In January 2026, the Department for Education confirmed that AI tutoring tools will be available to schools in England by the end of 2027. If you run a tutoring agency like I do, you probably read that headline and immediately felt your stomach drop.
I understand the reaction, I get it, I spent fifteen years in education before founding Capital Tuition Group, and I've built software for this sector too (as you probably know, as you’re here reading this at TotalOnboarder). But I promise you it’s not quite as scary as you might think. We’ve got the actual numbers, and panic toward this latest AI tutoring announcement is honestly the wrong response. The market won’t collapse, it’s just going to diversify.
What has the government actually announced?
Let’s just ignore the headline for a second, because the detail matters here. The DfE has invited edtech and AI companies to co-create tutoring tools with teachers, aimed initially at pupils within Years 9 and 10, across English, maths, science and modern foreign languages.
Trials begin later this year, and the idea here is that these finished tools will reach schools by the end of 2027. The target is to allow 450,000 disadvantaged pupils, specifically children on free school meals, to get hands-on time with these specialised products. The DfE has drawn the boundary in plain language too: these AI tutoring tools are there to complement "high-quality face-to-face teaching, not replace it." Which makes sense, if you’ve ever asked ChatGPT a question about the Great Wall of China, and promptly watch it as it makes up half the facts, and states Jeff Bezos was involved, somehow.
Why now?
This isn't the government muscling into a thriving market looking to curb its growth or trash a burgeoning industry. It's the government trying to fill a hole it helped create. The National Tutoring Programme, which subsidised tutoring for state-school pupils after the pandemic, was scrapped in 2024. According to the National Education Union, almost 60% of schools have reduced their tutor offerings because the funds just aren’t there anymore. As its general secretary, Daniel Kebede, put it in February, that decision "pulled the rug from under schools", further concentrating tutoring "among the most affluent."
Read that way, the AI education tool rollout is a public-sector intervention for pupils whose families simply couldn’t afford tutoring.
Is AI coming for private tutoring?
So not all doom and gloom then? No, and here are a few numbers that should help settle your nerves a little more, too. According to the Sutton Trust's Private Tutoring 2026 report, 29% of secondary pupils in England and Wales now receive private tutoring, up from about 27% in 2019 and 18% twenty years ago. This is a market that has grown steadily for two decades, and it now sits at the highest level the Sutton Trust has ever seemingly recorded.
That growth isn't evenly spread, mind you. In London, 45% of pupils are enrolled in some form of private tuition, whereas the rest of England sits at about 27% or so. Among the best-off families, it's 30%, compared with 23% for those in financial hardship.
And here's the telling part: the NEU anticipates that "wealthier families turn to human tutors to avoid AI-based learning". That’s likely because they know parents understand the risks that come along with AI hallucinations and value direct tutoring with experts who are trained in their respective fields.
The union raises that as an equity concern, admittedly, and they're right to, because it widens an already unfair gap. But it also tells you something plain about the demand. The households paying for private tuition aren't doing so for want of an alternative. They want a skilled specialist that’s accountable to them, invested in their child. A free AI tutoring tool aimed at a Year 9 pupil on free school meals, although helpful, still can't compete with that offering.
What AI tutoring can and cannot do
Let’s be fair to this technology, right, because dismissing it is as lazy as fearing it. AI tutoring is genuinely good at certain things: retrieval practice, instant feedback, adaptive difficulty, and patient, unlimited repetition of the material that simply needs drilling. It’s at its heart, a maths machine, doing huge sums and calculations based on a wildly large dataset. So handling times tables questions or advanced algebra is relatively child’s play compared to the calculations it needs to compute to output a single sentence.
For a pupil who would otherwise get nothing, that's a real and welcome gain, or at least if it works. And that "if" is doing real work: the NEU's position is that "the evidence for AI tutoring simply isn't there at the moment," and they've called for independent research, safety standards and professional oversight before it's leaned on. That's a sensible bar by any measure, and it's one reason the 2027 timeline might prove a little more optimistic than real.
Either way, the things AI tutoring can't do are predominantly the things parents are actually hiring tutors for. It can't read a teenager's face and notice they've quietly checked out. It can't ring a parent after a wobble before exams. It can't be held accountable, it can’t build a relationship across two academic years, and it can’t be that steady presence in the room when a child's confidence collapses. The DfE knows this, it's precisely why "not replacing teachers" is written into the announcement.
What this means for tutoring agencies
So the question has to be "am I clearly offering the thing the AI tutoring can't?"
In practice, that means leaning into everything human and verifiable about your service. The competition that should worry you isn't a chatbot; it's the agency down the road with properly vetted, well-trained, professionally onboarded tutors and the paperwork to back it all up. In fact, this is exactly the trust gap most parents don't even know exists.
As the national conversation fixates on AI and safety in education, parents start asking sharper questions about who, exactly, is teaching their child. Being able to answer them cleanly, with DBS checks, references, safeguarding, and qualifications, all in order in an easy-to-demonstrate way, is fast becoming the real differentiator. That bar has nothing to do with AI, and clearing it has never mattered more. Obviously, we're passionate about that here, but in a world where certification and authenticity are so highly valued, and it's so easy to spoof so much, compliance and having the proper checks behind what you do has become far more important than ever before.
AI tutoring in schools, though, well, that's inevitable really, it's coming, and for children who currently get nothing, it could be a genuinely game-changing toolset that helps them improve in a rapidly advancing technological world. It's not the end of our business though. The private tutoring market is at a twenty-year high, and the quality that makes it valuable, a trusted, accountable human being, is exactly the thing software cannot ship. So don't panic, not just yet.