If you have shopped for a tutoring app lately, you have met the pitch: personalized, neuroscience-based, tuned to read the way your child learns. This month a London startup called Medly AI raised 6.89 million euros, close to 8 million dollars, in seed funding, and the coverage repeated that promise almost word for word, describing a platform that intuits a student’s learning style and tailors its feedback to it. That one phrase is worth stopping on. Researchers tested whether matching teaching to a child’s learning style improves learning, found no credible evidence for it in 2008, and reconfirmed the null result in 2024. The branding is reaching for an idea the science retired almost two decades ago, and for a parent, the branding was never the thing to evaluate anyway.
AI tutors are everywhere, and nearly every one promises personalized, brain-based learning. Here are the questions parents are actually asking about what those promises mean, and how to tell a real learning tool from a well-marketed one.
Common questions
Do “neuroscience-based” learning apps work better than other ones?
Are learning styles real? Should a tutor match my child’s?
How do I know what kind of help my child actually needs?
Medly ran a study showing gains. Doesn’t that prove it works?
A new AI tutor sells ‘reads your child’s learning style.’ That idea failed testing in 2008 and again in 2024. Judge a learning tool by its mechanics, retrieval and real feedback, not by its branding.
What Medly raised, and what it claims
Medly AI is a London exam-preparation platform founded by Dr. Paul Jung and Dr. Kavi Samra, two UCL medical graduates who met in secondary school at age 11. Its 6.89 million euro seed round was led by Felix Capital, with existing investors Eka Ventures and Ada Ventures taking part, and it follows an earlier 1.7 million pound round in February 2025. The founders, who both describe growing up in low-income households, frame the company around access. As Jung put it: “Almost a third of 11 to 16-year-olds receive some form of private tutoring. For those who can afford it, it’s a powerful advantage. For those who can’t, it’s an equally powerful reinforcer of the status quo.” The platform reports more than 400,000 UK users and was named one of eight suppliers in the UK government’s AI Tutoring Tools Pioneers Programme in July 2026.
The pitch around the raise leaned on two ideas. Coverage described the product as built on “cutting-edge neuroscience and teaching theory,” running on a combination of large language models “tuned to read learning styles.” In an earlier announcement, Jung said the company had “developed a custom AI architecture” by “applying insights from my PhD in neuroscience,” and that “Medly AI adapts to each student’s unique learning style.” Hold onto those two claims, neuroscience and learning styles, because the evidence treats them differently from the way the marketing does.
On results, two different numbers travel together and deserve to be pulled apart. The headline figure in the funding coverage is a company-reported one: that 74% of students who used the platform for GCSE revision improved by at least one grade, and 41% by two grades or more. That is a before-and-after number from self-selected users, with no comparison group, which is the shape of claim that tells a reader almost nothing on its own. Separately, and to the company’s credit, Medly also has a genuine independent trial. A third-party evaluator, WhatWorked Education, ran a randomized controlled trial in GCSE science, and it found that students using the platform for 30 minutes a week over four weeks made more progress than a control group who spent the same time on their usual revision. The reported gains were modest, in the range of about 0.3 to 0.5 of a standard deviation by subject, and the trial was narrow: one subject area, four weeks, and roughly a third of students dropping out between the pre-test and the post-test, with no source confirming whether that dropout was balanced across the two groups.
Almost a third of 11 to 16-year-olds receive some form of private tutoring. For those who can afford it, it’s a powerful advantage. For those who can’t, it’s an equally powerful reinforcer of the status quo.
Laura Lurns · Learning Success expert
The branding is the noise. The mechanics are the signal.
Two instincts pull a parent toward a pitch like this. A founder with a neuroscience PhD sounds like a guarantee of rigor, and a tool that promises to teach your child “their way” sounds like exactly what a struggling child needs. Both instincts point at the wrong thing. The learning-styles promise is the clearer case, because it has been tested head-on. Pashler and colleagues reviewed the evidence in 2008 and found no credible support for the idea that matching instruction to a child’s preferred style improves learning, and a 2024 meta-analysis in Frontiers reached the same conclusion. This is one of the most durable null results in education. A product tuned to read and match a learning style is optimizing for something that was tested and did not hold up.
So what does make a learning tool actually teach, rather than merely occupy? The better-evidenced picture looks nothing like a learning-style match. It points toward effortful retrieval, a child pulling an answer out of memory instead of recognizing it on a screen; practice spaced out over time rather than crammed; and feedback that does more than mark an answer right or wrong, feedback that shows the next step in the reasoning. That is the direction the research on educational technology points, offered as direction rather than a settled verdict. It is also why our own exercises vary the task rather than drilling one format, so a skill sticks to the child rather than to one screen. What these mechanisms share is that they are invisible in a marketing headline, because they show up as harder work for the learner, not as a smoother, more engaging session.
