Luisa Serra’s Post

Quick RAG question: If you have a multilingual dataset in OpenSearch (English as the base articles + translations in other languages), and all articles are semantically chunked + embedded, what’s the best retrieval strategy for multilingual semantic search given that users query can be in any language and the translated articles in the open search chunks are limited except from the root articles “English”? Option A: • Always search only the English embeddings (base articles). • Even if the user queries in German/Spanish, run semantic search against English chunks. • Then map results back to the user’s language via metadata (root_article_id, translation fields). Option B: • Search all embeddings across all languages directly. • Let the multilingual model handle cross-language similarity. • Then deduplicate using the root_article_id to avoid returning duplicate translations. Ssssooooo my question.. stay with me here: Which do you think ranks better for relevance and language preference and why? Curious how others are handling this in real systems 👀… Especially with models like Snowflake Arctic / multilingual-MiniLM / LaBSE

Your customers speak 110 languages. So why are you still only marketing in one? 70%+ of your potential customers won't buy from you if you don't speak their language. That stat should terrify you. Because while you're reading this in English, your competitors are closing deals in 110 languages. The math is brutal: English-only website = 27% of the global market. That's it. You're leaving 73% on the table. Not with translators charging $0.25/word. With AI that keeps their brand voice intact. Weglot's custom language model changed the game: → Learns YOUR tone, not generic translation → Trains on YOUR content and edits → Remembers YOUR glossary forever → Gets smarter every time you review Real results from real brands: • Volant: +39% international revenue • Polaàr: 100 hours saved • HBO & Decathlon: Scaled globally The setup? 10 minutes. The impact? Almost immediate and lasting. Your sarcastic product copy stays sarcastic in Spanish. Your no-BS emails stay no-BS in German. Your personality travels. Stop thinking "someday we'll go global." Your customers already are. P.S. Every day you wait, someone's buying from your competitor because they bothered to say "hello" in their language.

𝐘𝐨𝐮𝐫 𝐜𝐮𝐬𝐭𝐨𝐦𝐞𝐫𝐬 𝐬𝐩𝐞𝐚𝐤 110 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞𝐬. 𝐒𝐨 𝐰𝐡𝐲 𝐚𝐫𝐞 𝐲𝐨𝐮 𝐬𝐭𝐢𝐥𝐥 𝐨𝐧𝐥𝐲 𝐦𝐚𝐫𝐤𝐞𝐭𝐢𝐧𝐠 𝐢𝐧 𝐨𝐧𝐞? 70%+ of your potential customers won't buy from you if you don't speak their language. That stat should terrify you. Because while 𝐲𝐨𝐮'𝐫𝐞 𝐫𝐞𝐚𝐝𝐢𝐧𝐠 𝐭𝐡𝐢𝐬 𝐢𝐧 𝐄𝐧𝐠𝐥𝐢𝐬𝐡, 𝐲𝐨𝐮𝐫 𝐜𝐨𝐦𝐩𝐞𝐭𝐢𝐭𝐨𝐫𝐬 𝐚𝐫𝐞 𝐜𝐥𝐨𝐬𝐢𝐧𝐠 𝐝𝐞𝐚𝐥𝐬 𝐢𝐧 110 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞𝐬. The math is brutal: English-only website = 27% of the global market. That's it. You're leaving 73% on the table. Not with translators charging $0.25/word. With AI that keeps their brand voice intact. 𝐖𝐞𝐠𝐥𝐨𝐭'𝐬 𝐜𝐮𝐬𝐭𝐨𝐦 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐦𝐨𝐝𝐞𝐥 𝐜𝐡𝐚𝐧𝐠𝐞𝐝 𝐭𝐡𝐞 𝐠𝐚𝐦𝐞: → Learns YOUR tone, not generic translation → Trains on YOUR content and edits → Remembers YOUR glossary forever → Gets smarter every time you review 𝐑𝐞𝐚𝐥 𝐫𝐞𝐬𝐮𝐥𝐭𝐬 𝐟𝐫𝐨𝐦 𝐫𝐞𝐚𝐥 𝐛𝐫𝐚𝐧𝐝𝐬: • Volant: +39% international revenue • Polaàr: 100 hours saved • HBO & Decathlon: Scaled globally The setup? 10 minutes. The impact? Almost immediate and lasting. Your sarcastic product copy stays sarcastic in Spanish. Your no-BS emails stay no-BS in German. Your personality travels. 𝐒𝐭𝐨𝐩 𝐭𝐡𝐢𝐧𝐤𝐢𝐧𝐠 "someday we'll go global." Your customers already are. 𝐄𝐯𝐞𝐫𝐲 𝐝𝐚𝐲 𝐲𝐨𝐮 𝐰𝐚𝐢𝐭, 𝐬𝐨𝐦𝐞𝐨𝐧𝐞'𝐬 𝐛𝐮𝐲𝐢𝐧𝐠 𝐟𝐫𝐨𝐦 𝐲𝐨𝐮𝐫 𝐜𝐨𝐦𝐩𝐞𝐭𝐢𝐭𝐨𝐫 𝐛𝐞𝐜𝐚𝐮𝐬𝐞 𝐭𝐡𝐞𝐲 𝐛𝐨𝐭𝐡𝐞𝐫𝐞𝐝 𝐭𝐨 𝐬𝐚𝐲 "𝐡𝐞𝐥𝐥𝐨" 𝐢𝐧 𝐭𝐡𝐞𝐢𝐫 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞.

