Don’t Stick To The Script: How To Make AI Feel More Like a Colleague Than a Tool

A Conversation with Sergio Llorens González

This special edition of the AI in Loc Newsletter celebrates a new chapter for one of our industry’s most accomplished localization leaders. On this Monday, November 3rd, Sergio Llorens González begins his new role at SAS Software, marking the next step in a career that has taken him from the world of video games to some of the largest technology companies on the planet.

Over the past fifteen years, Sergio has built and led localization programs at Big Fish Games, Amazon, Huawei, and Meta, bridging content, technology, and people across dozens of languages and markets. In this conversation, he shares what those experiences taught him about scaling global content, uniting teams, and how language itself can shape business outcomes.

We connected just days before his first day at SAS. What follows is a candid and insightful discussion about building localization from scratch, empowering linguists, navigating corporate complexity, and embracing AI as a natural evolution of language work.

Restarting from Zero

Stefan Huyghe: You’ve built and led localization teams and workflows at tech companies from Big Fish and Amazon to Huawei and Meta, and now you’re about to join SAS. What did each of these experiences teach you about starting from zero?

Sergio Llorens González: That’s a big question. My main lesson, and something I often share with colleagues, is: don’t rush. When you join a company that’s already trying to “do something with languages,” everything suddenly becomes urgent. But the first thing you need to do is slow down and look at the business.

Connect with the departments. Understand who they are, what they do, and what they actually need. Write it down, cover all the W’s: who, what, when, where, why, and how. Once you truly understand the priorities of each content type, you can start aligning localization with the company’s goals. That’s how you integrate yourself properly rather than running around putting out fires.

Once that’s clear, talk with upper management and every content owner. Draft a clear proposal for the support you’ll provide, define the workflow, how content will reach you, delivery times, costs, team setup, and success metrics. Only then start small: a few languages you can manage well and a small in-house core team to orchestrate.

That’s really where localization is going now, it’s more about orchestration than translation.

Stefan Huyghe: How do you decide whether to build localization in-house or outsource it?

Sergio Llorens González: You have to evaluate not just volume but also the value of in-house expertise. Someone internal understands the product, knows its quirks, and can react instantly, whether that’s editing live text or defining quality standards.

If you outsource everything, you create a wall between your company and the LSP. They’ll never know your product as deeply as someone on the inside. My advice is to always keep a core internal team, people who drive quality, maintain knowledge, and push both in-house and external efforts forward.

Building a Global Volunteer Network

Stefan Huyghe: Looking back, which setup that you built are you proudest of, and why?

Sergio Llorens González: The one I’m proudest of is also the most unusual and unorthodox: the Amazon Language Volunteer Program. It’s an internal network of more than two thousand volunteers covering over thirty languages.

When I joined Amazon in Spain, I was the only localization professional in the region. New teams were popping up everywhere, Prime Video, Audible, Kindle, Alexa, each growing fast and needing language support. People were improvising: sometimes Google Translating, sometimes asking a native speaker for help until that person left. There was no structure, no unified approach

So I created a volunteer program with an internal wiki and distribution lists where people could reach out when they needed help. Suddenly we became a localization radar, we knew what was happening globally, who was translating what, and where support was missing.

This visibility allowed us to guide teams toward proper workflows, bring them into our localization platforms, or even recommend hiring a localization PM if their volume justified it. It helped Amazon’s many teams act as one global community instead of isolated silos.

Tools, Workflows, and Simplicity

Stefan Huyghe: That’s fascinating. Who could have predicted that open communication would become such a central pillar in localization? In the early days, we were limited by technology. Speaking of that, how does technology selection play into building these programs?

Sergio Llorens González: A lot. When choosing a TMS or designing workflows, I always think about inclusion, how to involve everyone who touches content, not just translators.

A good TMS should bring together content creators, reviewers, PMs, and freelancers. It should centralize everything: style guides, terminology, context, instructions, so nobody needs to dig through email threads.

I also think about the freelancer’s perspective. They work alone, for multiple clients, across different topics. The simpler and clearer your system, the better the results. So my priority is simplicity, centralization, and communication, making everyone’s work easier and the final product better.

Scaling Localization Without Losing Control

Stefan Huyghe: What mistakes do companies make when they rush into building localization without a strategy? And maybe, what are some mistakes you’ve made yourself that others could learn from?

Sergio Llorens González: The biggest mistake I’ve seen, and one I’ve learned from myself, is scaling too early. You have to build structure before you build size.

If your workflows, reporting, or language materials aren’t clearly defined, every team and every language manager will start doing things their own way. That’s chaos. So first, standardize your methodology, your processes, your style guides, your quality assessment. Once that foundation is solid, then scale.