The real problem is not any one company, and it is not artificial intelligence in a tutor. It is a habit of mind that has been with ed-tech since long before the AI wave: engagement mistaken for learning. Minutes in the app, lessons completed, streaks kept, messages exchanged, all of it feels like progress and photographs well in a pitch deck, and none of it is the same as a measured gain in what a child knows. A neuroscience story and a learning-style promise are engagement dressed as science. The question that actually sorts the good tools from the rest is whether the thing is built to make a child think hard and retrieve, or built to keep them clicking.
Key takeaways
- Learning styles stay debunked: matching teaching to a child’s style has failed testing since 2008.
- Real but narrow evidence: Medly’s independent trial found modest GCSE science gains over four weeks.
- Judge the mechanics: retrieval, spaced practice and honest feedback teach, not neuroscience branding or streaks.
How to judge a tutoring tool your child actually uses
Here is the useful part, and it cuts in the company’s favor as much as against it. An AI tutor that asks a child questions, waits for an answer, and gives instant, specific feedback is closer to the mechanics that work than a points-and-badges app ever gets. Medly running a real controlled trial, against an active comparison rather than against nothing, is more rigor than most consumer ed-tech ever bothers with, and that is worth rewarding with your attention. The trap is not the product. The trap is letting the neuroscience halo and the learning-style promise do your evaluating for you, so you never ask the questions that separate a genuine tutor from a well-designed way to pass time.
So ask them. Does the tool make my child produce answers from memory, or mostly recognize and tap? Does it space practice and bring hard material back, or race to the next level? When my child is wrong, does it show the next step, or flash a red X and move on? And when a company puts a number in front of you, is it a before-and-after figure from its own users, or a result measured against a comparison group who did something else? A parent who asks those four questions is better equipped than any brochure adjective to tell whether a tool will teach.
None of this means an AI tutor is unable to help your child. It means the label on the box is not the evidence, and a founder’s credential is not a result. The tools that earn a place in your child’s week are the ones built around how learning actually works, effort, retrieval, spaced practice and honest feedback, whoever is selling them and whatever they call it. Judge the mechanics, not the story told about them, and you will rarely be fooled by the next buzzword either.
You do not need a neuroscience degree, or a founder’s, to make a good call about the tools your child uses. You need to trust two things: your child is capable of real learning, and real learning is built by effort and honest feedback, not conjured by branding. The villain in this story is not a company and it is not the technology, it is an old and profitable habit, engagement mistaken for learning, where minutes in an app and a science-sounding story stand in for a measured gain. That is the confusion Learning Success was built to cut through. Our All-Access membership opens an assessment that asks about the processing systems your child’s learning runs on, and a roadmap that names what to build first and in what order, worked alongside the reading, spelling or math instruction your child is already getting.
See what All Access gives your child
See MTSS for schools
Is your child struggling in school?
Get your free personalized learning roadmap
You describe what you see at home. We turn it into a plan you start this week.
- Answer 5 short questionnaires about what you already notice, 30–45 minutes at your own kitchen table
- Your child sits no test and gets no score: nothing to schedule, nothing for them to dread
- You do the answering, the AI does the writing, and a person reviews it before it reaches you
- Access all 40+ courses instantly: reading, math, focus, processing and more, with new ones added regularly
Why we use AI, plainly: it writes from a knowledge base our team maintains and audits. We work through it line by line and pull anything the evidence stops supporting. The roadmap you get on Tuesday reflects what we corrected on Monday, and a human still reads it before you do.
Your school district must evaluate your child free of charge if you ask in writing, whatever your income and whatever the outcome (US, 34 CFR 300.111 and 300.301(b)). That route takes time and answers a different question than you do. This one starts today, from what you already know.
Your answers stay yours. We do not sell your personal information, and we do not hand identifiable assessment data to outside AI companies to train their models.
A screener is a starting point, not a diagnosis. If your child might need formal accommodations (an IEP or 504 plan), or you suspect a vision, hearing or medical cause, pursue a professional evaluation too. That is the only route to those supports.
References
- EU-Startups — London’s Medly raises €6.9 million to democratise access to personalised education through an AI tutor
- tech.eu — Medly AI raises $8M to bring AI-powered tutoring to more students
- TNW — Medly AI raises $8M seed led by Felix Capital for GCSE tutoring
- EdTech Innovation Hub — Medly AI trial finds significant GCSE science gains across all three subjects
- UCL Enterprise — UCL startup raises £1.7 million to bring an AI private tutor to more students
- BusinessCloud — School friends raise £6m to give every kid an AI tutor
