rules further introduce requirements concerning electronic records, authenticity, integrity, examination, and data corruption. The review identifies persistent gaps in domain-specific datasets, explainability, standardised evaluation, interoperability, governance, and validation under real banking conditions. Future research should therefore focus on auditable, standards-aware, explainable, and human-supervised verification architectures.

Keywords— Letter of Credit; Documentary Credit; Document Verification; Intelligent Document Processing; Artificial Intelligence; Human-in-the-Loop