Act 1: The Snapshot

We begin with our feet firmly planted in the present and reflect on where the language services industry currently stands (spoiler: there’s no way to avoid this thing called GenAI).

But do we understand where we are and where we are headed? Are we asking the right questions, or should we just start with the “Why”? 

There is a sense of urgency that powers the varied market players. We explore new pathways, test new solutions, learn from the mistakes made along the way, course-correct, and start again.

The industry is evolving, and we are in motion. In the following pages, we examine what this movement looks like. 

Forget everything you thought you knew about AI in localization, at least for a moment. As the industry fixated on the ever-improving capabilities of machine translation to drive cost and time efficiencies, we at Deliveroo faced a different challenge: What happens when your brand’s distinct voice and customer-centric content just don’t align with the mainstream AI translation approach? This question led us to completely reshape our mindset (and ultimately our localization strategy), bringing us back to what truly matters: the person on the other side of the screen.

The Deliveroo dilemma: Not all content is fit for AI. Now what?

For many organizations, the initial foray into AI for localization often starts with content amenable to machine translation (MT). MT typically performs best with long-form, repetitive, or low-context content. However, Deliveroo’s content landscape is markedly different: As a mobile and web application-driven company, the bulk of our language localization requirements are for research-informed, short-form, high-stakes user experience content and marketing creative.

Even deeper content challenges arise from the way our business is structured. As a three-sided marketplace, we have to localize for three distinct audiences — consumers, riders, and partners — each with their own needs, goals, and brand voice expectations. A Deliveroo feature update commonly involves localizing multiple intertwined user journeys: We might be explaining a new payment process to a consumer, a delivery route feature to a rider, and an order management journey to a partner, all within a single product release. These pieces of UX content have far more important jobs than simply existing in a given language; they’re designed to guide users, build trust, and foster a seamless experience for our diverse global audiences.

It is this intricate web of overlapping features, high-stakes content, and the complex marketplace that makes our content so reliant on human judgment — and beyond the reach of generic AI solutions.

Given these circumstances, the idea of using a human-in-the-loop model might seem like the logical next step. However, we’ve found that this approach often misses the point of what AI should be. The sheer volume of human input required for post-editing, detailed training, and prompt design for generative AI risks negating any initial efficiency gains. In other words: If AI isn’t genuinely enhancing the customer experience in a way that couldn’t be achieved as efficiently or even more efficiently by other means, then its application risks becoming a drain on resources rather than a strategic advantage.

Having analyzed our content and concluded that the ROI for MT/AI and human post-editing just isn’t there for the vast majority of our requirements, we were left with a crucial question: If we can’t use AI to localize the bulk of our content, does that mean we can afford to ignore it altogether?

The answer, we quickly realized, was a definitive no. Our content may not fit neatly into the usual MT or AI applications, but the pace of AI innovation — and the growing demand for instant, intelligent customer interactions — left no room for hesitation. Embracing AI isn’t optional: It’s an opportunity to reimagine the customer experience in practical and forward-thinking ways.

This imperative led us to another fundamental question: If our usual content types aren’t a good fit, what if not only the solution but also the use case itself should be novel? This became the guiding principle for our exploration. Instead of forcing our existing content into an ill-fitting AI framework, we began to consider how AI could enable entirely new forms of communication and service that align with our brand and benefit our users.

Our journey into novel AI use cases has been heavily influenced by understanding our user base – what they expect, how they interact with our products, and where their pain points lie. This isn’t just about market research, but a deep, continuous learning process from real-life user interactions that has pushed us in new and exciting directions.

By listening closely to our customers, riders, and partners, we uncovered a powerful opportunity to make our marketplace even more connected through seamless communication. It wasn’t about translating official content, but about enabling spontaneous, real-time conversations between riders and consumers, celebrating the rich diversity of languages in our community. That insight led to a new AI use case at Deliveroo: an instant, multilingual interpreter. This is a dynamic, AI-powered bridge that helps customers and riders connect effortlessly, collaborate in the moment, and complete deliveries with greater ease.

The real value we’re finding lies in understanding the unique aspects of our situation and our strong sense of brand
and culture, which is highly user-attuned. Our approach is now leading us down a path of exploration and testing on our own terms, rather than simply adopting off-the-shelf solutions.

As members of the Experience team at Deliveroo, our mission extends far beyond merely translating words. We are, fundamentally, in the business of cross-cultural mediation and impact. Every phrase, every user journey, every interaction we localize is a deliberate act of bridging cultural divides and ensuring our global users feel seen, understood, and seamlessly connected to our platform. This isn’t a checkbox exercise – it’s a core strategic imperative that directly influences customer satisfaction, operational efficiency, and ultimately, our brand’s global success.

This positioning of our role has shaped our approach to AI. While the industry often focuses on AI for sheer volume and speed in translation, our journey has revealed that its true power for a brand like Deliveroo lies in its capacity to unlock entirely new avenues for customer experience and cross-cultural engagement. We’re not using AI just because it’s available; we’re leveraging it to solve unique challenges and create value in previously unimaginable ways.

The future of localization is full of opportunity — but it’s not just about embracing the latest AI technology. At Deliveroo, we’ve seen firsthand that when your content doesn’t align with the conventional machine translation approach, it’s not a limitation. It’s a chance to rethink your strategy. Rather than chasing translation speed as a success metric, we focus on what actually matters: how our content performs in-market.

For us, that means using technology to support real cultural connection, not just linguistic accuracy. By tailoring localization efforts to what resonates with specific audiences, we unlock measurable results: Stronger brand loyalty, better customer retention, and a more stable presence in each market we serve.

Read the full 132-page Global Ambitions: (R)Evolution in Motion publication featuring vital perspectives from 31 industry leaders on the ongoing AI-spurred (r)evolution.

Get your digital copy today