A Madrid-based bank. A financier in Seville. A middleman in Mexico City. And then suddenly, out of nowhere, a typical Spanish-language investigation ends up in Vietnam. Nearly every intelligence analyst can identify with the way in which a routine investigation starts in country A and within a few days, leads into countries X, Y and Z. Often, these third country leads contain crucial intelligence that is vital to the success of a project. At Neon, we are lucky to have a team that speaks most of the world’s languages, so we are equipped to go wherever an investigation takes us.
Relying on machine translations can result in mistranslations, false leads and in the worst-case scenarios, inaccurate intelligence. No tool can replace proficiency in a foreign language; there are hundreds of instances where multilingual analysts succeed where popular translation tools fail.
Machines Don’t Do Non-Literal. Humans can.
Idioms or non-literal phrases such as “break a leg” and “piece of cake” are common elements of speech in most languages, but are a key stumbling block in machine translation. These expressions do not make sense when translated literally and may lead to false intelligence being provided by an investigator. For example, in a recent investigation we came across the Russian phrase “да не гони” on a forum discussion about cryptocurrency trading. The phrase is literally translated as “Don’t speed [drive fast]”, but in colloquial speech it means something similar to the English “get out of here!” or “seriously?!”.
What’s in a name?
Using the correct name of a subject is crucial to the success of any investigation, however, it is not always as simple as using the name provided in the project brief. In many regions, people use only part of their full name; in Spain and much of Latin America, for example, names traditionally consist of one or two given names followed by two surnames: the father’s then the mother’s. A Spanish-speaking analyst will know that it is common for native speakers to use only their father’s surname. For instance, if the subject of the investigation was a Uruguayan CEO called “Jaime Ortega García”, deep web searches using the name “Jaime Ortega” would likely yield better results.
Similarly, an analyst proficient in Slavic languages such as Polish or Ukrainian would be aware that native speakers often use diminutive forms or nicknames instead of their full names online. For example, a Polish woman called “Katarzyna” would likely be known as “Kasia” to her friends. Cultural insights like these are invaluable when trying to uncover more information about a subject’s personal life.
Facebook, Instagram, Twitter and… StayFriends
Social media searches are a key element of any online investigation, with valuable insights often buried in posts or comments from years ago. However, the term “social media” encompasses more than just the traditional Silicon Valley giants. An analyst with foreign language skills understands that each country has its preferred social media sites. For instance, a multilingual analyst investigating a Polish entrepreneur would know to check NK.pl as well as LinkedIn to establish their professional background, while an investigator with cultural or linguistic knowledge of Germany would be aware that StayFriends is a popular social media site in the region. These platforms, along with many others, are little-known to those unfamiliar with the target language.
With hundreds of billions of pieces of foreign language content being posted online each month, the challenge for monolingual intelligence analysts conducting transnational or global investigations is daunting.
In the high-stakes world of boardroom intelligence, only a human analyst can truly source, validate and contextualise raw information and turn it from data, to information and ultimately, intelligence.