Webinar Recap: Lost in Translation? The New Rules for Global Brands Scaling at Speed

LILT CMO Matt Singer and Lenovo's Senior Digital Marketing Leader Ashwin Naiksatam on how global brands can stop trading quality for speed and build a localization model that keeps brand voice intact across every market.

Every marketer knows the power of a distinct brand voice. The harder question in 2026 is how to protect that voice across dozens of markets without slowing down. In a recent webinar hosted by Marketing Dive, LILT Chief Marketing Officer Matt Singer and Ashwin Naiksatam, Senior Digital Marketing Leader for Acquisition & Growth at Lenovo, tackled that tension head-on and made the case that brands no longer have to sacrifice quality for speed to produce multilingual content at scale.

The numbers set the stage. A 2026 survey from LILT found that 96% of the people responsible for translation say quality is “mission critical,” yet only 57% believe they currently maintain a consistent brand voice across languages. That gap between ambition and multilingual execution is exactly where global brands are getting lost in translation.

How global brands keep brand voice consistent across languages

Ashwin opened with a familiar challenge for any global organization: the constant tension between centralized efficiency and local-market relevance. Lenovo had long operated in a decentralized way, but it became clear the team needed to optimize for three things at once: speed, quality, and savings.

Lenovo onboarded LILT in Europe first, specifically to solve quality challenges like inconsistent tone of voice. The results reframed what “efficient” could mean. The team moved from human review of all content to roughly half, and is now at the point where the AI handles the overwhelming majority of the work, all without sacrificing quality. A LILT specialist worked directly with Lenovo’s editorial team so the engine understood the brand’s tone of voice, its messaging, and the nuances of each locale it localizes for.

“The proof has been in the quality and the savings. As we keep working with LILT, the engine is learning even faster.” ~ Ashwin Naiksatam, Senior Marketer, Acquisition & Growth, Lenovo

That trajectory is backed up by Lenovo’s broader results with LILT. Across its global eCommerce business, Lenovo now runs 60+ domain-specific AI models across 39 key languages, and saw delivery timelines fall by up to 60% while cutting average annual translation costs by more than 50% versus its legacy provider.

How continuously-learning AI translation models lower cost

Matt stepped back to explain what makes LILT different. Most companies today start with a general-purpose LLM and put humans in front of it to check the work. LILT offers two paths instead: a brand can bring its own model, which LILT fine-tunes to optimize translation, or LILT builds a proprietary model specific to that brand.

The distinction matters. Where some competitors fine-tune only in weekly or monthly batches, and aggregate learnings across multiple customers, LILT’s models are proprietary to each brand and updated continuously. Every correction a human verifier makes feeds back into the system, so accuracy compounds over time. That feedback loop is what allows LILT to steadily reduce the amount of human verification required, which is where the savings come from.

It’s also why a single company often runs multiple models: separate ones for product, marketing, HR, and other functions, each tuned to maintain a consistent brand, respect compliance requirements, capture cultural nuance, and launch in multiple languages faster.

Transcreation vs. translation: localization as customer experience

One of the sharpest points of the session was the shift in how marketers should think about multilingual content. In a digital-first world, it isn’t a back-end task; it’s a core part of the customer experience, spanning the entire journey from search to conversion.

Ashwin made the case for transcreation over literal translation: not converting words directly, but preserving meaning and impact so a customer experiences the whole journey in their native language. That’s what protects brand strength. Lenovo, for instance, positions its laptops as “affordable premium,” never “cheap” or “budget,” and uses a different voice depending on whether the audience is buying for work, gaming, or design. With tone of voice built in, engagement has improved at scale.

“As we say, if you lose the week, you lose the revenue.” ~ Ashwin Naiksatam, Lenovo

Matt framed this as LILT’s core value proposition: multilingual content is far more nuanced than “translation,” and that complexity (steps, workflows, approvals, brand-guideline mapping, cultural nuance) is precisely what makes it hard for marketers.

How AI agents are changing the localization workflow

Historically, a localization professional was the glue between a content creator and the system doing the translation. Agents disrupt that. As Matt described it, you can now explain exactly what you need and hand over your guidelines; the agent asks clarifying questions, you reach alignment in conversation, and it then autonomously runs the workflows.

For Lenovo, that flexibility pays off across very different touchpoints. The website offers large content libraries for the engine to draw from; ads need far less copy but far more precision. A brief can now carry the character limit, the audience, and the psychographics. Those elements are what Ashwin called the “secret sauce” that lets the system articulate the right message more consistently. A promotion called a “sale” in the UK doesn’t translate directly into French, but transcreation handles it, and Lenovo can now deploy that messaging in under 24 hours, freeing marketers for strategy, reporting, and analytics rather than rote translation.

Submit localization from the tools teams already use (130+ connectors)

Speed is a leadership mandate, but localization has traditionally been a bottleneck. Matt pointed to LILT’s 130+ connectors, including Adobe and Figma, that let teams submit projects directly from the applications they work in every day and get transcreation done natively. The payoff isn’t just convenience: managers get the insights they need to focus on the decisions that matter most, and teams avoid the constant context-switching that introduces errors. Lenovo is now beginning to scale its use of connectors across the business.

Why human-in-the-loop review still guarantees quality

Even as agents handle the vast majority of content, Ashwin was emphatic that the remaining sliver is what guarantees quality, and offered the session’s most memorable analogy:

“If you ask a chef, no matter how much he teaches his team, the food doesn’t go out to the customer unless it’s tasted. Human intervention is the last part of quality assurance, and it will always remain.” ~ Ashwin Naiksatam, Lenovo

That human-in-the-loop philosophy is the same engine behind the model’s improvement. As one of Lenovo’s eCommerce leaders put it in LILT’s customer story, the retraining loop quickly began lifting already-high quality levels:

“The AI model did all the heavy lifting and continues to do so, which is important for us running across so many countries and languages with only limited dedicated support resource.” ~ Angus Cormie, Director & GM, EMEA eCommerce, Lenovo

What's next: connectors, MCP, and the company LLM

Looking 12 to 18 months out, both speakers expect the relationship between marketing and localization to grow more creative and more intelligent, with agents increasingly suggesting ideas grounded in a brand’s tone of voice and prior conversations.

Matt sketched a third way of working that’s emerging alongside connectors and the LILT platform itself: the company’s own internal LLM. He predicts that within one to two years, most enterprises will have a white-labeled LLM on every employee’s desktop, and LILT is building to support it. Through its Model Context Protocol integration, LILT is starting to embed multilingual translation directly inside AI assistants, so that accessing translation becomes as routine as opening a cloud app, just inside a company’s walled-off environment.

Key takeaways for marketing leaders

The throughline of the conversation was a reframing: localization is no longer a cost center or a back-office chore. It’s a growth lever that, done right, lets a brand show up consistently and quickly in every market it serves. Lenovo’s experience shows the upside is real: dramatic gains in speed and cost without compromising the brand voice that earns customer trust, when AI does the heavy lifting and humans stay in the loop where it counts.

--------------------------------------------------------------------------------

See how Lenovo unified translation with LILT to reduce costs, increase speed, and scale its global eCommerce experience.

Read the Lenovo customer story · Book a demo with LILT

--------------------------------------------------------------------------------

Want to watch the full webinar?

Catch the complete conversation on demand here.