Search the Moravia Blog

Blog

Chatbots 101: Six Ways They Will Impact Localization

Posted by Lee Densmer on Mon, Jan 16, 2017 @ 05:20 PM

Multilingual chatbots

You may already be using chatbots but not know it. Have you ever sent a text message to get an instant bank balance or a weather report?

A chatbot is software used within a messenger service (like Facebook) that automates things like making a dinner reservation, setting an alarm, adding an appointment to your calendar, or finding and sharing news. You could also text with a bot to order food, shop for clothes, and find restaurants.

Chatbots are not new: first there was Microsoft's Clippy, and then there was Tay, Microsoft’s teenage chatbot who wound up tweeting racist and offensive messages.

Look at this Domino’s pizza bot that lets you order a pizza by tweeting the pizza emoji to @dominos. They’ve got this down to a five-second exchange. Also, Taco Bell has released a chatbot that allows you to order and pay for food through an automated chat conversation.

Some others include:

And here’s a fun one:

(For a full list, go here.)

It’s predicted that anything an app can do, a bot will be able to do in the future — and more easily. 

With the ongoing global boom in mobile messaging apps (WeChat, Facebook Messenger, and Slack), bots are becoming increasingly relevant. This fast growth of chatbot applications and interactions leads us to think that users like to (and indeed may sometimes prefer to) interact with bots. Many people find that sending a text message is much easier than opening up an app, logging in, navigating through menus, and selecting options. Chatting one-on-one with a bot is also more personal. (Ironic, since that bot is not a human.) And the target market is clear: the Economist reports that over 2.5 billion people use messenger apps, and that many teenagers now spend more time on smartphones sending instant messages than perusing social networks.

Six potential impacts to localization

  • Speedy localization will be essential. With users wanting to do more and more complex things via bots, and with more bots being developed and updated all the time, localization programs will have to keep pace. An agile localization workflow will be imperative.

  • Market entry decisions may not be obvious. Releasing bots in the same markets as apps may not make sense. The markets with the highest utilization of chatbots might not be the same ones that heavily use applications — demographics and usage patterns differ. Data and trends need to be analyzed so enterprises can make decisions about which markets to enter with which products.

  • Translators will have to stretch their skills to learn this new medium. Translators have to take into account the source language, functionality of the bot, and the way users deal with it in the target environment in order to be successful at their work. Pavel Doronin discusses this in the Jan/Feb 2017 issue of MultiLingual Magazine (page 57).

  • The shorter text will impact the translator’s approach. The message has to be concise to fit small-sized menus and brief notifications. It might be difficult to translate text into, say, Spanish — which can expand up to 30% when translated from English — and have it fit the length limitations.

  • Speech recognition technology, in all languages, will be challenged to improve. Verbal inputs will be an important way of interacting with chatbots when we aren’t in a position to type (while driving, for example). While the technology has come a long way, more advancements are needed to separate speech from background noise, handle slang, and respond appropriately. It gets more complicated when you think about the need for speech recognition in many languages.

  • Emojis pose lots of challenges. These are a major component of any SMS communication, and their usage grows every day. But there are technical issues: do CAT tools and QA tools support emoji? If they don’t, they will be displayed as unrecognized symbols. Also, translators need to be emoji-fluent because emojis are often used to express a concept without any text at all. (This point also comes from Pavel Doronin in MultiLingual.)


One thing is for sure: chatbot development will increase as user preference shifts towards them, and as more and more of the world engages with SMS applications. Software developers have to consider the implications of this medium for their global business and the impact on localization. While the full effects of the chatbot explosion on the localization industry have yet to be determined, it’s worth starting to think about the challenges above.

Further reading:

Topics: Localization Technology