This research paper explores the challenges and intricacies of translating metaphors in economic texts from English to Arabic, particularly through the prism of machine translation. It examines the challenges of conveying figurative language across different cultural contexts. A comparative analysis of twelve English metaphorical examples and their Arabic translations, taken from business websites and presumably to be machine-generated, is carried out. These translations are then compared with those produced by DeepL, a neural network-based translation system. The study highlights the strengths and weaknesses of machine translation. It shows that there is continuing need for the expertise of human translators who bring cultural and contextual sensitivity that machines currently lack in relation to figurative language. Ultimately, the research argues for a balanced approach that leverages the efficiency of machine translation alongside the nuanced understanding of human translators to ensure effective cross-cultural communication in business.