Salesforce, Atlassian, and Coursera Share Localization, NLP Insights at SlatorCon

Salesforce, Atlassian, and Coursera - SlatorCon March 2022

The latest edition of SlatorCon Remote featured localization buyers from Salesforce, Atlassian, and Coursera, who presented their takes on current language service and tech demand.

CRM giant Salesforce was represented by Caiming Xiong, VP and Managing Director of AI Research. As R&D lead for the company who collaborates with clients across the globe, Xiong shared insights on the use of natural language processing (NLP) in the context of hybrid contact centers (i.e., call centers that combine human agents with NLP applications such as chatbots).

He cited four areas where NLP can enhance the customer experience (i.e., knowledge extraction, conversational analysis, dialogue, and summarization) and how clients, big and small, were able to scale using their no-code / low-code CRM platform.

Noting how the goal of any contact center is to build a strong customer relationship and enhance customer loyalty, Xiong said it must provide three things off the bat.

  • Instant response – Waiting half an hour on the line or hearing a recorded message saying, “We only provide customer support from 9:00 AM to 5:00 PM, Monday to Friday,” no longer cuts it.
  • Accurate response – Queries should lead to the correct answer.
  • Cross-lingual support – As so many companies already support multiple languages, customers are less tolerant of being disconnected because they happen to speak a different language.

Hybrid contact centers are trying to solve or improve these three aspects, Xiong said, leaving the human agent more time to handle complex tasks; and humans, in general, to handle what AI cannot, such as mapping out the vision for the business.

Xiong also shared details on how Salesforce uses machine translation in their work with big international companies that support customers in many languages. Watch the rest of the presentation via Slator’s Video On-Demand.


Widely regarded as Australia’s first unicorn company, Atlassian develops collaboration software that facilitates teamwork for more than 200,000 businesses worldwide in 190 countries. The company was represented at SlatorCon by Melanie Heighway, Manager of the Product Localization Team, who gave the audience an inside look into the three facets of the company’s product localization workflow.

  • Traduki – The Esperanto word for “translation” describes the step where content is pushed to or pulled from code repositories into the translation management system (TMS).
  • TMS – Managed by the in-house localization team, Atlassian’s TMS includes a translation productivity (aka CAT) tool for translators.
  • MLVs – The company works with multilingual vendor partners for translation, review, and quality assurance.

Heighway shared the opportunities she and her team used to evangelize localization during financial year planning, innovation week, hackathons, and so on. As many of Atlassian’s engineers speak English as a second language, they understand the importance of localization — and have actually stepped forward to collaborate with Heighway’s team to enhance the customer experience through internationalization.

She made the case for securing executive sponsorship for localization by listing three reasons why they would be interested in supporting localization / globalization: they are mindful of OKR (objectives and key results) and want to grow; they care about the customer experience; they want to measure success.

Heighway also shared the company’s future localization plans. Watch the rest of the presentation via Slator’s Video On-Demand.


E-learning platform, Coursera, has some 97 million registered learners, 80% of whom are located outside the US, according to Localization Program Manager, Nora Duong — and they expect that percentage ex-US to grow even higher.

Duong outlined three types of e-learning localization at Coursera.

  • Subtitle-only localization – Only the subtitles of lecture videos are translated
  • Full course localization – Lecture video subtitles, in-video questions, reading materials, quizzes, assignments, and so on are translated. (Lecture audio remains in English.)
  • Full course localization with voice-over – Similar to the preceding, but lecture audio is dubbed in a new language.

She walked the SlatorCon audience through how a localized course is created, including the parts of the workflow where they work with language service providers (LSPs): course translation and importing.

Localization success is measured two ways at Coursera, Duong said, via direct data (i.e., ROI and revenue; contracts and deals; user growth) and “juicy” data (platform interactions, flagged feedback, user stories).

Learn more about how Coursera works with LSPs to make their products and content more accessible to global users by watching the rest of the presentation via Slator’s Video On-Demand.