Personalised English Learning: Why One-Size-Fits-All Fails

Every week I speak to adult learners who have spent months — sometimes years — in an English course that has not moved them forward the way they hoped. When I ask what the course covered, the answer is almost always a standard progression: present simple, past simple, conditionals, modal verbs, future tenses. In order. For everyone. The learner sitting next to a total beginner gets the same unit on the present perfect as the person who already uses it naturally and just needs to stop dropping the auxiliary in fast speech.

That is the fundamental problem with generic English instruction. It is not that the content is wrong — present perfect is genuinely important. It is that the sequencing, the pacing, and the specific errors drilled are optimised for nobody in particular. Personalised English learning fixes that by starting from your actual level and gaps, not the assumed average.

Why generic courses fail most learners

A generic course is designed for an imaginary average learner. That learner does not exist. Real learners differ across at least four dimensions that a fixed syllabus cannot handle at once: their current level (someone can have C1 reading and A2 speaking), their first-language background (which predicts which errors will keep recurring), their goal (business writing versus conversational fluency require very different vocabularies), and the specific errors they have already fossilised — the patterns that feel right to them because they have been producing them uncorrected for months or years.

When a course ignores these differences, learners either sit through material they already know, grinding through boredom, or find themselves pushed into content that is too far above their current production level to absorb. Both outcomes produce the same result: time spent without proportionate improvement. In my experience, the learners most harmed by generic instruction are not complete beginners — they are people at B1 and B2 who have been stuck at that level for a long time, because nothing in their course has targeted what is actually holding them back.

A course that treats every learner the same is not fair — it is just wrong for most of them. Real fairness means meeting people where they actually are.

Generic vs personalised: a side-by-side

The table below shows how each dimension of English learning looks under a generic syllabus versus a personalised path. The differences are not cosmetic — they change what you spend your time on and how quickly that time converts into real progress.

Start with a real diagnosis

The first step in building a personalised path is also the one most learners skip: a genuine diagnostic. Not a self-assessment — people are reliably poor at judging their own level, tending to underrate their reading while overrating their speaking. Not a multiple-choice placement test either, which measures recognition but not production. A useful diagnosis asks you to actually produce English: write a short email, answer two or three spoken questions on a recording, summarise a short passage. That output reveals the errors you actually make, not the errors you can spot on a page.

Once you know your real starting point, you can set specific targets. If your reading is B2 but your writing is B1, you have a skill gap to close, not a level problem. If you can use the present perfect correctly in drills but drop it consistently in free speech, that is a production problem under pressure — a very different drill requirement from someone who has never learned the form at all. This is the kind of granularity a generic course cannot offer, and without it you are just guessing which part of the syllabus to spend more time on.

The step-by-step method guide covers how to structure that work once you have your diagnosis.

How your first language shapes your errors

One of the most consistent patterns I see across learners is that first-language background reliably predicts certain error types. This is not a stereotype — it is linguistically well-documented, and knowing your own pattern is genuinely useful because it tells you where to focus before you have made every mistake yourself.

Spanish and Portuguese speakers generally transfer subject-verb-object word order cleanly and have good intuitions for verb conjugation, since their own verb systems are highly inflected. The areas that catch them out are articles — "a," "the," and zero article are used quite differently in English than in Spanish or Portuguese — and subject pronoun omission, which is grammatical in Spanish and Portuguese ("Estudio inglés") but wrong in English ("I study English," never just "Study English").

German speakers often produce grammatically ambitious sentences early, thanks to a shared Indo-European grammar base, but struggle with English verb positioning in subordinate clauses (in German, the verb goes to the end; in English, it stays in second position regardless of clause type) and with the absence of the formal/informal pronoun distinction that Sie/du marks in German.

Mandarin and Japanese speakers face a structurally different challenge: their languages lack grammatical articles and often omit plural marking, so "a dog," "the dog," and "dogs" all require conscious attention that feels entirely unnatural. Aspect marking — the difference between "I eat" and "I am eating" and "I have eaten" — is also encoded very differently, making the English tense-aspect system a recurring source of error.

A personalised learning path uses this knowledge from the beginning, not after you have drilled the same article exercises as someone whose first language is French and who already has a native feel for when to use a definite article.

Sources: Cambridge English — ELT research; Council of Europe — CEFR.

Where AI personalises well

Adaptive learning technology has made genuine progress in two areas where personalisation was previously difficult to scale: adjusting difficulty in real time, and tracking error patterns across many sessions.

A well-designed adaptive system does not show you the same exercise after you have answered it correctly three times in a row. It moves you on, raises the difficulty slightly, and comes back to the point later at intervals designed to strengthen long-term retention — the spaced-repetition principle that language learning research consistently supports. This kind of pacing is something no human instructor can deliver consistently for thirty students at once; a well-configured algorithm does it without effort.

Error tracking is the second real strength. If you consistently confuse "make" and "do," write "I am agree" instead of "I agree," or drop the "s" from third-person singular verbs in fast writing, a system that logs every response will surface those patterns within a few sessions and weight future exercises accordingly. It does this without the memory limitations and note-taking overhead a teacher faces in a busy group class.

The honest limit of AI personalisation is that it optimises within the content it has been given. It can tell you that you made an error; it is much less good at telling you whether fixing that error should be your priority right now, given your specific goal. That requires something closer to judgement.

Where a teacher personalises better

Personalisation by an experienced teacher is not about covering the same material more slowly or more quickly. It is about changing what you practise, based on what matters for your situation. A teacher working with someone preparing for a university interview and a teacher working with someone who needs to chair project meetings in English should give those learners almost entirely different vocabularies, different speaking tasks, and different feedback priorities — even if both learners are nominally at B2.

The same applies to error prioritisation. Not every error matters equally. A learner who occasionally misuses the past perfect but is otherwise fluent and clear does not need months of past-perfect drilling — the communication cost of that error is low, and their time is better spent elsewhere. A learner who regularly uses "since" where they mean "because" is creating genuine comprehension problems for listeners. A teacher recognises this; an algorithm generally cannot, because the algorithm has no model of what the learner is actually trying to communicate or who they need to communicate with.

This is what I mean when I say that personalised adaptive learning and personalised teaching are complementary rather than competing. Use the adaptive system for what it does efficiently — pacing, drilling, error logging. Use structured feedback from a teacher or programme for what requires contextual judgement: which gaps to close, in what order, for what purpose. The combination is considerably more effective than either alone.

For the mechanics of how to build that feedback into your practice, see the guide on when and how often to practise for maximum retention.

Building your personalised path

In practice, a personalised path looks like this. First, produce some English — write, speak, do not just choose answers from a multiple-choice list — and get that output assessed against the CEFR scale, skill by skill. Second, identify the two or three error patterns that recur most frequently and that carry the highest communication cost for your specific goal. Third, choose materials and drills that target those gaps, at a difficulty level slightly above your comfortable production level, not so high that you cannot understand the input. Fourth, review the pattern every four to six weeks: which errors have reduced, which persist, what the next priority should be.

That cycle — diagnose, target, practise, review — is not complicated. What it requires is honesty about your starting point and a willingness to drill the things you find difficult rather than the things that feel comfortable. Most learners default to practising what they are already good at, because it feels like progress. Real progress comes from the uncomfortable end of the spectrum.

If you want a starting point that does this without requiring you to build it all yourself, our free track is built around exactly this approach — it assesses where you are, targets the grammar and vocabulary gaps most common at your level, and corrects your own sentences the way an instructor would rather than just marking them right or wrong.

Start the free English track