What is Language Operations?

什么是语言运算?

2021-01-26 14:25 unbabel

本文共800个字,阅读需8分钟

阅读模式 切换至中文

As a multilingual entrepreneur, I’ve always been fascinated with the idea of building universal understanding. If that sounds somewhat lofty and ambiguous, let me explain. Many organizations find themselves limited in the markets they serve because of language. Even though English can give you a little more access in business, the idea that English is universal is an illusion. It’s realistically only spoken by a quarter of the world’s population. That leaves a lot of untapped opportunity on the table. Unbabel has found that most global organizations need to embrace language to maintain customer satisfaction and loyalty (more on that later). The opportunity is much greater than customer service alone. In fact, I’d argue that language is one of the largest unspoken biases in business. As the world changes and companies become more global and digital, they can’t afford to ignore the fact that they’ll need to communicate in multiple languages. So how do we solve this problem? Today, most organizations still rely on a siloed approach to language. For example, the marketing team might use translation services, while the sales team hires native speakers, and the customer service department staffs up with native-speaking BPOs. There’s no single language strategy that spans the entire organization. That leads to massive inefficiencies that can be solved with technology. That’s where our vision for language operations comes in: it’s a cross-disciplinary function that helps global businesses communicate more effectively with their multilingual customers and stakeholders. Here’s how it can work for every department across the organization. Spotting the early opportunity for language operations Language is the most basic, universal human connection point. At Unbabel, our big idea is to provide an AI-powered, human-refined translation layer that eliminates the need for native speakers in every market but still provides high-quality translations. We started by solving the problem of multilingual communications in customer service because the opportunity cost of not providing multilingual support is too great. Today, customers can come to online businesses from anywhere in the world. During the pandemic, in particular, global cross-border shopping has increased – online sales grew 21% from January to June of 2020, compared to the same time period last year. If something goes wrong, how many of these people can get support in their native language? According to a recent study from Intercom, only 28% of them. That’s despite the fact that 70% of people would feel more loyal with native language support – 35% of them even said they’d switch products altogether to get native support. However, the opportunity for language operations doesn’t start and end with customer service alone. Operating at a local level, at scale Much like we’ve seen the rise of operations-for-everything powered by technology (think revenue ops, DevOps, marketing ops), Unbabel imagines a new category of language operations, in which AI-powered, human-refined machine translation technology serves as the centerpiece to power an entire organization’s language strategy. Language operations leverages existing tools in the technology stack (in customer service, that’s tools like Zendesk, Helpshift, and more) to help everyday people communicate