Partnering with localization teams to achieve internationalization compliance on time every time means working closely together – especially as your processes and theirs integrate (Gen)AI. Open an ongoing dialog with localization, testing, and design colleagues. They are sure to have other suggestions for how to close the gap between you and the multilingual products that grow out of original coding. And, if you let them, the localization team may even offer to generate your developer notes automatically – a definite win for both teams.

Based on 53 in-depth interviews with organizations that have implemented a continuous localization model, here are four actions that your team can take to better support localization colleagues.

•     Elevate localization colleagues to join in product design. Help them shift left by carving out a role to participate – either from the beginning or at critical junctures – for sign-off on large language models (LLMs), UI design, and code iteration. This will benefit your own team by ensuring that internationalization bugs never occur in the first place. It will also help avoid messy and expensive code rework by enabling you to architect for local marketing requirements up front.
•     Collaborate to determine how continuous localization needs to be. You can be as fast as you want, but understand what your speed will cost the organization as a whole. Address the question of running localization in parallel with development. Interview participants were split on the issue of starting localization before UI is complete. Some firms find that it’s worth it after analyzing the waste. However, others wait because they believe that too many things can go wrong or will require rework. At the end of the day, you and your colleagues must choose the route that delivers high-quality code to your customers the fastest without stretching development, testing, content, or localization people or budgets beyond their limits.
•    Support localizers to build automation roadmaps. Ensure that infrastructure and technology function efficiently for your localization colleagues – and for your team. For example, the ideal solution may be to call on a function that encapsulates the text to be translated, along with all comments, instructions, and screenshots matched with unique string identifiers. This is a tall order and may not be totally possible at first, but work backwards from the end goal to see what you can achieve.
•     Recognize that context is a huge roadblock for efficient localization. A translator or localization program manager may have to be able to see the path and associate the name of the screenshot to the key – for example, when there are several “cancel” buttons, but not all are to be translated with the same phrase. The translation may depend on the meaning of the rest of the words around it – its context. As one interviewee describes it, “If you can’t link string IDs back to specific developers or where they appear in context on screen, it’s like sorting rice grains.”

If you approach internationalization compliance on the same level as compliance for security and accessibility, you shouldn’t have to think much about localization. Why? Because your code will be clean and internationalization-compliant from the start. Just make sure to: 1) Support localization teams’ participation in product and LLM design; 2) maintain infrastructure (including version control systems); and 3) automate screenshots and context information for use by testing and localization teams.