The Problem

Every school leader knows this problem because they see it every day. One teacher for thirty or more students means one pace for the whole room. However good the teacher is, they can't watch each student's understanding form in real time, so the first real sign that a student is struggling is usually the exam itself, weeks after it could have been fixed.

Students aren't short on material anymore. Between video lessons, practice apps and every new AI chatbot, they have more than they can get through. But none of it knows what a specific student actually understands, so a student can put in hours and still walk into the exam with the same gaps they started with.

Parents are often the last to know. Most only really find out how their child is doing at the parent-teacher meeting, by which point a term's worth of small gaps has quietly become a big one.

A market that can't supply enough good teachers

This isn't a perception problem. It's a supply problem. In Punjab alone, roughly 125,000 teaching positions sit unfilled against a workforce of 310,000, and more than 15,000 schools currently have no teacher assigned at all.¹ Nationally, close to a third of public-school teachers have had no professional training in modern teaching methods,² and the primary-level pupil-teacher ratio has been estimated at nearly 80 students for every teacher.³ When a school does find and keep a genuinely dedicated teacher, that teacher is still one person carrying a workload built for several, and every vacancy left unfilled elsewhere adds to it. It's little wonder good teachers report feeling stretched thin and demoralised, or that schools struggle to hold on to them at all.

Chasing the wrong signal

Faced with that shortage, students and parents look for a shortcut, and the shortcut they've found is personal brand. A teacher with a large following, a polished reputation, or a track record of past results becomes the default choice, especially for students in senior grades preparing for board or O/A-Level exams. What that choice usually buys, though, is a bigger class, not more attention: the same teacher spread across more students, often in a tuition-centre format the school has no visibility into. The pattern shows clearly in the data: one peer-reviewed study found that 62% of secondary students in Lahore, 54% in Faisalabad, and 34% in Peshawar were already paying for tuition outside school hours,⁴ echoing an earlier province-wide finding that private-tuition uptake was rising steadily even a decade ago.⁵ Families are, in effect, paying twice for an education the classroom alone isn't finishing. Star teachers also arrive with liabilities most schools aren't built to absorb: negotiating leverage, scheduling demands, and a loyalty that runs to their own brand rather than to the institution.

Administrators are flying blind between exams

Meanwhile, a school's own dedicated but uncelebrated teachers are often its best asset, and its most invisible one, because there's rarely a way to show, in real time, what they're actually achieving with their students. Administrators can't check academic performance until a test or exam happens; by then, a term's worth of small gaps has become a big one, and it's usually the administrator, not the specific teacher or topic, that parents blame. Schools can typically only demonstrate progress through exam results, at most a couple of times a year, with no standing feedback mechanism for parents in between, so nobody, teacher, administrator, or parent, can really see what a student has learned until it's too late to act on it.

How Sukrat AI Works

Most AI tools wait to be asked a question, which means a struggling student first has to know what they don't know. Sukrat works the other way. It figures out what a student needs to learn next, builds the lesson and practice for it, and adjusts as they go, the way a good tutor would, without waiting to be asked.

What Changes With Sukrat

All of this changes when Sukrat is used as a teaching aid across the whole school, not layered on beside it for the few teachers who have time to experiment.

In practice, teachers set the curriculum inside Sukrat, and from there it does the groundwork described above automatically. Every lecture gets scaffolding and reinforcement, every concept gets its own micro-assessment instead of waiting for a unit test, and every student's progress is measured against the actual standard the curriculum provider sets: Cambridge, IB, or Matric/FSc past papers, not a generic benchmark. A student who struggles gets targeted answers and extra coaching immediately, before the gap has a chance to compound.

Teachers carrying less bureaucratic overhead spend more of their time on the part of the job they signed up for: teaching, mentoring, and the judgement calls that still need a person. That, in turn, is what makes a school a place people want to work. Parents notice the same shift from the other side: a school that can show exactly how their child is doing, in granular, real-time detail, rather than asking them to wait and trust. Over time, that combination is what builds a school's reputation: recognised by parents who want to enrol their children there, and by teachers who want to work there because the job finally feels sustainable.

The Technology

Sukrat's intelligence has two separate jobs: deciding what a student needs to learn, and actually teaching it. Most AI tools have one, or the other. Sukrat is the first to have them both work together. An intelligence layer decides what a student should learn next, built on how students learn. A second layer of large language models generates the content itself: explanations, questions and feedback, grounded in that student's own coursework.

That split means a new course with its curriculum can be created in minutes. Sukrat already teaches Cambridge O and A Levels, International Baccalaureate, and Pakistan's Matric and FSc curriculum.

For Education Leaders

Sukrat takes the repetitive work off your teachers' plate: practice, quizzing, and diagnosis. In its place, it gives them live visibility into exactly which students need help and where, so their time goes into the interventions that actually move a student forward.

What Sukrat Delivers

Sukrat is being deployed with schools across Pakistan's Cambridge, IB and Matric/FSc systems, reaching more than 4,500 students from September 2026, with more schools joining. Each one adopted Sukrat because their own teachers and administrators saw what it did for their students.

In early pilot testing, before the product had much content loaded, all 22 students still returned for a second session in the same week, all 22 completed a full AI-generated lecture, and several asked for more content before the school could add it. Even with only a limited set of features released so far, the pilot cohort averaged over 30% weekly usage. A parent consultation with nine families led to requests to extend access beyond the pilot.

Built With a Real School, Not in a Lab

It's worth being direct about how Sukrat came to exist, because it's part of the pitch. The product wasn't specified on a whiteboard and shipped out to see if it would work. It was built and rebuilt over 3 months inside a live classroom, with a real teacher setting real lessons and real students working through them week after week, not a simulated cohort, and not a demo dataset. That is where the pilot numbers above come from. It's also why Sukrat's technology works in two separate layers instead of one: deciding what a student needs to learn, versus actually teaching it. That distinction wasn't a design decision made in advance; it became obvious from watching real students get stuck in real lessons.

Beyond diagnosis and practice, Sukrat gives students the full toolkit a subject needs. None of it replaces what a school already teaches. It makes sure none of it goes to waste.

This means Sukrat can finally provide personalised learning at scale, tailored to each student's learning style, something that was not possible before, due to human and technological limitations.