AI tutoring platform Medly has raised $8 million as it expands personalized exam preparation in the UK and United States — another sign that artificial intelligence could begin turning individualized academic support from an expensive advantage into a service available at mass scale.
WHAT’S HAPPENING
London-based Medly AI has raised an $8 million seed round led by Felix Capital, following rapid growth of its AI-powered tutoring platform.
Founded by former NHS doctors Paul Jung and Kavi Samra, Medly says more than 400,000 students have already used the service in the UK. The platform currently supports GCSE, A-level and International Baccalaureate preparation and has begun expanding into the United States with SAT preparation.
Instead of simply giving students answers, Medly is designed to provide explanations, practice questions and personalized feedback based on how an individual student is performing.
The company says its system uses multiple large language models and can work with handwritten answers, diagrams and step-by-step problem solving — areas where ordinary conversational chatbots can struggle.
Medly has also been selected for the UK government’s AI Tutoring Tools Pioneer Programme, receiving a £300,000 research and development contract to help test AI tutoring in real classrooms.
WHY IT MATTERS
Private tutoring has always had a scaling problem.
A good tutor can personalize explanations, identify weaknesses and spend additional time with one student.
But that requires another human being.
And that makes individualized tutoring expensive.
AI potentially changes that equation.
A software-based tutor can theoretically provide personalized support to thousands — or eventually millions — of students simultaneously.
That does not mean an AI tutor replaces a teacher.
It means students could increasingly have access to an additional layer of individualized help outside the classroom, regardless of whether their families can afford a traditional private tutor.
The UK government is already exploring exactly that possibility.
Its AI tutoring initiative is designed to determine whether safe, curriculum-aligned AI tools could eventually provide personalized support to as many as 450,000 disadvantaged students a year.
WHO BENEFITS
Students who cannot afford private tutoring may have the most to gain.
A student struggling with algebra at 9 p.m. does not necessarily need another classroom lecture.
They may need someone — or something — capable of recognizing exactly where they became confused and explaining the concept differently.
AI could make that type of individualized assistance available on demand.
Teachers could benefit too.
If AI systems can handle portions of practice, feedback and repetitive explanation, teachers may be able to spend more classroom time on instruction, motivation and students requiring direct human attention.
The UK government’s program specifically emphasizes that AI tutoring should support teachers rather than replace them.
WHO LOSES
Traditional tutoring businesses could eventually face significant competition if AI systems become good enough to deliver useful personalized instruction at dramatically lower prices.
But there is another risk.
Education cannot simply assume that personalized AI equals better learning.
Medly has reported that roughly three-quarters of surveyed students improved by at least one predicted grade, but those results come from the company and should not be treated as independent proof that AI tutoring caused the improvement.
That distinction matters.
AI tutoring tools still need rigorous testing for accuracy, teaching quality, student dependence, privacy and whether students are actually learning rather than simply becoming better at completing assignments with AI assistance.
WHAT HAPPENS NEXT
The race is now moving beyond general-purpose chatbots.
Education companies are beginning to build AI systems specifically around curriculum, assessment, student progress and teaching methods.
Governments are beginning to test them.
Investors are funding them.
And students are already using them.
The real test will be whether these systems can consistently produce measurable learning improvements without weakening the role of teachers or the student’s responsibility to think.
If they can, the economics of tutoring could change dramatically.
For generations, personalized academic support has largely depended on how much additional human attention a family could afford.
AI could begin separating those two things.
And if individualized tutoring becomes inexpensive enough to reach nearly every student, one of education’s oldest advantages may no longer belong primarily to the families who can pay for it.
