Why Workplace English Stalls

Companies pay for English classes, count who turns up, and wonder why their staff stay stuck at the same level for years. A weekly class is too little time, nothing is practised between lessons, and nobody measures the English people actually use at work. The fix is the same cycle as the rest of this series: find the errors in the real work, practise them little and often, and read the work again.

Spanish firms take language training more seriously than most. In 2010, one in five of those that ran training courses counted languages among the main skills they trained for. By 2020 it was one in eight, still about twice the EU average (Eurostat, 2020).

The need has not gone away. EF’s ranking, built from people who took its free online test, puts Spain 36th of 123 countries in English, in its “moderate” band (EF EPI, 2025).

Language gaps cost business. In a European survey of nearly 2,000 exporting small and medium-sized firms, 11% said they had lost a contract because of a lack of language skills (ELAN, 2006).

The first problem is time. Cambridge English reckons on about 200 hours of guided study to move up one level. Two 90-minute classes a week over a 30-week year give about 90 hours, and Spanish workers average about 13 hours of company-funded training a year across all subjects (FUNDAE, 2025).

Class time alone will not move most people a level in a year, so what happens between classes decides the result.

The second problem is forgetting. In a US study of 733 people tested on the Spanish they learned at school, most of what they lost went in the first three to six years, and almost none of them had practised in between (Bahrick, 1984).

The third problem is the plateau. In a study of adults who left English courses at a language school in Barcelona, lack of time was the most common reason, and most of those who left were at intermediate level or above (Evans and Tragant, 2020). Intermediate is where progress slows and a course stops feeling worth the effort.

Spaced practice is the answer to time and forgetting. A meta-analysis of 48 experiments found that spacing practice out had a medium-to-large effect on second-language learning, and longer gaps helped people remember for longer (Kim and Webb, 2022).

Corrective feedback is the answer to the plateau. Meta-analyses of spoken and written feedback found that correcting learners’ errors improves their accuracy, and getting learners to correct themselves worked better than the teacher simply rephrasing (Lyster and Saito, 2010; Li, 2010; Kang and Han, 2015). The gains lasted.

I have taught English for manufacturing companies and in private classes for large corporate clients, and the errors are predictable. Spanish speakers writing in English tend to make the same ones: “actually” used to mean “currently”, “assist” a meeting instead of attend it, a sentence that starts “Is important” with no subject. An English teacher carries that map in their head. Written down, it lets a system spot the errors that matter in someone’s real emails and reports.

So the design starts with the work. The learner’s own emails, reports or recorded calls are read against that error map, and the teacher checks what the AI flags. Each person gets a short task most days, built from their own mistakes, and the same errors come back at spaced intervals until they stop appearing. The feedback asks the learner to fix the error before it explains it.

FeedForward Learning already reads English writing this way, with a teacher approving everything the AI says, and its spaced practice is being built.

Every few months the work is read again. The company gets a measure it recognises: fewer errors in the emails its clients read, reported by team and by skill, never by name. The learner sees their own progress against their own past.