Thirteen functioning legal knowledge subjects. Two papers of 180 single-best-answer questions each. FLK1 sat across two sessions of 2 hours 33 minutes on one day, FLK2 on another. Somewhere in there you are also meant to be accumulating two years of qualifying work experience, keeping a job, and remembering what a resulting trust is.
Nobody revises for that with a highlighter and a ring binder any more. Not successfully, anyway. And that isn't because highlighters stopped working — it's because the assessment itself changed shape, and legal education quietly rebuilt itself around the new shape while most candidates were still worrying about which textbook to buy.
The exam moved first. The teaching followed.
The old route to solicitor qualification in England and Wales was, at its heart, a taught course with assessments bolted on. You attended, you were examined on what you'd been taught, and the syllabus you sat was the syllabus your provider had delivered. Coverage was somebody else's problem.
The SQE inverts that relationship. The SRA sets a functioning legal knowledge specification and assesses it centrally at computer-based test centres. Nobody is obliged to teach you anything at all. You could, in theory, walk in off the street with a qualifying degree, two years of QWE and a library card. The regulator's position is that it assesses competence, not attendance.
That single design choice is what pushed technology to the centre of SQE exam preparation. When the exam is centrally set, breadth becomes the enemy. When it's 360 MCQs across Business Law, Dispute Resolution, Contract, Tort, Legal System, Constitutional and Administrative Law including retained EU law, Legal Services, Property Practice, Wills and Administration of Estates, Solicitors Accounts, Land Law, Trusts and Criminal Law and Practice — you are no longer trying to understand a subject. You are trying to hold thirteen of them at recall-level readiness simultaneously, on one particular Tuesday.
Human memory is spectacularly bad at that without help. Which is precisely the problem software is good at.
Why "how to become a solicitor UK" stopped being a linear answer
There's a second effect worth naming. The old route was sequential: degree, then course, then training contract. The SQE route is deliberately modular — you can sit assessments and build qualifying work experience in almost any order, provided you've met the SRA requirements before admission. Paralegals qualify while working. Career-changers revise Trusts at 6am before the school run.
A cohort that fragmented is a cohort that cannot all sit in the same lecture theatre at 9am on a Wednesday. Asynchronous, mobile, offline-capable study tools aren't a luxury for that cohort. They're the only delivery mechanism that fits the life.
What the technology is actually doing to your memory
Strip away the marketing language and most credible SQE revision platforms are doing three things, all of which come from decades of cognitive psychology rather than from anything specifically legal.
Retrieval practice. Being tested on material is not merely a measurement of learning; it is learning. Pulling the elements of misrepresentation out of your head is a substantially stronger memory event than reading them off a page. This is the single most robust finding in the whole field, and it maps almost perfectly onto an MCQ-based assessment. A question bank isn't a revision aid bolted onto your notes. It's the primary mechanism.
Spaced repetition. Material re-tested at expanding intervals sticks dramatically better than material crammed. Doing this by hand across thirteen subjects and several thousand discrete points of law is administratively impossible. Doing it in software is trivial — the engine simply tracks what you got wrong, when you last saw it, and how confident your answer pattern looks, then surfaces it again at the right moment.
Diagnostic granularity. This is the underrated one. A paper mock tells you that you scored 61%. Tagged analytics tell you that you scored 61% because your Land Law is fine, your Trusts is fine, and your Solicitors Accounts is quietly haemorrhaging marks on one specific sub-topic. Those are completely different pieces of information, and only one of them tells you what to do on Saturday morning.
The candidates who struggle aren't usually the ones who worked less. They're the ones who worked hard on the seven subjects they already liked, and found out in the test centre which six they'd been avoiding.
A worked example: Priya and the disbursement problem
Priya is a paralegal in Leeds, sitting FLK2. She's competent, disciplined, and doing roughly ninety questions an evening. Her overall accuracy sits at 72% and has been stuck there for three weeks.
Her instinct was to do more questions. What the analytics showed instead: Solicitors Accounts accuracy of 54%, against high-70s everywhere else in FLK2. Drill into that, and it wasn't Solicitors Accounts generally — Trusts of client money, breaches and the residual balance rules were all fine. The failures clustered almost entirely on one fact pattern shape.
The pattern went like this. A firm acting on a purchase pays a £310 search fee from the office account on Monday. On Thursday the client transfers £2,000 "on account of costs and expenses". Question asks which entries are correct. Priya kept treating the £2,000 as a single undifferentiated receipt and reaching for the client account every time, rather than separating out what was properly a reimbursement of a disbursement already paid from office money. Same underlying confusion, dressed in a dozen different scenarios: money received in advance for something not yet incurred versus money received to reimburse something the firm has already funded.
That's a forty-minute fix. Not a three-week grind. She read the relevant rules once, worked through eleven tagged questions on that exact sub-topic, got the last seven right, and her FLK2 accuracy moved four points inside a fortnight — because that one blind spot had been silently taxing every mixed mock she sat.
No amount of re-reading a textbook would have surfaced that. The pattern was only visible because something was counting.
The AI question, honestly answered
Every conversation about technology in legal education in 2026 arrives at generative AI within about ninety seconds, so let's deal with it properly rather than politely.
Used well, an AI tutor solves a genuine problem: the follow-up question. You get an MCQ wrong on remoteness in Tort. The explanation tells you why option C is correct. But your actual confusion is one level down — you don't see how this squares with the eggshell skull principle, and there is nobody in your kitchen at 11pm to ask. Historically that gap either got closed at a tutorial you may not have had access to, or it didn't get closed at all.
