AI Localization Think Tank Looking Forward to 2026 | Part 1

As we enter 2026, the AI Localization Think Tank went from reflection to anticipation. The year ahead promises to solidify changes already in motion and bring new ones. But what changes can those be?

To explore what lies ahead, we asked each member three questions:

• What major change do you expect to see in the field in 2026, and why?

• What emerging opportunity or risk should the industry prepare for in 2026?

• Which innovation, behavior shift, or market force do you believe will define 2026?

The volume and depth of perspectives once again exceeded a single publication. We’re therefore sharing the insights in two parts.

This is the first part. The second part will be published on January 22.

Aaron Bhugobaun

Technical Production Operations Manager, CDSA Studio Chair, and Creative Technologist

What major change do you expect to see in the field in 2026, and why?

Agentic AI with stronger contextual understanding. I expect to see more of this in 2026.

What emerging opportunity or risk should the industry prepare for in 2026?

Agents are great because they take away a lot of “drudge” work, but ultimately their capabilities will increase over time, as we’ve seen with LLMs. How companies implement these capabilities will ultimately decide how humans are impacted.

Which innovation, behavior shift, or market force do you believe will define 2026?

I’m watching the US market very closely, specifically how companies introduce more AI into traditional human managed workflows and how that affects the US workforce. (as there employment law is more relaxed EU/UK) As mentioned.

Andrea Ballista

Co-founder, CEO

What major change do you expect to see in the field in 2026, and why?

An increase of various AI tools and AI solution to deal with multiple aspect of the language market: to be able to use at best all the tools to solve specific use cases need to be solved by next-gen of orchestration and agentic platforms driven by clear requirements and by the possibility to use multiple techs to fine-tune and update the most efficient solution for each specific use case.

What emerging opportunity or risk should the industry prepare for in 2026?

The speed of the tech improvements is definitely introducing the need to quickly and substantially transform current processes, insourcing and managing new competences. Just a small set of medium/large companies will be able to cope with the challenge and there will be a clear split between who is able to manage and integrate new tech and workflows following the brands and publishers needs and who does not have resource/time to adapt to new scenarios.

Which innovation, behavior shift, or market force do you believe will define 2026?

More “voice first application”, more “agentic solution” supported by synthetic voice quality measurement capabilities, more blending between LTP and LSI, ability to solve “use cases” and the skills to continuously update them.

Balazs Kis

Co-founder, Chief Evangelist

In 2026, I see an opportunity and a threat for the language industry: Language professionals will emerge as members of a larger, more complex profession than just translation and localization.

The divide between humanities and STEM will get narrower. This will make it easier for language professionals to assume their power over large language models — a power only they possess.

The threat I see is the looming AI bubble that will burst sometime n 2026. Although it isn’t entirely certain, many analysts expect it to happen and produce a fallout similar to the crisis of 2008.

If — or when — the AI bubble bursts, a lot of capital disappears from the market. I don’t think LLMs or AI services will disappear but providers will put a market price on most services. As a result, many smaller users might be priced out of the AI game. This will accelerate the process of “sobering up”, which means that large portions of the general public become aware of the limitations of current AI systems.

We know that there are limits to what and how large language models can learn; how far their representation of meaning and the world can go; and we know they are a dead end to the road to artificial general intelligence (AGI). This will become known to the general public, and hopefully AI literacy will make it to public education.

Limited as they are, LLMs will remain very powerful automation tools in language work. But they are not the technology that bring about “singularity”. That will be something different, and it will definitely not happen in 2026. 

Belén Agulló García 

Executive Consultant of Innovation

When we reflect about our industry and think about the future, we tend to forget that we do not create anything, we transform what others create. And most of us offer services, not products. It doesn’t matter if you are an LSP, LSB, or a freelance translator: services is what we offer, just to different stakeholders. And interestingly enough, some of the language technology providers out there are also offering the service on top of the product, which kind of tells you the level of automation that can be achieve at the moment, and the real needs when it comes to localizing content.

