How to choose English training for a multinational team

The short answer

The best English training for a multinational team is the one that trains each person in the English of their own function, under some pressure, with progress a manager can actually read. No single format does that for everyone on the team. A cohort course, a tutor and a self-serve app each win in a different band of the same team, and the buying mistake is almost always the same one: choosing one format for all twenty people because procurement wants one invoice.

I should say up front that I run a one-person company that sells one of these formats. Read the last section with that in mind. Everything before it is what I would tell a friend who has to spend a training budget next quarter.

What teams get wrong when buying

Most corporate English purchases go wrong before the vendor is chosen, because the brief is wrong. Four patterns come up again and again.

Buying for the certificate, not the meeting. The team is already B2 or C1 on paper. What they cannot do is hold their ground when the quarterly review turns hostile, or push back on a supplier without sounding rude, or write the memo that gets read rather than skimmed. A course that ends in another certificate solves the problem the team does not have.

Buying "Business English". Generic Business English teaches the average of every job, which is nobody's job. The controller does not need to "circle back". She needs to defend a Q4 variance to a board that reads numbers faster than sentences. The supply-chain lead needs to say no to a supplier in a way that keeps the supplier. Same language, different worlds. A programme that serves both by serving neither is what you get when the syllabus was written before anyone asked what the team does.

Buying seat time. Ninety-minute Thursday classes feel like training. They are also the first thing cancelled when the quarter gets busy, and attendance is the metric everyone reports because it is the only one the format produces.

Buying one format for everyone. The three people who present to clients, the eight who write reports, the two who are still at B1 and the one senior engineer who is fluent but silent in English meetings do not need the same thing. Treat them as one cohort and you will over-serve some and lose the rest.

Six criteria that actually separate the options

Use these as questions to every vendor, in this order.

1. Does it adapt per learner, in level and in function? Not "we offer levels A1 to C2". Ask whether two people on the same team, one in compliance and one in product, would receive different material on day one. If the answer involves the phrase "we can customise on request", it is a no.

2. Does it use the team's actual vocabulary? The documents your people produce, the meetings they sit in, the objections they field. Ask to see one sample unit built from something like your real work. If the sample is about booking a hotel or "negotiating a contract" in the abstract, it will not transfer.

3. Can people practise under pressure? Reading and listening are cheap to provide and comfortable to consume. The skill that gates careers is speaking while being interrupted, challenged or asked "why not the other approach". Any format worth paying for makes people do that, repeatedly, with someone or something that does not let them off the hook.

4. How is progress evidenced? Attendance is not progress. Ask what a manager sees after week four and week twelve: scores on real tasks, recorded role-plays, a written piece before and after. If the only evidence is a completion certificate, you are buying seat time with extra steps.

5. Does it fit the working week? Five sessions of fifteen minutes survive a bad quarter. One ninety-minute block does not. Ask what happens to a learner who misses a week, and whether the programme retunes or simply moves on.

6. What happens at week twelve? Enrolment is easy; completion is the product. Ask for the vendor's completion rate, defined as people who finished the programme rather than people who started it. Many vendors will not have that number to hand, which tells you something too.

The three formats, compared fairly

There are really only three shapes on the market, whatever the branding says. Each has an honest case.

In-person or live-online cohort courses

A trainer, a group of six to twelve, a fixed schedule, a syllabus.

Where it wins. Shared vocabulary and norms across a team that must present together; presentation skills that benefit from a live audience; cultural and register questions that are best argued about in a room; leadership buy-in, because a visible course signals that the company takes the language policy seriously.

Where it fails. Personalisation is structurally limited: one syllabus for the group means the material is right for nobody in particular. Scheduling across time zones is brutal for distributed teams. The pressure is real but rationed: in a group of ten, each person speaks for a few minutes an hour. Costs scale with headcount and with the trainer's calendar, and the programme ends when the contract does.

Buy it for: the client-facing five, or a leadership team that needs to present a strategy in English next quarter.

One-to-one tutoring

A tutor, one learner, a schedule they set together.