For international candidates the value is sharper still. A significant slice of the SQE cohort is doing this in a second or third language, and the cognitive load of parsing an unfamiliar English legal construction is real load that isn't available for legal reasoning. Being able to ask "explain the difference between an equitable and a legal interest in land" and get the answer in Mandarin, Japanese or Korean, then return to practising in English, is a genuine accessibility gain rather than a gimmick. It's one of the reasons multilingual explanation has become a standard expectation of any serious platform — the Ant Law SQE Question Bank, for instance, lets you interrogate any question in the bank conversationally in your own language rather than leaving you stuck with a two-line explanation.
Where it goes badly wrong
Now the warning, and I'd rather over-egg it than under-egg it.
- General-purpose chatbots invent English law. Fluently. They will produce a case name that sounds exactly like a real one, attach a plausible year, and state a proposition that is subtly or entirely wrong. In a discipline where the difference between two answers is one word in a statutory test, that is not a small risk.
- They are trained on a global corpus. Ask a broad question about "contract law" and you may quietly receive American doctrine. Consideration, promissory estoppel, remedies — the divergences are exactly where the exam probes.
- Asking is not learning. Reading a beautifully clear AI explanation produces a strong feeling of understanding and a weak memory trace. The feeling is called fluency illusion and it has ruined a lot of otherwise good revision. If you didn't retrieve it, you didn't learn it.
- It can't tell you what the SRA will actually ask. Only the specification can do that, and only sqe.sra.org.uk carries the authoritative current version.
The practical rule I'd give any candidate: use AI as a tutor sitting beside a curated, syllabus-mapped question bank, never as a substitute for one. The question bank supplies the ground truth and the retrieval; the AI closes the comprehension gaps the explanations leave behind. Reverse that order and you are learning law from a very confident stranger.
What technology still can't do for you
It would be dishonest to write an article about educational tech without drawing the line clearly.
| Technology handles well | Technology handles badly or not at all |
|---|---|
| Scheduling review of 13 subjects at scale | Building the professional judgement SQE2 assesses |
| Surfacing blind spots you can't feel | Sitting a client interview with a live role-play actor |
| Simulating 180-question timing pressure | Recording and evidencing your qualifying work experience |
| Answering follow-ups at 11pm in your own language | Character and suitability — an SRA process, full stop |
SQE2 in particular resists automation more than people expect. Five skills — Client Interviewing with attendance note and legal analysis, Advocacy, Case and Matter Analysis, Legal Research, Legal Writing and Drafting — assessed orally and in writing across five practice areas. You can drill legal research technique in an app. You cannot learn to hold a nervous client's confidence, or to stop talking over an assessor, without doing it out loud with another human being who will tell you honestly that you were rushing.
Similarly, qualifying work experience is not a study problem. QWE is about doing real legal work and having it confirmed by a solicitor or COLP, within a framework the SRA sets out in detail on sra.org.uk. Software can remind you to log a matter; it cannot generate the two years.
And on SQE pass rates: they are published by the SRA after each sitting, they move, and candidates who have taken a prep course have historically performed differently from those who haven't — but the causation there is muddier than anyone selling anything would like you to believe. Read the SRA's own statistical reports for the current figures rather than trusting a number you saw quoted in a forum eighteen months ago.
How to judge whether a platform is actually helping you
The market has more noise in it every year. If you're trying to identify the best SQE question bank for your own revision — and "best" here genuinely means best-for-you, not best-in-general — these are the questions I'd put to any product before paying for it:
- Is the bank mapped to the SRA specification, subject and sub-topic? Not "covers Land Law" — actually tagged, so you can practise co-ownership in isolation when co-ownership is what's broken.
- Are the questions genuinely single-best-answer? Real SQE1 items are applied. A fact pattern, a plausible-looking distractor set, one defensible answer. Anything that reads like a definition quiz is training you for a different exam.
- Do the mocks replicate the timing ratio? A 90-question, 180-minute sitting reproduces the real minute-per-question pressure without demanding a whole day. That pressure is the thing most candidates under-rehearse.
- Does it tell you what you got wrong, or just how much? A wrong-answer book, favourites, personal notes and subject-level mastery analytics are the difference between practice and revision.
- Does the engine choose your next question, or do you? Left to your own devices you will re-practise what you're good at. That's human. Good engines prioritise unseen items, past errors and low-accuracy topics whether you fancy it or not.
- Can you use it in the ten minutes you actually have? On a phone, on a train, without your laptop. The revision that happens beats the revision that's optimal.
Apply those six honestly and you'll narrow the field fast. If you want a second opinion beyond the Ant Law SQE Question Bank, CELE SQE is worth a look — but interrogate it against the same list rather than taking anyone's word, including mine.
The point of all of it
Technology hasn't made the SQE easier. Thirteen subjects is still thirteen subjects, and the SRA is still going to put 180 questions in front of you and take away five hours and six minutes of your life in a single day.
What it has done is remove the excuses that used to be structural. You no longer need to be in a particular building, in a particular city, at a particular hour, to get high-quality applied practice with proper feedback. That's a meaningful redistribution of who gets to become a solicitor in England and Wales — and it's the strongest argument for taking these tools seriously rather than treating them as a supplement to "real" studying.
Here's your next step, and it's a small one. Pick your weakest FLK subject — the one you've been quietly not revising. Sit twenty tagged questions on it tonight, timed. Look at what the analytics say afterwards rather than at the score. You'll likely find your problem is narrower and more fixable than the general dread suggested.
You can do exactly that at antlaw.ai, on web, iOS or Android, with 14,000+ syllabus-mapped FLK1 and FLK2 questions and an AI tutor that will take your follow-up questions in whichever language you think fastest in. Check sqe.sra.org.uk for current sitting dates, fees and the authoritative specification — and if you get stuck, [email protected] is a real inbox with real people behind it.