The increasing reliance on chatbots for legal translation purposes has been discussed in a number of recent scholarly studies. Nonetheless, so far only general LLMs (such as ChatGPT) have been scrutinized in the field of legal translation. This contribution is aimed at exploring the accuracy and reliability of legal translations performed by field-related chatbots, such as ChatLaw and Notebook LM. In addition, it wishes to investigate whether corpus-based post-editing can increase the quality of automated target texts while, at the same time, keeping the translation process seamless and expeditious. To this aim, the ChatLaw and Notebook LM AI-driven systems are prompted to translate an extract of an employment contract from Italian into English. An ad hoc corpus of employment agreements is composed and used for post-editing purposes. More precisely, chatbot-generated target language, collocations, and language patterns are attested in corpus data, where also word frequencies are accounted for. The findings indicate overall satisfactory results, where chatbot-driven target texts feature minor shortcomings, such as slight collocational or colligational issues, redundancies, and some inappropriate literal renditions. In this scenario, corpus-based post-editing is easily and swiftly mainstreamed, thus producing satisfactory results without resulting in a time-consuming process. Thanks to AI and corpus-driven tools, legal translation becomes less lexically demanding and more expedited. Nonetheless, human supervision and legal expertise remain central to guarantee the high quality and reliability of the results.