Another common mistake is not paying enough attention to the source text. That’s our raw material. If the source is unclear or inconsistent, those problems multiply across every target language.

I always spend time with content creators, helping them write unambiguous text, use consistent terminology, and think ahead for translation. The more we collaborate upstream, the smoother everything goes downstream. In the end, localization is a team effort, not an isolated function.

When Localization Drives Engagement

Stefan Huyghe: That’s such an important point. People in charge of localization really do need to be linguistically aware. What product launch stands out as the most complex localization challenge you’ve ever faced

Sergio Llorens González: One I’ll never forget was at Amazon, when we rebranded Amazon Premium to Amazon Prime in Spain, France, and Italy. It was a huge operation with no clear owner, teams had shifted, people had left, and some of the content was orphaned.

I suddenly found myself responsible for hundreds of strings, scattered across systems, some implemented, some not, with no one quite sure where everything lived. So I took ownership, combed through each string manually, and updated them one by one.

It was hectic, but it taught me the importance of ownership in localization. Sometimes, you just have to step up and fix it yourself.

Looking back, I can’t help but think how AI could have helped, flagging brand names, distinguishing between “Prime” as a term versus as a brand. But at the time, it was all manual. I remember finishing the updates late at night just before the change went live.

Stefan Huyghe: So much of our industry has been about finding ways to gain traction outside the translation bubble, to get a real seat at the table. Can you remember a time when localization decisions had a direct impact on product results or engagement?

Sergio Llorens González: Yes, absolutely. One clear example comes from my time at Amazon with Kindle Direct Publishing (KDP). That’s the platform where independent authors upload and publish their books on Amazon. It has a big help section, a technical setup process, and all the steps for uploading, pricing, and releasing a book.

Every year, our marketing teams ran a contest for independent authors, whoever’s book was read the most would win prizes, interviews, or even a publishing deal. But here’s the crazy part: on the KDP platform itself, there was no mention of this contest at all. None.

Marketing was doing their thing, landing pages, social posts, campaigns, and product was focused on the platform experience. But authors who were already on KDP had no idea the contest even existed.

I decided to dig into the data. I looked up search terms in the help pages and saw thousands of authors searching for “contest,” “writing competition,” and similar phrases in multiple languages. Then I checked with customer support and learned they were getting a flood of inquiries about it.

So I went to the leadership team in Seattle and said, “We’re missing something big here.” We created a help page explaining the contest and added dynamic UI pop-ups in the publishing flow, simple reminders like, “If you’re entering the contest, remember to add this keyword or meet this price condition.”

The results were immediate: conversion rates went up, support requests dropped, and participation soared. That’s when it really hit home, localization isn’t just about language; it’s about understanding user intent and connecting the dots across teams.

Multilingual Empathy and Cultural Intelligence

Stefan Huyghe: That’s a great example of how localization can drive engagement. You, of course, speak an impressive number of languages, far more than me. How has language itself shaped the way you run localization?

Sergio Llorens González: Knowing several languages changes everything. Most people assume that as a translator or localization lead, you just need to master your source and target languages. But the truth is, every additional language gives you perspective.

When you’ve studied multiple language families, Romance, Germanic, Slavic, you start anticipating where things will break. You can spot issues that might not affect your language but could create problems elsewhere, like declension, word order, or gender agreement.

So even if you’re not fluent, knowing the basics helps you raise the right questions early. It helps you help others, and it connects you more personally with your colleagues.

Stefan Huyghe: I completely agree. I also think that speaking languages from very different families gives you a broader perspective. Maybe Romance languages share certain issues, but those are different from what you face in Slavic or Asian languages.

You were a translator earlier in your career, did that experience make you more empathetic toward translators and their challenges? Did it make you a better manager?

Sergio Llorens González: I like to think it did. I’ve been lucky to experience localization from every angle, freelancer, in-house translator, tester, and later, program builder.

Working as a freelancer gave me empathy. I know what it’s like to sit at home, working alone, juggling multiple clients, dealing with inconsistent instructions. It’s why I always strive to simplify communication and centralize resources for the people I manage.

I also ask for feedback, what’s working, what’s not, where they struggle. In some cases, freelancers didn’t fully understand the product or context of what they were translating, and that’s critical. So I started doing briefings and open Q&A sessions with them, showing my face, inviting questions, breaking the wall between project managers and translators.

My message is always: “I’m here for you, and you’re here for me.” The better your experience, the better my results. We’re partners.

Stefan Huyghe: That empathy really shows. And I think when you speak many languages, you also become more sensitive to the cultural and political sides of language. In my home country of Belgium, for example, language is deeply tied to identity. It’s the same in Catalonia. How do you handle those cultural and political dynamics in localization?

