The State of

Localization
A CENTIFIC REPORT

2023 EDITION
 THE STATE OF LOCALIZATION REPORT

CONTENTS
3 Introduction
by Jonas Ryberg

4 Localization in an Economic Downturn

by Wei Zhang

6 Quality and the Customer Experience

by Giovanna Conte

9 Localizing AI
by Vincent Swan

13 How AI Localization Can Help Global Brands Scale

by Sergio Bruccoleri

15 The Human Factor

by Alba Guix

18 About the Authors

19 About Centific
Introduction THE STATE OF LOCALIZATION REPORT

by Jonas Ryberg
Chief Globalization Officer

The field of localization has evolved rapidly. But overall state of localization is making tremendous
has it “arrived”? strides and becoming impossible to ignore in the
news, whether it's Google’s cloud-based translation
On the one hand, localization is integral to most hub or Meta’s new AI tool for translating primarily
business’s global growth aspirations. If you’re a oral languages.
business that aspires to expand its operations
across multiple countries, you can’t afford to Where Localization is Headed
ignore localization. Our new report takes a close look at the state
of the industry, ranging from the impact of the
Language services was estimated to grow to COVID-19 pandemic to the challenges that
more than $57 billion by the end of 2022. This businesses face when localizing products such
growth makes sense: globalization is as strong as voice assistants that use artificial intelligence.
as ever, and as businesses increase their presence We’ve tapped into the thinking of some of our
across hundreds of countries, being able to best subject matter experts to provide thoughtful
communicate with customers who do not speak insights into where localization is headed in 2023.
the same language is a fundamental need. But This report is written for professionals who
localization is more than language services, and believe in
localization done poorly can undercut a brand’s
global growth by alienating potential customers. the power of localization to solve
business problems – or who are
curious about its positive impact.
Localization encompasses many moving We’ve based our thinking on our
parts, including the user experience – the
basic user interface, choice of sound and experiences helping businesses build global brands
color palettes for a site, and yes, language. across hundreds of markets in some cases. The
In short, localization is about creating practitioners who’ve contributed essays have
lovable experiences everywhere. strong first-hand experience collaborating with
clients, localization teams, and technologies such
as our OneForma platform to scale localization
Localization still has a long way to go. According efforts.
to a recently published survey of European
consumers, 57 percent had experienced cultural Thank you for your interest in our report. I hope
problems while using a UK brand’s local website; you find it useful and enlightening. As always,
24 percent of this group said these issues contact us at Centific with your own thoughts
reduced purchase likelihood. Fortunately, the and questions.

3
Localization in an Economic Downturn THE STATE OF LOCALIZATION REPORT

by Wei Zhang
AVP of Localization Technology

Businesses around the world are managing through an economic downturn that just gets worse by the day.
In the fourth quarter of 2022, the International Monetary Fund (IMF) predicted that global growth would
slow to 2.7 percent in 2023. The IMF also said the world economy was “historically fragile,” and the financial
markets were “showing signs of stress.”

Businesses, seeing the bumpy road ahead, are under more pressure than ever to cut costs and operate
more efficiently, while maintaining quality and protecting time to market. What does that mean to their
localization efforts? And how can localization help a business not only survive a recession but also
manage to improve their globalization efforts?

Here are two trends:

Localization is effective at helping Outsourcing large-scale localization

companies expand into new and projects will become normal for
bigger markets. brands.
Many companies are targeting overseas markets With workforce reductions happening (and more
for growth. That’s because in a downturn, it is likely coming), we’ll see an increase in the
common for global businesses to try to press outsourcing of jobs. At present, many companies
their advantage over competitors by achieving still retain large localization teams in house. Many of
economies of scale. But doing so is enormously those positions and the work might be outsourced
challenging amid supply chain disruptions and to save money.
geo-political instability that threatens the free
flow of goods and services. We already saw a surge in people entering the
localization market through self-employment
Global companies that already serve multiple during the pandemic and we expect the surge
local markets have an advantage because it is to increase. As a result, businesses will tap
less expensive to grow an existing infrastructure into a broader book of diverse talent globally,
than to build new markets from scratch. That’s including people who possess more breadth
why now is the time to lean into localization – and and depth of language expertise from different
here’s why: Localization makes a company more countries, domain expertise, and knowledge of
relevant in the markets it serves and is necessary local cultures. This will make globalization more
to expanding your existing customer base in effective in the long run. For more insight into this,
diverse markets. After all, approaches that work read Alba Guix’s essay in this report, “The Human
in the United States won’t succeed as well in Factor.”
Spain or Japan and vice versa. For more on this,
read Jonas Ryberg’s post, "What Is the Future of
Globalization" 4
 THE STATE OF LOCALIZATION REPORT

