F
rom an upbringing in a trilingual home to finding himself in a Zen Buddhist monastery and then stumbling into an unexpected language career, Bureau Works CEO Gabriel Fairman’s life has rarely followed a straight line. Fascinated by the philosophies, attitudes, and ideas that have guided civilization, he eventually found himself working with languages — a binding element at the center of all those human concerns.
Now, that central fascination with humanity informs his work guiding Bureau Works, a company that presents itself and its translation management system (TMS) with the openness of a friend at the local pub. It’s a refreshing approach in a field grappling with the implications of artificial intelligence (AI) and machine assistance, and Fairman took the time to tell MultiLingual all about it.
To begin, could you tell us how you found yourself in this line of work? Was language work always in the picture, or did you discover it somewhere along the way?
I was born and raised in São Paulo, Brazil. I grew up speaking Portuguese, the native language of Brazil, with my mother; Spanish with my father, who was from Argentina; and English because I went to an American school.
I was always involved with languages. My mother was a translator, and I used to watch her type out her translations on a typewriter when I was a little boy. I also watched her transition to using a computer, a Toshiba, when I was around seven years old.
I began helping my mother review some texts and proofread when I was around 13. I studied Simplified Chinese in college and lived in southern Taiwan for a while. I also studied a little bit of Italian and lived in Italy for approximately six months as part of a college exchange program. I picked up a little bit of French along the way from friends.
So, yeah, language was always in the picture. In college, I studied the death and rebirth of human agency — that’s my major. It focuses on the intersection of psychology, philosophy, and interpretation theory, especially on how our awareness of language shapes our ability to make true decisions. I was always fascinated by linguistics, the nature of language, awareness, and the philosophy of language.
I ended up starting Bureau Works more by accident than anything else. After college, I was very lost. I lived in a Zen Buddhist monastery for nearly half a year and was really unsure of what I was going to do next. Disillusioned by academia, I decided to start translating, and that was the beginning of where I am now.
Your background really drives home why Bureau Works has a company culture that emphasizes personality and quirkiness over a sterile, businesslike presentation. Could you tell us how that shapes your day-to-day work experience?
Even though I admire what large organizations can accomplish, I think most corporate environments don’t favor honesty. That’s just my personal perspective, based on my experience working in corporate settings. I believe it’s often about putting on a façade, playing a game, and revealing very little of who you truly are. It’s a necessary condition of large-scale organizations, but it just isn’t for me.
I feel that by lowering my guard and being more authentic, I invite others to do the same. That creates an environment where people can be more human. It doesn’t mean that it’s just peace and love, though; we are insanely committed to excellence. It’s just that we don’t have to lose sight of our humanity as we go about it.
People have good days and bad days. People go through things like heartbreak, the loss of loved ones, and other challenges. In my opinion, being human is very hard. There’s a lot of suffering involved. But when you’re surrounded by people who are truly committed to the same ideals, it makes a difference.
The key thing is that most companies hire people for specific jobs. Employees are hired to perform tasks, and it often feels transactional. People are paid to do something, whether they like it or not. If things are going well, they might be well-paid; if not, perhaps not as much. It’s usually just a transaction.
Do you think that approach to work informs the way your customers experience the Bureau Works platform?
Yes, absolutely. Our team members are deeply invested in their work because they share certain key values. They’re committed and passionate, and they genuinely want to see our customers happy.
For instance, the last time I checked, our customer support satisfaction numbers were at 99%. We don’t have any policy that punishes people for not achieving flawless customer support, but it’s implicit. Our team wants to ensure that customers have successful experiences on our platform.
I think company culture plays a key role in this, especially in terms of listening and continuously improving — not just to close deals, but to make sure our users stay excited and happy about our product. I believe it also influences how our engineers think. They’re more prone to innovative thinking because they’re working from a deep love for developing cool software and solutions.
On that note, language work often requires a high degree of creativity as linguists attempt to capture the essence of a text in its original language. Do you think that Bureau Works’ approach to business helps nurture and encourage that creativity, and if so, how?
That’s a controversial question because our approach strongly supports the collaboration between machines and humans. Typically, especially among translators, machines are seen as enemies of creative translation. However, we believe in “agentic translations,” which represent the next step in the evolution of translation, particularly with the advent of generative AI (GenAI).
GenAI has brought a significant paradigm shift, but I don’t think most people fully understand its impact. Many see it as an enhancement to machine translation (MT), improving accuracy through large language models and deeper contextualization, which is true to some extent. But we see it differently.
By using GenAI to listen carefully to how humans work and think, we unlock a new level of creativity through collaboration with machines. Typically, people think working with MT limits creativity compared to working freely. In my experience, though, when I work with our context-sensitive translation — which leverages GenAI — the engine “learns” from me, lifting much of the heavy linguistic work off my shoulders. This frees up my mind for more creative thinking.
There’s a common belief that MT is great for support content and technical literature but terrible for creative copywriting. I don’t think that’s true anymore. GenAI can offer translators multiple perspectives at once, which is invaluable for creativity. Creativity, after all, is about seeing things from different viewpoints and making critical decisions about which perspective makes the most sense.