Nonetheless, when we think about our future or we’re trying to predict what will happen to us, we should not gaze at our own navel, as we won’t find the answers there. We should instead look outside of our bubble and analyze the behavior of our clients and stakeholders, and that of the clients of our clients, aka, the end users. Reports like this one by StartUs Insights shed some light into what consumers are expecting from brands. Keywords such as personalization, transparency, trust, local relevance, cultural nuance, data privacy, sustainability or emotional connection point to the differentiators that will make brands stand out in a very crowded and competitive market. It seems like language, content, and emotional connection are important for consumers, who will in the end decide the fate of all the brands out there with their wallets.

With this in mind, the major change that I expect this year is that we will all become more user-centric and less technology- and process-centric. While workflows and technology are essential to what we do, they are not strategic at a business level. To me, the biggest opportunity for our industry is to become experts in our local users and leverage those insights to then inform the operationalization part of what we do. To do that, it is key to leverage LLMs and different levels of automations, including agentic workflows, paired with business intelligence teams that can turn massive amounts of data into meaningful insights that will positively impact business outcomes. And then, of course, we need to be able to measure those outcomes that are relevant for the business.

Personally, I’m curious about one specific behavior shift that I see on social media, which is the rise of auto-translation features. We see it on LinkedIn (with one button you can translate a post into your the language that you have set as primary in your profile), but also Instagram, YouTube, and TikTok. Even if you don’t localize your content, now Meta and Google do it for you (without your consent), and it’s up to the user to turn it on or off. And it’s not just text, but also voice now. So how will this impact how users perceive brands? And how does it feel to lose control over what’s localized and how? Does it really matter to anyone? To be seen… 

Coral Diez Carbajo

MT and AI Strategist

What major change do you expect to see in the field in 2026, and why?

The biggest event on the horizon for 2026 is a hard “face-the-facts” moment about the real costs of technology, with AI front and center. In the past years, companies have prioritized innovation and development over the cost of their per consumption-based solutions. However, organizations are reaching a point where the costs for AI services can no longer be sustained without showing clear, tangible results. The central issue is that the first wave of generative AI (often launched too early) failed to deliver meaningful productivity. When vendors responded to the low quality of results from this first wave of GenAI by pushing agentic AI, this only intensified the problem: agentic AI’s systems run up costs by dramatically increasing the iterations and thus the expense per business function. Consequently, in 2026, many companies will begin scaling back their AI agents because they simply cannot justify the cost per consumption and the economic benefits to management. 

What emerging opportunity or risk should the industry prepare for in 2026?

There will be a growing divide between "AI in platform" and "AI as platform." The immediate opportunity for mid-tier and large enterprises is the passive adoption of AI offered by the "AI in platform" model. Major platform vendors, such as Salesforce, Adobe, Canva, and NetSuite, are increasingly baking AI directly into the core processes of their established systems. For organizations relying on these solutions, this approach suggests they "don't need to do much" and can simply wait for new AI functionalities to become available, positioning them as a consumer of AI services. However, this reliance on AI being embedded within core platforms presents a significant risk: it locks organizations back into the high-cost, consumption-based models pushed by vendors. The real opportunity for strong market differentiation lies in the "AI as platform" approach. This requires developing a highly architected platform capable of using multiple AI models and implementing orchestration to apply the most efficient model at specific, costly points within known business workflows. While this requires more effort and knowledge (and is currently preferred by larger organizations) the industry needs to be cautious that the simplicity of consuming AI within existing platforms does not lead to cost complacency while competitors pursue true efficiency gains through custom-built "AI as platform" architectures. 

Which innovation, behavior shift, or market force do you believe will define 2026?

The landscape of technology in 2026 will be defined by two profound internal shifts: the preservation of institutional knowledge and the restructuring of data. Internally, the most critical innovation is that organizational memory becomes durable and computationally preserved. This is a fundamental change from the past, where knowledge was fragmented and faded when humans left, to a system where agents learn continuously from both staff and other agents, effectively making institutional memory an asset independent of staff turnover. Complementing this, knowledge stops being isolated files and starts to become structured meaning. Through a transformation process utilizing knowledge graphs and embeddings, disconnected data is normalized into structured representations that create a unified semantic fabric. This shift allows for the automatic processing of data and the creation of agentic workflows, enabling crucial cross-language, cross-format, and cross-departmental retrieval of meaning.