How Businesses Can Operate More Efficiently

Operating in a downturn means applying localization to be more efficient and cost effective. Here are
some ideas for businesses to consider:

Diversify the talent pool. Becoming Adopting automation for file

more relevant to local markets management, translation validation,
requires businesses to rely on a project management, and cost

more diverse team of localization experts monitoring will significantly reduce the cost of
who possess knowledge of cultures and localization production.
communication styles in countries where a
Machine translation will play a
business operates. This approach can be done
bigger role. During recent years,
most efficiently by complementing human
machine translation technologies
expertise with a technology platform used to
manage projects and assets. For example, have dramatically improved, becoming a great
freelancers can easily onboard directly and find solution to help reduce localization cost, while
tasks in their related areas. An outside partner providing fast translation coverage. Adopting
with a ready-made team of localization specialists machine translation technologies requires
and technology platform can help. engineering setup, post-editing, linguistic
evaluation, engine training, testing, and code
Enhance business-intelligent
integration. But once these elements are in place,
technologies for localization
their value can be extracted immediately and in
decisions. Which features, pages,
some cases, indefinitely.
videos, or documents should be localized into
which languages? Businesses can and should Conclusion
use business intelligence and user research to The 2023 economic recession will have a big
localize the high-traffic, top-priority content and impact on the localization market. But it will also
locally relevant content. bring more opportunities for businesses. By
diversifying the organizational talent pool and
Localization should start with the adopting technology wisely, businesses will best
content creation stage. When creating position themselves to succeed globally.
new documents or software resource

files, businesses should try to re-use existing

content/strings, reduce the number of documents,
identify and remove the “local market relevant only”
content from the localization list, use localizable
file formats, and create comprehensive style guides
and a terminology list to not only increase the
linguistic quality, but also reduce costs and
turnaround time.

5
Quality and the Customer Experience THE STATE OF LOCALIZATION REPORT

by Giovanna Conte
Head of Global Quality Management

Providing a more personalized, high-quality Personalization is essential to a better CX/UX.

customer experience (CX) is the new mandate According to Nimdzi Insights, 60 percent of
for localization and for global businesses. users are likely to buy a personalized product
recommendation they see on an online store. All
During recessionary times especially, satisfying of this means that the localization professional
existing customers with a better experience is needs to support global businesses by delivering
more cost-effective than acquiring new ones, a more personal experience that results in higher
which is urging businesses to put a higher priority levels of satisfaction from one country to the next.
on delivering better CX. As a result, businesses are
connecting localization to broader CX requirements, This shift has a profound impact on how businesses
including the online user experience (UX). approach localization of content and experiences.
For example, it used to be that businesses relied
on a single, language-specific style guide as the
standard for localization. Businesses used this
style guide for localization instructions, and any
large localization services provider would need to
know that style guide by heart.

This top-down style guide ensured localization

was standardized and the basics of communication,
such as grammar and syntax, were consistent;
but, businesses often created content that was
awkward and disconnected from how people
really communicate. As a result, localization
efforts might have been technically accurate, but
they did not contribute to a better experience.

As times changed, brands have taken a bigger

look at the totality of the UX, from language to
experience, only to realize the need to provide
higher-quality content that responds to what
people want based on their experience. All this
has brought us to a new era – one that is laser-
focused on delivering human-centric, intelligent
experiences.