Our software enables this kind of nuanced creative expression while working closely with machines, which is the essence of “agentic translation.” The idea is that you’re no longer translating on your own. You’re working with AI agents that guide, coach, and flag potential issues, creating a much richer experience. Language is about seeing the same thing from different angles, and this technology allows us to focus on delivering the true intent behind the content, not just on getting it into another language.
I find this approach beautiful, though I admit it’s controversial.
Likewise, linguists are often required to think more analytically or academically. How does Bureau Works accommodate those ways of working?
Adopting more analytical and academic ways of thinking requires either deep specialization in areas like patent law or organic chemistry, or exceptional linguistic versatility. It’s like a chef knowing when to add that perfect pinch of salt — small adjustments can make a huge difference.
Our platform is designed to support both subject matter and linguistic expertise. By reducing the “edit distance” — the number of changes a translator needs to make — we’re seeing incredible results. For example, with English-to-Spanish translations, we’re down to a 5.5% edit distance, meaning only 55 words out of 1,000 need to be changed. This low edit distance makes every adjustment more significant and reduces the need for extensive rounds of proofreading.
Our platform also includes semantic runtime verification, which acts like a “sidekick” for translators, flagging potential mistakes, omissions, or awkward phrases in real time. This reduces anxiety in the translation process, making it easier to focus on quality. In the past, I often felt anxious about whether I was working fast enough or making silly mistakes. Now, with the help of these tools, I feel more relaxed and ready to be creative and precise.
Most people underestimate the cognitive burden we carry daily. Translators face an enormous mental load, constantly juggling details from previous segments or documents, glossaries, and translation memories (TMs). This burden can hinder our ability to make critical decisions. In my opinion, less cognitive load leads to better decisions. Our goal is to free up mental bandwidth, allowing translators to focus on the important, creative aspects of their work.
When we free up mental bandwidth, great things happen. Just like returning from a vacation with a clear mind, reducing unnecessary cognitive load allows for sharper thinking. That’s what we aim to do at Bureau Works.
The simple feel of a platform factors so much into how someone experiences their day-to-day work. How does Bureau Works create an experience that its own employees find enjoyable, given its less-formal atmosphere?
As for the platform itself, its design, responsiveness, and user interface (UI) all contribute to the day-to-day work experience. We aim for a clean, crisp UI and try to make things as simple and intuitive as possible. We definitely believe in minimalism.
Our engineers are incredibly creative, always rethinking the process from first principles to drive the best user experience. While existing concepts like TM and term bases are useful, they can also lock us into outdated ways of thinking. We have to balance industry standards with innovation to create software that makes sense now and for the future.
If we focus only on how the industry has done things, we’ll build decent software, but we won’t have much fun or challenge ourselves. On the other hand, if we get too abstract, we’ll end up with something that may appeal to a niche group but isn’t practical for most users. We aim to bridge that gap by blending innovation with industry needs, always guided by the principle that “less is more.”
Another key factor that impacts not just the responsiveness of the platform, but also our ability to innovate, is our commitment to consistently refactoring code. Why? Because the goal isn’t just to hit specific targets — it’s to create great software. And great software often requires a significant amount of refactoring as we’re constantly learning new things.
That learning needs to be integrated into what we’ve already built, or we need to adapt our existing systems to reflect our current understanding. It’s an ongoing process of aligning where we are now with what we’ve previously built. I don’t think we consciously think about it all the time, but from a business perspective, frequent refactoring can become expensive quickly.
However, our approach isn’t based on short-term cost-benefit analysis. Our focus is long-term, building what we believe to be the best TMS available. Of course, “the best” is relative — it’s the best based on our understanding of what matters in localization. It doesn’t mean it’s categorically the best; that’s always up for discussion. But it’s the best for us — a system we can look at and know we’ve done our absolute best.
Our company culture supports this mindset. It encourages ideas and constructive challenges, rather than being siloed or overly focused on rigid objectives and key results (OKRs). While I understand the importance of structure in large organizations, I believe it can also make things more rigid.
Which Bureau Works features do you consider especially innovative or exciting?
What excites me most about Bureau Works is our ability to deliver results. At the end of the day, everyone wants better, faster, and cheaper solutions, and we excel at this by integrating various concepts.
For example, we have end-to-end automation. You can create a project and get costs within seconds. The platform’s responsiveness is exceptional — page loading speeds are top-notch. Once a project is approved, tasks can be assigned algorithmically to vendors, allowing them to start immediately. As tasks are completed, the next workflows kick in, payables are calculated automatically, and performance metrics feed into user profiles. These metrics are then used to drive further job placements, ensuring a continuous cycle of improvement, where you work with top performers.
Automation is great for results, but augmentation is even more interesting. Our “context-sensitive translation” uses a micro-trained engine that learns in real time from the translator’s actions. The user determines the TMs, glossaries, and MT settings for specific content. This creates a tailored experience in which the engine provides suggestions; and as the translator makes changes, the knowledge base updates accordingly.