Trust Your Translation? Last week at LSPN San Francisco, we launched something I am truly proud of: “The BIG CheQ”. Today, many companies simply don’t know how their translations are being done. Are they being carried out using machine translation or not? Have they been told the process behind it, or how the translation engines were created and tuned? Is an LLM being used or an NMT? That’s where The BIG CheQ comes in. The BIG CheQ is not another platform, it’s a comprehensive translation audit that independently evaluates, benchmarks, and verifies translation quality. It gives companies a clear view of where they stand and how they can improve. The excitement and engagement around The BIG CheQ have been incredible. It’s clear that this is a game changer for organizations that want to bring accountability and assurance back into their language workflows. The BIG CheQ replaces uncertainty with insight and trust with governance.

One of the biggest lessons I’ve taken from great business leaders like John Legere (former CEO of T-Mobile) is that trust isn’t something you claim, it’s something you prove. That’s exactly what BIG IP & Legal Solutions - A BIG Language company and BIG Language Solutions are doing with The BIG CheQ. It brings real transparency and accountability to an area that has too often operated behind the curtain. By giving organizations an independent benchmark for translation quality, The BIG CheQ helps teams move from assumption to assurance and from uncertainty to trust. Exciting to see this launched at LSPN San Francisco last week and to be part of a team driving that change. #BIGCheQ #BIGLanguage #BIGIP #TranslationQuality #Trust #Innovation #Leadership #LSPN #AI #LanguageIntelligence

Ever notice how the smartest-sounding people often make the least sense? I stumbled across a thread a while back that ruined half of business writing for me. Because once you see it, you can’t unsee it. If someone sounds full of shit, they probably are. And the proof isn’t emotional - it’s linguistic. English didn’t start as the language of empire or intellect. It started as a working-class mashup, the tongue of farmers, fighters, and builders. But around a thousand years ago, that changed. When the Normans invaded England in 1066, they brought their French, a Latin-based language used by rulers, priests, and scholars. For centuries, the upper class spoke French and Latin. The lower class spoke English. And that divide got baked right into the DNA of the language. That’s why we say ask but lawyers say inquire. We say help but bureaucrats say assist. We say lie but officials say misinformation. The words of power stayed Latin. The words of life stayed Germanic. And to this day, that’s how you can tell if someone’s trying to connect — or control. Because Germanic words are short, muscular, concrete & hit your gut. Latin words are abstract, multi-syllabic, intellectual & float above your head. That’s not just linguistics. It’s psychology. Your brain processes simple, common words faster. You trust them more. They feel like truth. It’s why political speeches, sermons, and great copy all rely on Anglo-Saxon language. You don’t win people with “optimization of scalable frameworks.” You win them with “get more done, make more money, sleep better.” One reads like a strategy deck. The other reads like real life. That’s why the best direct-response copywriters like Schwartz, Ogilvy & Halbert all leaned hard into Germanic rhythm. They wrote for the body, not the boardroom. If you can’t say it simply, it probably isn’t true. Because the moment your message starts swimming in Latin fluff, you’ve lost the sale and the reader. Go through your copy. Circle every word over two syllables. Ask if it says something you would say out loud. Most won’t survive that test. Almost anything you can express in Latin, you can express in Germanic and when you do, it converts better. The first sounds smart. The second makes money. Language is leverage. And the fewer syllables between your message and someone’s heart, the faster they buy. So next time you write, don’t “utilize persuasive linguistic mechanisms.” Just tell the truth and make it land.