6
 THE STATE OF LOCALIZATION REPORT

Language effectiveness (e.g., translation quality) across markets is as important as ever. But quality
means satisfying users, and this goes beyond language. For example, localization professionals need to
understand language in context of the entire UX. A description of a table accompanied by an image of
someone putting their feet on a table might be factually accurate and complement the text well, but the
photo might be offensive to cultures such as Japan, where this is considered disrespectful. In contrast,
the same image might be perceived as expressing relaxation in France.

This means that localization teams need to operate in a more cohesive, integrated fashion, instead of
working in silos. Where quality was once a key step in the production cycle, it’s now become a mindset
and design principle shared across teams, where each team is deeply involved in the quality of the
localized end product experience.

How this looks in practice:

Content creation teams liaise with localization teams by preparing content that can be easily
adapted and customized for each market to avoid errors and rework down the line. The central
team needs to have oversight into either hiring or outsourcing local talent. But from there, the
central team needs to be willing to defer more to the local teams on matters of style, content
creation, and overall experience rather than mandate a top-down approach.

The central team still needs to manage the overall quality of the experience. For instance, a
central team needs to oversee formal ways of measuring customer satisfaction, such as surveys
and focus groups, so that there is an agreed–upon way to collect feedback across multiple markets.
The central team needs to work with local teams to manage the actual process. At Centific, we
have found that user surveys and focus groups are the most useful methods of improving what is
important. We work with our UX team to include the multilingual customer experience in research
and share this information with our translation team, thus creating circular communication.)

7
 THE STATE OF LOCALIZATION REPORT

Briefings have become key between local customers and localization teams to ensure that
experts on the ground remain in sync with customers’ needs. Training and certifications are
also crucial to training linguists and getting translations right the first time.

Automated checks and customized solutions are adopted to avoid recurring errors or
unacceptable behaviors. This is where artificial intelligence (AI) can play an essential role,
so long as a diverse team of humans are in the loop to guard against bias creeping into AI.

Localization experts need constant upskilling because language and customs are dynamic.
For instance, emoji and stickers are constantly falling in and out of favor from one country to
the next. It’s essential that localization experts understand these nuances to deliver a great
experience.

Localization needs to be measured against quality more than accuracy of content. Quality
scores help the organization get visibility into the quality of translated content with no need to
understand the target language.

To ensure that work is performed with consistently high standards of quality, integrated
teams need a central technology platform to manage everything from workflow to defining,
reporting, and tracking the success of global experiences. A single platform ideally should
collect and analyze end-user feedback, such as customer ratings and reviews (which Alba Guix
discusses in this report), in order to give the localization teams a constant feedback loop.
Teams also need to be supported by proper tools, such as social listening, in order to monitor
customer feedback in real time.

Bottom line: Localization is no longer just translating software but providing a great experience that is
personalized to the needs of customers in multiple markets. As a result, localization means teams need
to act in a more coordinated fashion, applying a broader complement of skills and technology.

8
Localizing AI THE STATE OF LOCALIZATION REPORT

by Vincent Swan
VP of Technical Solutions

Products fueled by artificial intelligence, such as voice assistants, are becoming smarter and more
pervasive around the world. At the same time, their proliferation raises concerns about bias. AI localization
can play a crucial role in making AI-based products more inclusive and relevant to the needs of multiple
markets.

AI Localization Defined
AI localization is about localizing AI products, tools, and services to work in their respective markets.
Localizing AI means training AI with localized data – in other words, collecting and curating data
sets that respond to cultures in different markets. In addition, localization experts increasingly use AI
and machine learning engines as essential tools for enhancing end customer experience and driving
customer satisfaction and loyalty.

In this context, localization is no longer about just translating files, text, and software for human
consumption. Language translation is part of AI localization, but AI localization also includes localized
experiences that people love no matter what country they live in – experiences that resonate based on
their own cultures.

Ultimately, AI localization helps AI deliver memorable experiences.

As localization specialists make voice assistants, chatbots, computer vision applications, and other
AI-based products relevant to different markets, there are some important points to consider:

Pervasive AI Products Create New Challenges for AI Localization

The global market for AI-fueled products is exploding. Chatbot-related products are expected to grow at
a compound annual growth rate (CAGR) of 22.8 percent between 2022 and 2028 according to Stratview
Research. AI in computer vision will achieve a 46.9 percent CAGR from 2021-2027 according to Insight
Partners. And these are just two examples.

