Innovation and Change Management in Localization: Lessons from AI Deployment
In our March edition of Elevate Innovate, we explored how innovation helps companies advance through various levels of their evolution. This session was hosted by Dave Ruane, Director of Client Solutions at Lion People Global, featured M&A Principal Partner at Lion People Global, Olga Blasco, and amazing guest – Cristina Anselmi, an independent professional specializing in localization, machine translation, and AI implementation.

The Origins of an AI Localization Initiative
“Disruption makes people uncomfortable,” said Olga Blasco, and this statement perfectly encapsulates the journey of Cristina Anselmi, who has been at the forefront of AI-driven localization and machine translation in the gaming industry, where she led a major 6 year initiative to deploy AI (NMT) across an enterprise company.
It didn’t start with an elaborate plan managed by with BI, spreadsheets and detailed projections. Instead, it was an experiment, fueled by curiosity and a passion for technology. When she first entered the localization industry, statistical machine translation was the prevailing AI technology – far from ideal for the gaming sector. However, in 2017, the introduction of Google Neural Machine Translation (NMT) sparked her interest in “what if”.
At the time, an unused account with another translation provider presented an opportunity. Cristina took the initiative to train models and test their performance, quickly realizing the potential impact of this technology. Enthused by her findings, she shared them with her team and management, proposing an official trial. But innovation, especially in gaming, often meets resistance. While a few people were excited about the experiment, the majority were skeptical. Before moving forward, she realized, “We needed to think about change management.”
Navigating Resistance: From Skepticism to Adoption
Every major technological shift faces resistance. “Are you going to rip the band-aid slowly or quickly?” Blasco asked, highlighting the dilemma organizations faced when machine translation became an unavoidable reality. In 2017, the shift from statistical to neural machine translation was not just an upgrade – it was a fundamental change. Many feared for their jobs, unsure of how AI would reshape their roles.
Anselmi’s approach was strategic: she didn’t demand blind trust in AI but instead relied on incremental implementation. With management support, she built a task force representing different functions, ensuring a holistic approach. Within six months, the project moved from the experimental phase to full-fledged deployment. “We didn’t have a specific point we needed to reach. We just said, ‘Let’s do it and see how it works.’” This flexibility allowed them to fine-tune the system without the pressure of immediate success.
The Future of AI Deployment: Can We Still Innovate?
“Today, AI adoption is no longer a choice,” Anselmi states. In 2017, she had the luxury of experimentation, but the modern AI landscape demands structured, data-driven strategies. “You need to rely on data and not let biases – positive or negative – cloud your decision-making.” The challenge isn’t just about implementing AI, it’s about managing human responses to it.
Organizations that prioritize transparency and structured adoption will find it easier to navigate the inevitable disruptions AI brings. As AI continues to evolve, businesses that embrace incremental change and involve stakeholders early will be best positioned to thrive in the digital age.
The Power of Flexibility: Adapting Content for Success
One of the key factors in achieving success with content strategies is the ability to be flexible and open to change. Cristina shared that when the team explored various content types, it became evident that creative content, in particular, demanded a dynamic approach. “There are so many different kinds of creativity,” she noted, emphasizing the need to test different formats and styles. To navigate this complexity, they embraced trial and error. Testing different methods helped to determine what worked best for each scenario, allowing to pivot quickly if something wasn’t effective.
“Be bold and try, but don’t be afraid to declare failure and change it”
Was the guiding principle. The willingness to experiment and refine strategies based on feedback ultimately led to a more effective content framework that could adapt to various needs.
From Migration to Transformation: Overcoming Workforce Fears
Transitioning from traditional translation workflows to post-editing required new skills and a mindset shift. Initially, many professionals feared automation would replace them. “There was – and still is – a lot of fear around this technology,” admitted Cristina.
Instead of imposing change, the team focused on showing how jobs could evolve. Providing tailored training, setting clear guidelines, and consulting industry experts eased the transition. Listening to those affected by the shift proved essential. “It’s not about enforcing change, it’s about working together to make it effective.”
This collaborative approach increased efficiency and helped professionals appreciate how technology could enhance, rather than replace, human expertise.
Customization and Collaboration
As the project expanded, the team faced a critical decision: use multiple providers or stick with one. They chose the latter, and worked on fine-tuning for different franchises, languages, and content types. “Even with one provider, it was like having several”. Customization ensured optimal performance without the complexity of juggling multiple platforms.
Over six years, the program scaled from a few projects to over 200 engines, fueled by trust and collaboration. Initially met with reluctance, the initiative gained momentum as people saw its benefits. Managing this growth required a structured approach, so the team created a task force to gather insights and develop key performance metrics.
Watch a full session recording here:
Advancing AI in Localization: Challenges
Beyond human evaluation, discussions emerged around AI-driven quality estimation. Stakeholder education also became essential. AI isn’t magic; it requires expertise and careful implementation, Cristina emphasized. Misuse – such as applying AI without understanding localization challenges – could lead to broken game code or poor translations. The solution was balancing automation with human expertise.
Conclusion
Successfully implementing AI-driven change in enterprise localization requires bold leadership, adaptability, and collaboration. Overcoming resistance isn’t about forcing change but guiding people through controlled pilots, clear communication, and data-driven decision-making. As Cristina put it:
“You’re never 100% ready – be bold, take the risk, and adapt as you go.”
Despite AI’s power, the human element remains critical. AI’s success depends on teamwork, feedback, and trust. As Olga highlighted:
“Disruption makes people uncomfortable, but when managed well, it leads to transformation.”
Balancing AI advancements with human expertise, continuous learning, and an openness to change will drive the future of localization forward.
Join us for future Elevate sessions and stay ahead of the curve in innovation, leadership, and business growth!