Key Language Industry Insights from SlatorCon Remote March 2024

SlatorCon Language AI Conference Summary March 2024

The first SlatorCon Remote conference of 2024 was held on March 20, and there was no shortage of memorable insights to reflect upon. The event began with an introduction by Slator Managing Director Florian Faes, in his “industry health check.” 

Faes walked attendees through the changes in direction we are witnessing in the language industry, which finds itself at a critical crossroads influenced by AI technological advancements, market pressures, and evolving client expectations.

On the mixed financial health of language service providers (LSPs), backed by data that includes Slator’s Language Service Provider Index (LSPI), Faes said the industry shows overall stability, with the super agencies reporting some growth. 

A few companies had a tougher year, however, with nearly 40 percent of LSPs reporting negative growth, “the highest we’ve ever had,” remarked Faes, along with two recent bankruptcies: WCS Group and Lengoo. “We see that undifferentiated providers really are struggling,” added Faes.

Aya: A Massively Multilingual Dataset and Language Model

The keynote presentation was delivered by Marzieh Fadaee, a Senior Research Scientist at Cohere, who focused on the outcome of a project called Aya. 

Fadaee described Aya as a single, large, multilingual language model on which many researchers and evaluators from multiple countries collaborated. The model covers many languages that most datasets do not include, “making Aya particularly valuable,” said Fadaee.

Fadaee invited audiences to try the dataset and model, available under an open-source license, and to sign up to use Aya in 22 languages in the “Aya Cohere Playground.”

APIs, AI, and Strategy

Florian Faes moderated a panel that included Lynn Nguyen, Customer Success Operations Director at DeepL, and Dierk Runne, Senior Manager of Localization & Systems at HubSpot. Faes set the stage for the discussion mentioning DeepL’s free “Report: The state of translation and localization in 2023-2024,” which includes two HubSpot case studies.

Starting from that context, Nguyen explained that DeepL was trying to understand the market it serves. To her, the survey validated that boosting customer engagement with localized content makes sense and pays off. Runne added that that is the case at HubSpot, where they have built “an internal set of localization tools into which DeepL is deeply integrated.”

Slator’s Alex Edwards, Senior Analyst, moderated a group discussion that started with the premise that LSPs are under pressure to offer all language services “or die by API,” in the words of main panelist Manuel Herranz, CEO of Pangeanic

Herranz explained that “die by API” referred to businesses wondering how they could leverage AI technology and how to connect via API to something that could potentially be very good at translation. He added that LSPs came to understand AI not as machine translation, but as automation, and that those who created it don’t really care much about the translation industry.

Chris Menier, CEO of Transifex, discussed the challenges and opportunities in modernizing the localization process, highlighting its current cyclical and inefficient nature, “filled with human touchpoints and creating synchronization issues between source and localized content … I posit that we can change this journey and make it more efficient with the use of AI, without sacrificing quality.”

Julia Cassidy, Head of Global Experience & Localization at SurveyMonkey, highlighted the firm’s current approach to localization, focusing on local market expertise and utilizing a translation management system (TMS) to manage various content types and complexities.

A panel discussion moderated by Slator’s Head of Advisory, Esther Bond, brought together Wada’a Fahel, Senior Director of Engagement Marketing-Transcreation and Global Ops at Zendesk; Katherine Elkins, Author, Researcher, Director and Professor at Kenyon College Digital Humanities Lab; and Gabriel Fairman, Founder and CEO of Bureau Works.

The panelists talked about their own experience and interest in generative AI, which Fahel described as being “inevitable.” They were all in agreement on the advantages of AI for speed and scalability in localization, recognizing the role humans play in shaping the interaction with the technology.

Shivali Goel, AI/Machine Translation Lead at Adobe, closed the conference with an account of Adobe’s localization journey, the learnings along the way, and where things are going. 

In what Goel described as the expansion phase at Adobe, around 2019, the realization was that the globalization team was small, and there were many products that needed support and localization. “We are seeing a shift now, moving towards more self-help models and moving to … continuous integration, continuous localization, and continuous development,” observed Goel when referring to how translation is integrated in the software development cycle.

If you missed SlatorCon Remote March 2024 in real-time, recordings will be available in due course via our Pro and Enterprise plans.