As AI products become more widespread, localization professionals need to train them to be more
responsive to global audiences. For example, chatbots need to handle customer interactions in multiple
countries and cities, which requires localization professionals to train chatbots on how to read customer
interactions and deduce whether the bot can help them or if they need a human being.

Part of the challenge is training the chatbot to understand when a question is too complicated to address,
which isn’t an easy task when you’re working in multiple languages. For instance, the chatbot needs to

9
 THE STATE OF LOCALIZATION REPORT

discern local slang and accents to understand what someone is asking.

But localization professionals also need to train the chatbot to understand important issues of context,
tone, and intent:

• Context: A chatbot in Florida shouldn’t recommend snow chains for tires, but a chatbot in Quebec
should. A retail chatbot needs to know the difference between, “I need a self-help book” and “I need
help.”

• Tone and intent: Understanding tone goes hand in hand with intent. Different cultures express
themselves differently. A chatbot needs to discern when a customer is truly becoming upset versus
being direct, or whether a customer is calling because they need urgent help from customer service.

Knowing how to translate content for multiple

languages is a necessary skill, but it isn’t
necessary for a localization expert to be effective
as AI-based products evolve in local markets.

The same holds true for any AI-based product

such as synthetic voice, smart speakers, and
voice search. The hit movie Us, directed by
Jordan Peele, illustrates the challenge.

A victim of a crime cries out to her voice assistant for the police. In a moment of satiric commentary
about our dependence on technology, the voice assistant begins playing a famous rap song about
the police. The assistant responded based on the data it had been given. But it obviously failed to
understand context and tone.

Similarly, computer vision applications need to understand context. Computer vision is making great
strides in helping stores safeguard themselves against theft. It’s also helping large, public spaces such
as airports manage crowd control and security. But computer vision requires effective data annotation
to understand how to evaluate a scenario.

For instance, let’s say an airport wants to train a computer vision model to spot a knife hidden in carry-on
baggage. The model needs to know what a knife looks like. And this means all the knives in the world,
ranging from a carpenter’s knife to a balisong. It also needs to know how to distinguish between a knife
and other objects – like the difference between a carabiner and fold-out knife for example. It’s easy to
see the challenge for localization here: training the application to discern context in airports ranging
from Heathrow in the UK to France’s Charles de Gaulle Airport.

10
 THE STATE OF LOCALIZATION REPORT

Bias
AI localization is at the center of an ongoing
conversation about the bias problem with AI.
Businesses continue to struggle with AI-based
products that are fraught with bias, as these
examples illustrate. Amazon even recently
abandoned an AI-based recruitment software
because it was biased against women. As it turns
out, the data on which the AI algorithm was based
favored male candidates. In other news, a Google
Photos labeling feature was found to be biased
against darker skin tones.

AI bias happens because people bring their

inherent biases into training AI to do its job. For
example, Amazon’s recruitment software was
designed to rank job seekers on a scale of one
to five stars. It was developed with the input of
10 years’ worth of resumes, most of which came
from men. As Reuters reported, “In effect, Amazon’s
system taught itself that male candidates were
preferable. It penalized resumes that included
the word ‘women’s,’ as in ‘women’s chess club
captain.’ And it downgraded graduates of two
all-women’s colleges, according to people familiar
with the matter. They did not specify the names
of the schools.”

These issues can creep into AI localization, too.

For instance, machine translation technology can But how about “I am your new physician”?
produce results that reveal gender bias. Gender A machine will pick Arzt or Ärztin based on
bias can happen when you are translating an whichever translation is statistically more likely
expression that is gender-neutral in the source and on what they have seen more often in their
language but needs to be gender-specific in the training data. As a result, a physician is more
target language, such as English doctor to German likely to be identified as male, reflecting bias.
Arzt (male doctor) or Ärztin (female doctor).
Now imagine that simple translation bias being
If the machine is translating a sentence containing repeated many times in chatbots, smart speakers,
such a word from English into German, how and so on. The magnitude of the problem and its
does a machine know which gender to use? potential impact on business is significant.
Sometimes the answer is easy. The sentence,
“She is your new physician” will be translated
using Ärztin.
11
 THE STATE OF LOCALIZATION REPORT

