Saudi Arabia's Vision 2030 construction programme is increasing the scale and complexity of tendering across infrastructure, buildings, tourism, transport and giga-project supply chains. Conventional bid evaluation remains vulnerable to inconsistent scope interpretation, fragmented tender documents, manual arithmetic checks, evaluator subjectivity and weak traceability between technical and commercial decisions. This review examines how artificial intelligence can support tender normalization and bid evaluation without displacing professional judgement or procurement accountability. Evidence published from 2020 to 2025 was synthesised across construction procurement, contractor selection, construction supply-chain digitalisation, natural language processing, explainable machine learning, supplier selection and AI-assisted bidding. The literature indicates that the principal opportunity is not autonomous award selection but the creation of a governed decision-support layer that converts heterogeneous tender submissions into a common comparison basis. Natural language processing can extract qualifications, exclusions, scope assumptions, programme commitments and contractual risks; rules and structured analytics can reconcile quantities, currencies, unit rates, provisional sums and arithmetic differences; and explainable models can support technical, commercial, sustainability and supplier-risk scoring. However, automation also creates risks of hidden bias, data leakage, inconsistent criteria, overreliance, algorithmic collusion and weak legal defensibility. This review therefore proposes a six-stage framework comprising data ingestion, document intelligence, tender normalization, multi-criteria evaluation, human governance and award recommendation, with project-performance feedback used to recalibrate future models. For Saudi Vision 2030 projects, the framework offers a route to faster, auditable procurement at scale while preserving fairness, local-content considerations, competition and accountable human approval.