Sergio Llorens González: You have to treat every language, regional or not, with the same respect you give to any major market. Look at it from a cultural relevance standpoint.

When I was at Amazon, we decided to localize a few campaigns into Catalan. One of them was for Sant Jordi, Saint George’s Day, the patron saint of Barcelona. It’s a huge local celebration, and it would have been tone-deaf to publish content about it in Spanish only.

So we supported marketing in creating Catalan versions for those key events. It wasn’t about politics, it was about experience. For users in Catalonia, it made all the difference.

Once you decide to localize into a language, you have to commit to it long-term. You can’t start and then pull back later, that would frustrate users and hurt trust. So treat local languages like you would any major language launch: strategically, respectfully, and consistently.

Linguists, Business, and the Rise of AI

Stefan Huyghe: Do you think linguists make better localization managers, or is business experience more important to succeed in this field?

Sergio Llorens González: Both matter, but if I had to choose, I’d say language experts make better localization managers. It’s harder to teach someone the nuances of language work, the tools, the terminology, the quality side, than it is to train them in business reporting or project metrics.

A strong linguistic foundation gives you a better sense of context and impact. You understand why something matters before you learn how to measure it. So yes, I’d root for the linguist to lead the team.

Stefan Huyghe: You once compared learning CAT tools to integrating AI. Why do you say AI is easier?

Sergio Llorens González: Because we’ve lived through so many waves of technology already. We’ve seen the arrival of CAT tools, then the TMS era, then automation, and now AI.

Back when Trados first appeared, everything was manual. You had to study how to create translation memories, parse files, update terminology, manage formats, it was clunky and time-consuming.

AI, by contrast, feels like a colleague you can talk to. You ask questions, check your work, and get help instantly. It’s more intuitive.

Stefan Huyghe: Wouldn’t you agree that “speaking LLM” has become a new language skill in itself? I’ve found that the more I use large language models, the better I get at communicating with them, like learning how to work with a colleague.

Sergio Llorens González: Exactly. That’s the only real skill you need to master now. You don’t have to remember where to click or how to format files. You just need to speak clearly and naturally, something linguists are already trained to do.

It’s like learning a new dialect of communication. The more precise you are, the better your results.

Stefan Huyghe: You see what behaviors work, what communication styles get the best results, how to give instructions, just like learning to work with a new colleague. This really is a language of its own.

The Future: AI, Community, and Constant Reinvention

Of course, our industry, like every other, is changing fast because of AI. Which old localization practices do you think will disappear completely?

Sergio Llorens González: Several, especially those that involve long, manual text assessment and rewriting.

For example, I once had to standardize hundreds of Spanish texts that mixed usted and tú, formal and informal styles. That kind of work used to be incredibly tedious: reviewing every sentence, watching for verb conjugations and tone. AI can now do that in seconds, analyzing and rewriting while preserving meaning.

The same goes for keyword detection or content moderation. When I worked at Huawei, we had to flag swear words in user comments, sometimes disguised with numbers or symbols. It required human intuition. Now AI models can identify those patterns much more reliably and at scale.

AI won’t replace us entirely, but it will take over the heavy lifting, the repetition, the cleanup, the drudgery, so we can focus on the higher-value work that requires judgment and empathy.

Stefan Huyghe: Let’s close by talking about the human side of all this, the worldwide localization community. Both of us have built strong networks over the years. What role has your global network played in your career?

Sergio Llorens González: A huge one. Honestly, most of my career moves came through connections, not cold applications. Someone referred me, or knew a hiring manager, or mentioned an opportunity that fit.

Networking doesn’t guarantee a job, you still have to prove your worth, but it gives you a chance to be seen. This industry runs on trust and reputation. Helping others, sharing knowledge, and staying connected always comes back around.

Stefan Huyghe: If you could mentor your younger self, the early-career Sergio just starting out in localization, what advice would you give him?

Sergio Llorens González: That’s easy: don’t stick to the script.

Don’t just do what your job description says. Be curious. Explore other parts of the process. When I started, I thought video game testing or dubbing would always work the same way, it didn’t. Everything evolves, fast.

So keep learning. Stay genuine. Help others. Think big. You’re not a small cog; you’re part of a larger mechanism that only works when we all grow together.

Stefan Huyghe: That’s beautifully said. I couldn’t agree more. Being valuable to others really is the key. Sergio, thank you so much for taking the time to share your story and your insights, and congratulations again on your new role at SAS. I look forward to buying you that coffee in person soon.

Sergio Llorens González: Thank you, Stefan. It’s been a real pleasure. Talk to you soon.