We believe that a Mindful AI approach is pronunciations of certain words in a target

key to fighting bias of all kinds in AI, and market.
this includes AI localization.
Doing so helps our clients make AI more inclusive.
We define Mindful AI as follows: developing Mindful AI also needs a process and tools for
lovable AI-based products that put the needs designing a solution with people at the center.
of people first. This is done by being inclusive, You can read more about that here.
as free of bias as possible, and building trust.
Mindful AI won't happen without people --
Mindful AI and Localization in Action
especially a diverse team of people -- being
involved in the creation of AI-based solutions. Our expertise with AI localization enables our clients
This approach is known as keeping humans in to make their AI applications more inclusive and
the loop. personalized, respecting critical nuances in local
language and user experiences that can make or
Mindful AI has many moving parts, but one break the credibility of an AI solution from one
of its most important components as far as country to the next.
localization is concerned is keeping humans
in the loop. Human-in-the-loop AI combines We design applications for personalized and
the strengths of people (e.g., creativity, insights localized contexts, including languages, dialects,
from ambient information, historical and cultural and accents in voice-based applications. That way,
context) with the strengths of machines (e.g., an app brings the same level of voice experience
accuracy, speed, and the ability to manage sophistication to every language, from English
repetitive tasks that people don’t want to to under-represented languages. The last mile in
manage). making AI pervasive and trustworthy is localizing
and personalizing content and experiences.
Ultimately, human-in-the-loop AI means relying OneForma enables that for our clients, thereby
on a globally diverse set of people whose mere transforming the industry while helping AI
involvement in training AI applications instills applications become more inclusive and equitable.
a more inclusive approach. At Centific, we rely
on globally crowdsourced resources with in-
Conclusion
market subject matter expertise, proficiency in
While more pervasive AI-based products will
200+ languages, and insight into local forms of
improve how people live and work around the
expressions such as emoji on different social
world, they also are fraught with challenges. AI
apps.
localization both compounds those challenges
and offers a solution, depending on how well a
Humans also need the help of technology to scale
business sources and trains a diverse localization
their work. For instance, our crowdsourced team
team to expand its skill set beyond language
uses our OneForma platform to teach AI models
translation.
how to make accurate decisions. LoopTalk, our
voice AI data generation capability, enables our
team to train voice recognition models to better
understand regional accents and non-typical

12
How AI Localization Can Help THE STATE OF LOCALIZATION REPORT

Global Brands Scale

by Sergio Bruccoleri
Director of Platform and Innovation

Many businesses think of localization as a tool that helps them reach new markets by making content
more relevant to their customers’ cultures and ways of communicating. And this is certainly true. What’s
intriguing, though, is how localization, with the help of artificial intelligence, achieves this outcome.

Consider the enormous challenge of being relevant to cultures at a time when there is an explosion of
social media content from one country to the next. The persistent and rapid creation of user-generated
content requires businesses to be equally nimble, if not more so, with their own content – meaning they
need to practice “instant localization.” Turnaround times must be quicker as content is originally created
in the source language. Publication of content in multiple markets and languages must be practically
instant.

Infusing AI as part of the localization process can accelerate publication times globally, while optimizing
operations such as task allocation, quality management, and even publication itself. Let’s take a closer look.

Where AI Fits Best in the Localization Workflow

In the era of instant localization, AI models have a multitude of uses depending on which step
the localization process they are leveraged in.

Neural Machine Translation

Neural machine translation (NMT) can be a strong ally when a brand needs to publish content in multiple
languages fast. Using publicly available models provided by Google, Amazon Web Services, or Microsoft
can be tempting, but these often require a bit of post-editing before final publication. Infusing NMT into
the final process helps accelerate publication times. But there are a few points to consider:

• Translator-in-the-loop: A translator must always review the content before publication to ensure it
respects the style and tone of the content and brand it is representing.

• Model training: The most common NMT models allow for fine tuning – which means customers
can upload their own translation corpuses to ensure the AI model learns the style and terminology
of the content and brand directly from the translators. This leads to shorter post-editing times as
more content is translated.

