For two years, a logistics manager in Poland avoided picking up the phone.
His English was good. He'd done the courses. He had the vocabulary. In spite of his language training and knowledge, he preferred email. It was much slower but also felt safer and more controllable. He could check his words before they left him.
Then a shipment got stuck at an American port. The problem was too complex for another email thread. So he picked up the phone, navigated the conversation, understood the problem properly, and found a solution that got the shipment released early, heading off a serious customer escalation in the process.
He was "absolutely categorically convinced," his L&D lead shared, that had he carried on with email, he would never have had the insight that changed the outcome.
He is not an exception. Across every multinational I've spoken with, there are people who have completed the language courses, hit the metrics, earned the certificates, and still default to the safer, slower option when the pressure is on. They had the knowledge, but lacked confidence in the moments that count.
That gap, between knowing a language and performing in it, is where the real value is won or lost. And most organisations are not yet designed to close it.
The world of work changed. The bar for human performance rose permanently.
I've been thinking lot recently about the shift from organisations optimised around tasks and knowledge to systems designed around judgement and capability. It's a shift I see playing out in every sector, every function.
AI has made knowledge abundant. It has not replaced the need for human beings to walk into high-stakes rooms and perform; under pressure, in another language, without a script.
When a large client relocated its operations from one European country to another, the business made English the operating language of the entire company, essentially overnight. For L&D, that meant taking a largely non English-speaking workforce to working proficiency in English, at scale under real-time pressure. As, their L&D Business Partner, put it: it wasn't a learning initiative or a benefit. It was a strategic decision to ensure business continuity and support global expansion.
Not every organisation faces that level of urgency. But the underlying dynamic is the same everywhere. Research by The Economist found that 90% of executives believe that if cross-border communication improved, profit and market share would follow. With the proliferation of AI tools, English is becoming the de facto lingua franca of even more global workflows, turning multilingualism from a soft skill, to a hard operational requirement.
The organisations pulling ahead are shifting focus from more training to building capability for the moments that drive results, They are also changing the way they measure from tracking module completions to business outcomes.
Most organisations are measuring the wrong thing. And the cost is showing up in the moments that matter most.
I hear a version of the same conversation in almost every L&D room I sit in. Completions are up. Time-on-platform looks good. Pass rates are strong. But then someone asks: can your people actually perform differently because of this? And the room goes quiet.
The Head of Talent and Development at a client’s logistics business operating across 46 countries stated the measurement problem directly: "The certifications are less impressive to them than they are to me. What the business cares about is profitability, customers retained, and time saved. Language underpins all of that. It's just hard to make the link from a completion report."
That difficulty is more than a reporting problem. It is a design problem. The system was never built to surface what leadership actually needs to see.
The data tells the same story at the human level as we discovered through Busuu’s Language Anxiety at Work Survey. Fifty-five per cent of employees say they can't effectively convey their message under pressure. Nearly half dread unplanned conversations. The knowledge is there. The capability to deploy it under real conditions is not, because traditional learning programmes weren't designed with that purpose in mind.
As we know from the often cited CSA Research study, in service industries a single language friction incident can reduce a customer's likelihood to recommend a brand by 30%. For a logistics business managing international freight and resolving problems in real time, that number has a direct line to revenue. For a tourism company with 66,000 colleagues in over 150 countries, the same dynamic plays out at extraordinary scale.
The Head of Training and Development at a global travel group client, named something I think about often: "People who feel insecure in their second language don't vocalise their opinions and that leads to less good decision making." When the cost of under-performing in a language is silence in the room, an unresolved escalation, a deal that doesn't close, it is never attributed to the language gap. But it should be.
The gap has persisted because of how learning has been designed, and that design is changing.
Here is what I've come to believe: the gap between language training and language performance is structural, not motivational. Language learning has always happened away from work, in a course, a classroom, or an app, and learners have been expected to carry it into the moment that counts. That model has a ceiling.
In my own thinking about how to design for the AI era, I keep coming back to the same principle: the most important information only becomes visible when you act. In developing language capability, the moment of truth doesn't happen in the lesson. It happens in the conversation, the meeting, the phone call, the negotiation and the presentation in front of an audience.
Language learning must evolve to keep pace with business needs. It must be redesigned to connect learning to the moments and contexts of daily worklife. Offering support that arrives before the moments that matter, and practice that mirrors real work, in real pressure, with the vocabulary and scenarios that are actually relevant to someone's role. It must be a system that builds capability continuously, rather than front-loading knowledge and hoping it transfers.
This is more than a feature update. It is a different model. The measure of success shifts from "who completed the course" to "who can now perform differently in the conversation that determines whether a deal closes, a market opens, or a team aligns."
Eighteen years of language learning science, a growing data engine, and a human layer that AI alone will never replace: these are the foundations on which that model is being built. The organisations working with us are already starting to see the power of moving from completion to performance.
The question I'd leave you with
Think about the highest-stakes moment one of your people will face in the next 90 days. A major client call. A cross-border negotiation. A leadership presentation in a second language. A conversation that determines whether a problem gets resolved or escalates.
Are they ready? Can they walk in, perform under pressure, and come out the other side with the desired outcome?
That is what language performance means. That is what L&D needs to be designed around. And that is the future we are building at Busuu, one where capability is developed intentionally, measured at the moment of performance, and connected to the outcomes that actually matter to the business.
The organisations that design for this now will outperform. The window to lead is open.