This engine draws deeper inferences through GPT-4, noticing shifts in tone, style, and formality, and even building glossaries in real time. It’s like a guide that keeps the translator on track, which solves a major frustration with traditional MT — making the same edits repeatedly.
For example, quirky, idiomatic content in English may not translate well into other languages like Portuguese, Japanese, or German, where a more business-casual tone might be required. Adapting this content over and over can be frustrating, but our engine learns and adjusts in real time, making the process more seamless and adaptable.
What’s remarkable is how it feels like the engine is “listening” to me as I work. It’s more than just adaptive — it’s interactive, making the translation process smoother and more intuitive. I’ve even described it as feeling like a “language flow architect,” where I’m shaping the flow of language between the editor and my mind.
All the augmentation we’ve developed is the most refreshing part of Bureau Works. We’re not just focused on a GenAI tool that proofreads or fixes tags, though we have that. The real focus is on creating a completely new experience for users by integrating GenAI, making it more of a communication tool than just another feature.
You frequently cite the Bureau Works team as the greatest cause of the company’s success. How does your team work together to create a product that so many companies and professionals rely on?
I think we’re very close as a team. We don’t have mandatory office days, yet many of us choose to go in. Personally, I prefer working from home, but we do organize retreats a few times a year in different parts of the world, bringing together people from various fields like support, solution architecture, engineering, product, marketing, and customer success. We genuinely have a good time at these retreats, and I think that’s because there’s real trust and camaraderie between the team members. The positive environment and shared sense of humor naturally foster collaboration.
When teams work together, it feels natural, and I believe that stems from good governance. If everyone is aligned with the same goal — building the best possible company and software — it creates unity. Conflict arises when different departments have conflicting goals. For example, if the sales team wants a feature that isn’t on the product team’s roadmap or the security team is focused solely on their priorities, fragmentation happens. It’s harder to agree on what’s important when everyone is working toward different objectives. However, when the entire company is focused on a few key, shared goals, things get done — and done well. This unity makes the outcome more sustainable and fulfilling, even though it might take longer to achieve.
I think many companies unintentionally encourage divisiveness by siloing departments. For example, a CFO may focus on cost-cutting while a CTO is concerned with computational efficiency, and these perspectives don’t always converge. But when goals are aligned, the process feels more organic and leads to more sustainable success.
A product is only as good as the needs it satisfies for its customers. How have you applied user feedback into your platform throughout the years while keeping the core vision and philosophy intact?
Applying user feedback to our platform is challenging because everyone wants different things. Many users don’t have a broad product vision; they just know they need a specific feature to make their work easier. We respect that and listen closely to user feedback. Since we entered the market later, we had to meet established benchmarks from other platforms right from the start. That pressure to meet industry standards made us better, faster.
Once we met those basic requirements — like parsing, segmentation, TM, and glossaries — we began to open more space for our own vision. Over time, by gathering and analyzing user feedback, we’ve been able to see a bigger picture of what our software can be. Our team — especially our CTO, Director of Software Engineering, and product leads — are all very forward-thinking, always considering future possibilities while adapting to current needs.
One thing I’m grateful for is how our CTO thinks. He’s not attached to past ideas. Even as we completed Bureau Works 1.0, he was already thinking about 2.0; and now that 2.0 is here, he’s already thinking ahead. This kind of forward-thinking allows us to take user feedback and abstract from it, creating deeper changes that improve the entire platform.
The core vision remains the same: to build a platform that minimizes effort in translating content from one language to another while maximizing the human experience. But how we achieve that continues to evolve based on the needs and feedback of our users.
I also want to emphasize that with “agentic translation,” we’re moving from the age of text to that of knowledge. Traditional translation captures text but not knowledge. With agentic translation, we’re creating a system where the machine understands the meaning behind the text, not just the words themselves. This opens up new possibilities, allowing translators to work alongside AI agents that guide and assist them in real time.
This collaboration elevates the process, shifting the focus from simply transferring words to conveying the true intention behind the content. It also creates more opportunities for creativity, critical thinking, and improved performance.
Is there anything you want to add?
I know some people feel disillusioned or worried about the rise of AI in translation, and I sympathize with those concerns. But for those willing to adapt, there’s tremendous opportunity. AI can help elevate the role of the translator into something more specialized — what I like to call a “language flow architect.” For localization managers, translation agency owners, and translators alike, the potential is huge. Our software plays a key role in helping facilitate this transition into the future of translation and localization.
Related Articles
This partnership addresses those friction points of industry AI promises by providing a structural paradigm shift e.g. moving away from generic software patches toward a…
→ Continue ReadingBy integrating Bureau Works, Keylingo is able to deploy next-gen tech enablement across its entire production chain, enhancing automation, streamlining workflows, and empowering linguists to…
→ Continue ReadingThe Bureau Works Blackbird integration enables automated localization workflows through no-code tools and platform interoperability.
→ Continue Reading