• For content, not for AI training: Machine translation (MT) is an ally for content publication, but
it’s not a useful tool to enhance or further train your own NMT model. Why? AI isn’t perfect, so
anything provided by an AI model is just a “guess” based on a backend algorithm fueled by another
13
 THE STATE OF LOCALIZATION REPORT

AI's training data. This means that using an NMT model to train another NMT model will lead to
potential bias and source model issues that can transfer to the target model, creating a model that
underperforms.
• Great tool against bias: Where MT cannot be used to train other MT models, commercial neural
machine translation engines can be embedded for AI training, but not as helpers for the translation
of content. Rather, they are more applicable as tools to verify that the translator is not using MT to
translate the content themselves. Using MT engines in conjunction with algorithms, such as the
Levenshtein Distance, can help estimate which translators might be using Commercial MTs. This
ensures translators are not being biased by using NMT to complete a translation task, which helps
ensure consistency and bias-free training data for NMT engines.

Task Allocation
When it comes to selecting the right translator for
the job, ML can use learnings from past performance
data to determine who is the best translator or has
the best allocation or response time based on previous
translation handoffs in the same day of the week and
time of day. This can lead to improved quality and
turnaround times for translation. But that data must
be part of the localization management process to
start with, which means data centralization from a
translation perspective, as well as from a handoff
and task management perspective.

Quality Assurance by Exception

Once a translation task is performed, machine learning can be used to study the performance of
your pool of translators in real time. Applying ML can help determine which translators might be
performing “differently” compared to the rest of the pool of translators. For example, machine learning
can help identify which translators are translating certain sources too quickly compared to the rest
of the translation pool. It can also help you spot reviewers that tend to use certain error categories or
severities in the review process more than others.

There are several tools on the market that can help create large indexes of your translation data,
including ElasticSearch, which provides some basic machine learning capabilities out of the box, but
also requires a consistent, centralized data source across the entire localization management workflow.

AI as the Ally
AI can be a strong ally in the localization workflow, from translation and review to final publication.
But keep this in mind:

• A translator-in-the-loop becomes essential to ensuring content fits the tone and style of the brand.
• A translator-in-the-loop approach helps ensure the content feels "local”.

Data centralization and the “platformization” of the localization management process are critical to
success and the best way to leverage machine learning techniques end to end. 14
The Human Factor THE STATE OF LOCALIZATION REPORT

by Alba Guix
VP and Head of Europe

The localization industry centers around people. People looking for localized experiences and people who
do the localization. What trends are affecting the relationship between the two?

New Ways of Working

Before the pandemic, remote work was limited to a few professions. The world was not set up for
online work. Most companies did not promote remote work as an option; many discouraged the model.
The pandemic changed everything. It enabled a new work style that let people work from anywhere.
The rise of remote working has created a global workforce that enjoys that enjoyed a more flexible,
balanced lifestyle suited to their personal needs. Along with it, the proliferation of remote collaboration
tools such as Mural, Teams, and Zoom has made it even easier for people to work and collaborate
remotely.

As a result, more people have sought the benefits of remote work, and this includes localization, which
relies heavily on work outsourced to independently contracted professionals.

Thus, localization has become a more attractive profession to more people, receiving an infusion of
new talent that come from different backgrounds and diverse talents beyond language or translation
skills, including unique industry knowledge and domain expertise.

This wider pool of localizers also brings their own knowledge on local languages, cultures, and products
that is relevant to localization. In a world where
user experience is key to determining the success
or failure of a product, bringing that in-market
product experience leads to superior localization.

Take this example of someone using PayPal in

Spain for online purchases. As a native Spanish
speaker that knows the local market well, they will
be very familiar with the content being translated
for this market, making them better equipped for
localizing similar payment products than other
linguists without that knowledge.

