English for game studios: why one course never fits

A game studio recently came to us with a request we hear more and more often: short English training for several teams at once, developers, QA testers, and artists. Not one course for everyone. Each group needed its own situations covered, plus the skills they all share, like presenting and running meetings. And they wanted to pay only for the parts each team actually needed.

Inside one company, the language changes with the job. A developer reviewing code speaks differently from an artist walking a lead through a model, a designer defending a mechanic under pushback, or a producer explaining a slipped milestone. Put them all in one business English class and most people spend half of every lesson on situations they will never face.

We have developed role-specific materials for game-development teams before, so the request landed on familiar ground. Much of the material already existed, which is the real advantage of an industry-specific provider over a generic one: the head start. We never reuse an old programme untouched, but starting further along means the adaptation goes to the team in front of us, not to building basics from zero.

Every programme starts with a needs analysis, the standard opening move in English for Specific Purposes and the step most providers skip. Two questions drive it: what do these people do in English at work, which meetings, documents, and calls do they dread; and where do they stand now, in CEFR terms, against what the job demands. Short conversations with team leads answer most of it, along with sitting in on a live call and reading what the teams already produce: tickets, patch notes, comment threads. From there the design runs backwards: we name what each person should be able to do by the end, and shape every lesson toward it.

The programme is assembled from blocks. Some belong to a single role and carry its vocabulary and situations. Others cover what nearly everyone in a studio needs: presentations, meetings, cross-team collaboration, writing. The client picks the blocks its teams need and leaves the rest.

Because shared blocks are reused across teams, the cost per head stays reasonable at scale. In a large studio, the block model can be used across five or more roles; the exact number of participants and blocks depends on the needs analysis. QA and artists can sit together in the same running-meetings session while each keeps the block built for their own work.

The material comes from the world the learners already live in: a conference talk, a set of patch notes, a thread where people argue over an art change. Attention is higher when the content looks like your own work, and the English transfers back to the job with far less friction.

Lessons are designed to maximise relevant learner speaking time, built around a task people will meet for real: running the stand-up, defending a decision, presenting the build, walking a stakeholder through a problem. In task-based activities, correction usually follows the speaking task so the learner can finish the communication first; immediate correction may still be used when it supports the task.

Mixed levels in one group are normal and planned for. Stronger speakers take the demanding roles in a role-play: the director, the stakeholder who needs convincing, the lead asking awkward questions. Newer speakers get sentence starters and preparation time, and the support falls away as they settle in.

Format follows the team. One-to-one gives the closest feedback and suits senior staff with specific needs. A small group offers what one-to-one cannot: a live meeting to run, a disagreement to handle, colleagues to respond to. AI runs in the background: it helps draft role-specific exercises faster and gives learners extra practice between lessons. Preparation gets quicker; the teaching stays with the teacher.

Progress shows up two ways: movement on the CEFR scale, and a before-and-after on the tasks themselves rather than a single final score. If the goal was clearer bug reports or calmer incident updates, that is what gets measured. A company paying to train a whole department deserves more than "everyone improved a bit."

This role-specific block model can also run inside a managed programme covering the wider company.

Start with one question: what does each team need to do in English, and where does it fall apart today? That answer tells you more than any average test score. Once you have it per role, choosing the right course is the easy part. This is regular work for us with game studios, and we are glad to compare notes.