Intertrust develops secure content & data platforms used by some of the most demanding media and technology companies in the world. So when we first got on a call with Technical Writer Liene Egle, there was a sense that the technical capabilities of a partner are equally important to what we can do linguistically.
When the company decided to pilot its product suite in Japan and Germany, speed and accuracy were paramount: 94 000 words of UI strings and technical documentation had to be shipped in both languages within a four‑week sprint.
Until then, localization was handled ad‑hoc by freelancers and multilingual employees. The method was slow, terminology drifted, and release managers lost time fixing markdown and *.json formatting before each launch. Intertrust needed a repeatable, engineering‑friendly process that could scale far beyond two languages.
The considerations on day one
From the outset, we knew success hinged on precision and pace. So we kicked things off with a real-world translator calibration phase: three teams per language tackled the same content samples, and Intertrust’s in-country reviewers handpicked the strongest performers. The selected teams took the lead, while the others remained on standby—ready to absorb overflow without delay.
To manage two radically different content types—UI strings in *.json and technical docs in *.md—we implemented a dual-platform setup built around SDL Trados Studio and Crowdin. It gave us the flexibility we needed, letting our translators work across formats while maintaining automatic synchronization between translation memories. UI and documentation evolved together, ensuring consistency in tone and terminology.
But our biggest engineering leap came from customization. Markdown, by nature, doesn’t play nicely with many CAT tools. We built a bespoke ruleset in Trados and paired it with Python scripts designed to cleanly parse markdown syntax, safeguard tags, and flag issues before they reached Intertrust’s QA team. These tools surfaced everything from broken code blocks to *json key collisions—problems that previously delayed launches.
To keep the quality bar high and the workflow lean, we added lightweight AI agents that scanned translations for placeholder mismatches and formatting anomalies. On average, these automations shaved two full days off every release cycle—time that used to be spent manually fixing deliverables before go-live.
By the second sprint review, markdown breakages had fallen to zero and reviewer terminology edits dropped by 62 %. Our tems were able to deliver localized versions that were ready for immediate publishing.
“In our business, no news is good news—and the silence after each drop told me we’d cracked it.”
Intertrust’s localization is now baked into their CI pipeline, turning translation into an early market‑testing lever rather than a post‑build scramble.
With the playbook operational, Intertrust is is equipped with smooth update delivery and access to testing new markets at any moment they decide to do so. The best thing about it, it neither breaks the code, nor the bank.
Native’s role as a localization vendor has shifted from fire‑fighter to long‑term partner—tuning TMs and QA agents so that “flawless” remains the default.
Localization is a competitive edge that your competitors are likely not capitalizing on.
There is a huge market gap for products that do not yet speak the language of their customers.
Let’s talk about turning your content into a competitive edge—before your next release clock starts.
© Native 2024