Bottom line: talent pools have become more

global and diverse.
15
 THE STATE OF LOCALIZATION REPORT

The Rise of the Feedback Economy to each person in each country. As many brands
We’re all living in a feedback economy in which have discovered, people will ignore – or even
customers’ ratings and reviews can have a worse, criticize on social media – products that
profound impact on a company’s reputation and fail to measure up. In this context, localization
its well-being. This is encouraging businesses supports a new sphere of business activity:
to figure out better ways to monitor, collect, and helping companies achieve stronger levels
learn from that feedback, which includes using AI of customer satisfaction, which manifests in
tools such as natural language processing. stronger ratings/reviews.

The feedback economy is a global phenomenon Creating content that is factually correct and
affecting global businesses. Its rise dovetails with grammatically sound is essential, but not
an increased focus on delivering more personal, sufficient. As Giovanna discusses in this report,
higher-quality experiences in local markets, which localization experts also need to create content
Giovanna Conte discusses further in this report. that is culturally relevant and more personal to
each audience. This requires subject matter
In the digital era, where access to product reviews expertise and knowledge beyond linguistic skills.
from other users and personalized campaigns is
the norm, users have become more demanding In addition, localization experts themselves
with their experience of those products and need to be evaluated against a broader set
services. What’s more, users are becoming more of metrics beyond the technical mechanics
vocal about sharing their positive and negative of creating accurate content. Which begs the
experiences with others. following questions: How relevant is their work?
And how well does this work support customer
For products to create a genuinely lovable satisfaction?
experience, those experiences need to be relevant
16
 THE STATE OF LOCALIZATION REPORT

Emerging Conversational Formats

The digital era continues to usher in different
forms of expression and customer interaction
between the brand and consumer. Other than
those mentioned above, brands and consumers
are also interacting everyday behind the firewall,
an example being the rise of chatbots.

With chatbots, the relationship between the brand

and consumer is immediate. It’s decoupled from
space and time if a brand uses a chatbot powered
by AI that can answer questions 24/7.

Chatbots are becoming more sophisticated

and human. For example, startup Tymely is
developing a chatbot that enables brands to
provide email and chat support services in a
more human, empathetic, and precise way. And
Soul Machines has developed a 3D digital person
currently being used on sites such as Calocurb Conclusion
that responds to your talk or type preferences, The pandemic and new economy models have
and its digital person even smiles back at you if increased and diversified the community of
you decide to use your camera during a session. translators and localizers worldwide, enabling
And, DeepMind has developed Sparrow, which businesses like Centific to benefit from a larger,
answers more complex questions using AI. more global workforce with other talents and
skills that make them more qualified to new
As chatbots become smarter and more global, localization demands.
localization experts will play an increasingly
important role in making them more relevant
to each market they serve. But the localization
expert here needs to expand their responsibility
beyond language translation. With chatbots, the
business needs to read customer intent.

When a business localizes chatbots, it’s not

just translating English that is important. The
localization professional must understand
different ways of expression. They, too, must
discern intent in order to train the AI model to
respond in a suitable way, whether the chatbot
is communicating with a customer in Mumbai
or New York.

17
About the THE STATE OF LOCALIZATION REPORT

Authors

Sergio Bruccoleri
Sergio Bruccoleri is the Director of Platform and Innovation for the Globalization
Services team at Centific, where he leads the R&D, Engineering, and Platform teams
behind Centific AI Data & Language Services offering.

Giovanna Conte
Giovanna Conte is the Head of Global Quality Management at Centific, where she
leads Quality Managers and Language Specialists teams across different geographies.

Alba Guix
Alba Guix is VP and Head of Europe at Centific, where she also leads the Global Shared
Services organization managing Resources, Quality, and Learning.

Jonas Ryberg
Jonas Ryberg is the Chief Globalization Officer at Centific, where he manages
Language and AI Data services.

Vincent Swan
Vincent Swan is VP of Technical Solutions at Centific, where he manages the teams
delivering innovative technical solutions & support ecosystems for AI & Globalization
Services.

Wei Zhang
Wei Zhang is an AVP of Localization Technology at Centific, where he focuses on
adopting new localization technologies for global customers.

18
About Centific
Centific is a global digital and technology services company. We design, build
and optimize human-centric intelligent digital platforms. Our core capabilities
are in data, intelligence, experience and globalization. To learn more about us
visit us at centific.com