Where it wins. The deepest personalisation available: a good tutor will work from your actual slides and emails. Accountability is high because someone is waiting for you. High-stakes preparation, such as a board presentation or a regulator interview, is best done with a human who can read the room you are about to enter.

Where it fails. Cost per learner is the highest of the three, so it is rationed to executives, which means the people with the most language risk, the mid-level specialists who write and speak every day, rarely get it. Quality varies enormously and is hard to evaluate before you have paid for ten sessions. Progress evidence is whatever the tutor writes in an email. And the learner practises with one person's pushback style; the real meeting has six.

Buy it for: two or three people with an identifiable high-stakes moment in the next ninety days.

Self-serve programmes and apps

Software that builds and delivers the programme, with or without human touchpoints.

Where it wins. It is the only format that can give each of twenty people material built from their own function at a cost that does not scale with a trainer's hours. It fits fifteen-minute slots, survives a bad quarter, works across time zones, and produces evidence automatically because every exercise is scored. The better ones now rehearse speaking under pressure with role-play partners that argue back.

Where it fails. Nobody is waiting for you, so completion depends on the programme's rhythm and on the learner's reason to finish. Beginners are poorly served: software is bad at the patient, forgiving repetition that A1 to B1 needs. Certification preparation is a different job, with a fixed syllabus and exam technique, and general-purpose apps handle it badly. And the human dimension of a live room, the shared team norms and the visible signal to leadership, is absent unless you add it yourself.

Buy it for: the distributed majority of the team who are already B2 or better and need function-specific practice they will actually do.

How the formats fit together

The decision is not which format, but which people get which. A pattern that works for a team of twenty in a company that has just switched to English:

• The two or three with a high-stakes moment in the next quarter: a tutor, six to ten sessions, tied to the actual event.
• The client-facing group: a short live cohort on presenting and handling questions, four to six sessions, with a real internal audience for the last one.
• Everyone at B2 or above who writes, reviews and sits in meetings: a self-serve programme built from their function, run for a fixed term with a visible finish line, reported to the manager on scores rather than attendance.
• Anyone below B2: a general course first. Function-specific training on top of missing foundations does not stick.

Budget follows the same logic. Tutoring is the expensive line and should be small. The cohort is a one-off. The self-serve programme covers the most people for the least money, and it quietly decides whether the language policy works a year from now.

What nicheAce does and does not do

nicheAce: language is a self-serve programme, so it belongs in the third bucket above and nowhere else.

What it does. It opens with a conversation about the learner's job, field and where they get stuck, and generates a fifteen-week journey from that: foundations, deepening, application, mastery. Every learning day is five exercises of about fifteen minutes each, five days a week: vocabulary retrieved before it is revealed, gap fills with real distractors, reading graded on meaning, a scenario, and a role-play with a character who has a goal of their own and pushes back. After each day the difficulty moves one step, up or down. Progress is scored per exercise and per day, so a learner can show a manager what happened in week four without a certificate standing in for evidence. It is built for people who are certified but not yet usable in their own function, which is the group most training budgets miss. If that is the gap on your team, the language page shows what a journey looks like.

What it does not do. It is not for beginners: B2 or an equivalent is the floor, and below that a general course is the right purchase. It does not prepare anyone for an exam. There is no live human and no team dashboard; evidence is per learner, and the learner shares it. It runs on Android in the EU today, with iOS next, and hands-free Audio Mode is live in English, French, German, Spanish and Italian. And it is honest about the effort: about seventy-five minutes a day, five days a week, for fifteen weeks. A team that will not make that time should buy something lighter and expect less.

I built it because I was the person the other formats missed: certified, silent in the room, and not senior enough to get a tutor. If your team has twelve of those, you now know which line of the budget they belong on.

A one-page checklist for the buying conversation

Ask each vendor, in writing:

1. Show me the day-one material for a compliance analyst and for a product manager on the same team. Are they different?
2. Show me one unit built from a document like ours.
3. How many minutes per week does each learner speak under challenge, and to whom?
4. What does a manager see at week four and week twelve?
5. What happens when a learner misses a week?
6. What percentage of enrolled learners finished the last three programmes?

The vendor who answers all six in plain language is the one to shortlist, whichever format they sell.