Historians who work on a given subject area and time develop the language skills themselves to do the work. Yes, they can out-source some translation work — but that's more the exception than the rule in academic work. If you are going to be seriously doing a lot of work where the materials are going to be in a given language, you have to acquire at least reading-knowledge of the language (which is easier than learning to speak or write in a language). Most PhD programs in History require demonstrating reading knowledge in at least one language; many require two, even if you are planning to do research in your native language (e.g., French and German are common requirements, because they are also the language of scholarly writing).

So in your hypothetical, the person finding the document is probably not a Classicist (and one wonders how they even know that said "letter" is what it is in the first place). But such people exist (they are less niche than you probably realize; many universities have Classics departments full of such people), and they are the once who will be capable of using the document.

If one works in these worlds, one also gets to know people who speak and read different languages and also know your subject matter. So I can't really read a word of Japanese, but I know a number of scholars at different levels of rank who are fluent in Japanese, and if I have a small task that requires translating a document that would be interesting to them, I can sometimes ask them to help.

In addition to this, most better known texts from antiquity that are available to study at university don't have just the text, but are edited. Classic texts usually survive till today because these texts were copied by hand and spread this way. This was done mostly in Medieval monasteries, and of course, writing errors happened and not every (maybe even just fragmented) version of an Greek or Latin ancient text is the same. So, if one of our Classical sources is to be printed (and sometimes also translated in modern languages), there is a lot of minuscule detective work necessary to do so. Which surviving fragment or text is the oldest? Do we know which text they had to copy from? Can one writing mistake be found in other versions of this text so we can establish which copy was spread where? All the information gathered is usually incorporated in a good text edition and helps the historian interpreting his source and defining the reliability.

But, and that is relevant to your question, people that are doing this research accumulate a tremendous amount of knowledge on how languages change and how the history of transmission of knowledge did and does work. And these people are the ones to look for when you need to work on a new source.