At LT4ALL in February, I listened to many discussions on how technology companies are extracting all the value from local languages. Extraction and exploitation go hand in hand, excluding local communities as these neo-colonial mega-corporations emulate the East India Company, adapting their models for digital domination. This excellent article on the Icelandic language summarizes the challenges that even wealthy indigenous speakers face against huge tech companies that push English as a homogenous digital default. Our goal is to empower communities, producing digital services that are syntactically and culturally effective, creating more options for non-English speakers. Have a great weekend #Culture #LowResourceLanguages #IndigenousLanguages

Lan-bridge has been a key player in the translation industry for 25 years, growing from a small office into an industry benchmark. Committed to quality first, we rigorously select professional translators and implement three layers of quality control, earning the trust of our clients and evolving from a translation vendor into a strategic language partner. Leveraging technology to optimize services, Lan-bridge builds terminology databases with CAT tools and adopts a collaborative workflow of AI pre-translation plus human refinement to enhance both efficiency and accuracy. Our intelligent systems provide full-process visibility, while big data enables continuous service iteration. Looking ahead, Lan-bridge will continue to uphold its commitment to quality, embrace technological innovation, deepen customer-centric solutions, cultivate interdisciplinary talent, and strive to build a language bridge connecting the world.

The Golden Translation Strategy – Clear, Systematic, and Reliable This strategy moves through two deliberate stages. In SL1 → SL2, the translator engages in deep analysis of the source text—its structure, syntax, and meaning—then mentally reformulates the ideas in the target language to prevent literal transfer and structural interference. In TL1 → TL2, a first draft is produced, then refined into a polished version through revision that enhances coherence, naturalness, and stylistic quality. This method is “golden” because it mirrors the natural cognitive path of understand → rethink → draft → refine, ensuring translations that are accurate, readable, and methodologically sound. It also provides trainees with a disciplined, step-by-step framework that reduces both linguistic and meaning-related errors. Mohamed El-nemr

𝐓𝐡𝐞 𝐭𝐡𝐢𝐧𝐠𝐬 𝐀𝐈 𝐰𝐨𝐧'𝐭 𝐝𝐨 𝐟𝐨𝐫 𝐲𝐨𝐮 (𝐛𝐮𝐭 𝐚 𝐡𝐮𝐦𝐚𝐧 𝐭𝐫𝐚𝐧𝐬𝐥𝐚𝐭𝐨𝐫 𝐰𝐢𝐥𝐥!) Sure, AI can translate text from one language to another. But here's what it won't do, and what the best human translators bring to the table: ✨ They're 𝐲𝐨𝐮𝐫 𝐬𝐞𝐜𝐨𝐧𝐝 𝐩𝐚𝐢𝐫 𝐨𝐟 𝐞𝐲𝐞𝐬 𝐨𝐧 𝐭𝐡𝐞 𝐬𝐨𝐮𝐫𝐜𝐞 𝐭𝐞𝐱𝐭. Spotted a typo or grammar mistake in your original? They'll flag it before you publish it on your website for everyone to see. ✨ They explain their choices. Need to know why they chose one term over another? Great translators will walk you through their reasoning with 𝐜𝐥𝐞𝐚𝐫, 𝐥𝐨𝐠𝐢𝐜𝐚𝐥 𝐚𝐫𝐠𝐮𝐦𝐞𝐧𝐭𝐬. ✨ They give you options. When there's more than one good solution, they'll often present 2-3 alternatives and 𝐥𝐞𝐭 𝐲𝐨𝐮 𝐜𝐡𝐨𝐨𝐬𝐞 𝐰𝐡𝐚𝐭 𝐟𝐢𝐭𝐬 𝐲𝐨𝐮𝐫 𝐛𝐫𝐚𝐧𝐝 𝐛𝐞𝐬𝐭. ✨ They catch the formatting issues. Broken links, missing spaces, inconsistent formatting that would look terrible if published? 𝐓𝐡𝐞𝐲 𝐬𝐩𝐨𝐭 𝐭𝐡𝐞𝐦 𝐚𝐧𝐝 𝐟𝐢𝐱 𝐭𝐡𝐞𝐦. ✨ They 𝐟𝐥𝐚𝐠 𝐚𝐦𝐛𝐢𝐠𝐮𝐢𝐭𝐢𝐞𝐬 before they become problems. If something in your source text could be interpreted multiple ways, they'll ask for clarification rather than guess and risk getting it wrong. And let's not forget the truly irreplaceable part: 𝐭𝐡𝐞 𝐰𝐚𝐫𝐦 𝐡𝐮𝐦𝐚𝐧 𝐜𝐨𝐧𝐭𝐚𝐜𝐭. That friendly "How was your vacation?" at the start of an email. A genuine smile in a video call. The things that make work feel less like work and more like collaboration between actual people. What are the little extras you appreciate in your industry? #Translation #Localization #LanguageServices #ServiceOriented