Speed, scale & savings – Uber’s India based tech team helps cut down app localisation time and costs to 25%

Uber’s Bangalore-based global scaled solutions (GSS) tech team enabled launching operations in new cities across the globe much faster by cutting down the time and cost needed to localise the app to regional languages to a fourth of what it was earlier, backed by industry-leading machine learning and AI. This added efficiency has helped Uber localise the app in many more languages and markets.

Localisation of the Uber app to regional languages enables drivers, riders, and in general platform users to understand and navigate Uber content on mobile and web or any medium with ease and comfort. The GSS teams working on localisation are based out of Uber’s tech centres in Bangalore and in the US.

The localisation process earlier involved hiring bilingual agents and feeding days of inputs into systems to translate the Uber app into local languages, which is now done through a combination of Generative AI and monolingual agents, expanding Uber’s language processing automation. As much as 85% of the localisation process is touchless now, and doesn’t require human intervention, with a target to reach 90% touchless localisation by December.

The team is building its own customised large language models now, and has brought the error margin next to naught, with a need for human intervention in the translation process only under very exceptional circumstances.

The Uber app is now available in a total of 63 languages globally, including 8 in India alone.

Commenting on the development, Megha Yethadka, Senior Director – Program Management, and Head of GSS, Uber said, “Our mission for localisation is to make our products feel local to everyone, everywhere and we are constantly innovating on our internationalisation/localisation tech and operating models to enable magical experience for the users. This helps customise the app in local and regional languages to help earners and customers in new markets navigate through it with ease, all in a matter of few hours in many cases”

An example of this was in Ukraine and Poland, where the team quickly stepped in to translate key communication and offer help and safety for Ukrainians within 24 hours as Uber provided free rides from border areas between the two countries.

The team also enabled translation of the app to Punjabi in the US and Canada within days to help truck drivers driving on Uber Freight, many of whom hail from Punjab.

A large part of Uber Globalisation team is based out of India. India hosts the 2nd largest concentration of localisation and globalisation employees outside the US who work on managing the process of creating style guides for 60 languages, driving the localisation testing, and expanding our language processing automation.