in any language. It’s a discipline that can help businesses grow more effectively, using data to drive decisions. For example, language operations can help a business understand exactly which languages are in highest demand so that organizations can support customers better and build stronger relationships. Though the pandemic has led to many companies working remotely, for a company trying to find a new market for a product, local language and cultural norms still matter. As technology democratizes language even further, every company will be able to function like a multinational corporation and do business anywhere. In the near future, we’ll see AI powering business operations beyond the customer service organization – across functions like sales, marketing, and more. As a result, more companies will be able to test product fit in more markets, especially as language becomes an asset, rather than a barrier to doing international business. Even if the head of the Japanese office is an English speaker in Austin, Texas. Looking to the future It’s not just Unbabel that believes technology has become the great equalizer for language around the globe. There’s a surge in demand for language learning apps like Duolingo, and growing excitement around multilingual artificial intelligence models like GPT-3, as well as more recent advancements from Google and Facebook. In other words, we’re all seeking universal understanding in different ways. In 2021 and beyond, I anticipate that the category of language operations will grow as research in the machine learning field for language continues to take off. The ultimate goal? Empower organizations of any size to reach the scale and measurability of a sophisticated, multinational corporation.
作为一名多语言企业家,我一直致力于构建人们能够相互理解的方式。如果这听起来有些学术和模糊,请允许我来解释。 由于语言的原因,许多组织发现自己在服务市场中受到限制。在商业活动中,虽然英语能提供给你更多的机会,但是有这样一个错误的想法,英语是各国通用的。实际上,世界上只有四分之一的人口说英语。这就留下了许多待发掘的机会。Unbabel发现,大多数全球性组织需要相关的语言来维持客户的满意度和忠诚度(稍后将有更多内容说明)。这一机会比单靠客户服务业务要大得多。 事实上,我认为,在商业中,语言是最大的隐性偏见之一。随着世界的变化,公司变得更加全球化和数字化,他们不能忽视的一个事实是,他们将需要使用多种语言进行交流。 那么我们如何解决这个问题呢?今天,大多数组织仍然停留在使用单一的语言策略层面上。例如,营销团队可能使用翻译服务,而销售团队雇用母语人士,而客户服务部门则配备母语人士进行外包业务。没有一种单一的语言可以适用于整个组织的的运行。由于语言不同,大量的低效率问题随之产生,而这些问题可以通过技术来解决。 这就是我们对语言运作的愿景:语言运作具有跨学科的功能,语言运作能够帮助全球企业更有效地与他们的多语言客户和利益相关者进行沟通。以下是语言运作如何协助组织中各个部门的工作。 及早发现语言运作的机会 语言构成了人类最基本,最普遍的联系。在Unbabel,我们的宏观构想是,提供一个由人工智能驱动的,人工译后编辑的翻译层,使得在每个市场上,无需母语人士,但我们仍然能够提供高质量的翻译。因为不提供多语言支持的机会成本太大,所以我们从解决客户服务中的多语言沟通问题开始。 今天,身处世界各地的客户可以同时进行网上交易。特别是在新冠疫情大流行期间,全球跨境购物增加了。与去年同期相比,2020年1-6月的在线销售额增长了21%。如果在线销售出了问题,这些人中有多少人能得到母语交流?根据内部通信公司最近的一项研究,只有其中的28%的人可以以母语交流。尽管如此,70%的人会支持母语交流--其中35%的人甚至表示,为了获得母语交流,他们会完全更换产品。 然而,语言运用并不仅仅是开始和结束于客户服务业务。 语言在地方的大规模运用 正如我们看到的,由技术驱动的一切操作(比如收入操作,开发操作,营销操作)的兴起那样,Unbabel设想了一个新的语言策略类别,其中由人工智能驱动的,人工译后编辑的机器翻译技术是整个组织语言战略的核心。 语言操作利用技术堆栈中的现有工具(在客户服务业务中,例如Zendesk,Helpshift等工具)来帮助人们用各种语言进行日常交流。这是一个可以帮助企业实现更有效地增长的工具,利用数据来推动决策。例如,语言运用可以帮助企业准确地了解哪些语言最受欢迎,这样企业就可以更好地维持客户关系,并与客户建立更牢固的关系。 尽管新冠疫情大流行导致许多公司采用远程办公的方式,但对于一家试图为产品寻找新市场的公司来说,熟知当地语言和文化规范仍然很重要。 随着技术使语言进一步普及化,每家公司都能像跨国公司一样运作,能够在任何地方开展业务。在不久的将来,我们将看到人工智能在客户服务组织之外的业务运营中发挥作用--有了超越销售,营销等多种职能。因此,更多的公司将能够在更多的市场上测试产品的适用性,尤其是当语言成为一种资产,而不是进行国际业务的障碍时。即使日本办事处的负责人是得克萨斯州奥斯汀的一位讲英语的人,也可以轻松交流。 展望未来 不只是Unbabel相信技术已经成为全球语言的重要交流工具。人们对于Duolingo等语言学习应用的需求激增,而且对GPT-3等多语言人工智能模型以及谷歌和Facebook的最新进展也越来越感兴趣。换句话说,我们都在以不同的方式寻求普遍的理解。 在2021年及以后,我预计,随着机器学习领域对语言研究的继续深入,语言运用这一类别将会增长。我们的终极目标是什么?是使任何规模的组织都能达到成熟的跨国公司的规模和标准。

以上中文文本为机器翻译,存在不同程度偏差和错误,请理解并参考英文原文阅读。

阅读原文