Helena Moniz

President of the European Association for Machine Translation

We have not answered core questions as what makes us humans? How can we evolve as social beings and technology creators if we do not yet understand the creator and the creators impact in the near future. Are we willing to delegate our creators’ status to new creators in every aspect of our lives? At what cost? 2026 may be a step more to better understand the social and cultural beings we are and move to cultural awareness and fluid communication.

More languages means more cultures and we need to represent their richnesses, their voices, their multifaceted and multimodal ways of experiencing the world. We are very focused still in a single modality in translation, we need to expand to multicultural and multimodal scenarios. Every language on its own, existing per se, without always being compared to the mainstream ones.

2026 will bring new ways and more real pipelines on optimizing AI pipelines, making them simpler, but yet more comprehensive on integrating all in the broad spectrum of who we are and how we communicate. More domains, more languages, more tasks which were somehow forbidden places can be unveil, but still we need a lot of effort to do so, better metrics, better data and better ecological settings to understand the impacts of what we create.

A trend I would very much like to be explored is how we make AI Responsible and socially aware of our humanity. Guardrails, Red teaming, social awareness, group dynamics, agentic behaviors which can assist us not harm us, etc.. We are already noticing the surge of AI agents and we are accomodating to that reality. Understanding that we can not be automatic pilots in our own realms is fundamental. How to implement this is still a very unexplored terrain. A true parodox of AI, since we are using AI without fully understanding bias amplification, social implications and future impact in teh next generations.

A toast to training new generations and to the excitement of being questioned by students who research new ways to use AI in academia, industry and society. A toast to being challenged and having to discover together the answers. A toast to a better understanding that we humans are fragile beings with amazing capabilities! What makes us fragile is also what makes us amazing beings! A toast to creating ecosystems, such as this Think Tank, of different lens to learn from each other and create new ideas, products and new ways of research with societal impacts. A toast to a 2026 with a lot of challenges and opportunities to better understand who we are as humans! A toast to washing our eyes in every look and trying to understand the why and who!

Gabriel Karandysovsky 

Researcher, Content Creator and Consultant 

I’m personally very curious and invested in seeing what comes after the L-word. Depending on who you ask, our classic industry jargon is running out of runway, and I love seeing new approaches being tested, like “global experiences” (plural very much intended here). Other difficult-to-stick acronyms have been floated in 2025, but the trouble with acronyms is that you always need to explain them — and by the time you do, the conversation’s already moved on. Also… what’s wrong with “translation,” anyway? That’s what my mom remembers. Yours too, probably.

Ultimately, it’s less about the label we slap on the work than the why behind it. I’m watching for companies integrating translation/localization with disciplines such as UX, user research, product, and content creation. We’re inching toward multilingual content created from scratch — not necessarily groundbreaking in concept, but newly viable thanks to AI — a system that will live in parallel to the cascading, sequential models we rely on today. And the questions that will define that shift are simple but oh-so essential: How do we craft meaning? Why do we create the content we do? Who needs it and, more importantly, why?

Johan Botha

African Language Solutions expert

What major change do you expect to see in the field in 2026, and why?

In 2026 I expect the industry to shift from broad AI adoption to disciplined AI management. We spent the last few years rushing to plug tools into workflows. This year the focus will move to control, accuracy, trust, and perhaps even accountability (there goes my stupid optimism)

What emerging opportunity or risk should the industry prepare for in 2026?

The biggest opportunity lies in custom language models built for specific domains and communities. These will outperform generic models and unlock better quality and more inclusive language coverage. The risk sits on the other side of the same coin. Without proper oversight, the gap between well-supported and last- minute languages will widen.

Which innovation, behavior shift, or market force do you believe will define 2026?

Smaller language models and real-time multimodal (agentic) AI will change how content is created and delivered.