Why remediation is the wrong starting point — what Pakistan's early years gap is actually costing, and what must change.
Who we are. Taleemabad is a Pakistan-based education technology organisation working in 1,101 government schools across ICT Islamabad, Rawalpindi, Punjab, Balochistan, and Sindh — reaching 152,815 students. We deliver AI-enabled teacher coaching, structured lesson plans, and Rumi, a foundational literacy and numeracy platform, at $2–4 per child per year.
What this paper argues. Pakistan's learning crisis is not a remediation problem. It is an early years investment problem. 78% of children cannot read by age 10 — not because remedial programmes have failed, but because the foundational window between Nursery and Grade 3 has been systematically neglected. Fixing this requires redesigning what early years schooling is for, investing in teacher capability at scale, and reframing foundational literacy as the entry point to critical thinking in an AI-shaped world.
What we are calling for. Curriculum reform in the early years. Sustained coaching over one-time training. FLN skills embedded in regular classroom teaching — not in parallel programmes that end when funding does.
Every year, as children reach Grades 3, 4, and 5, the same crisis becomes visible. Teachers realise large portions of their class cannot read. Assessments confirm it. The policy response is almost always: what remedial programme can we put in place?
Taleemabad believes this is the wrong starting point. Remediation is a response to a problem that was already preventable. It means asking the system to do twice the work — teach children what they should have learned at age six, while simultaneously teaching them what they need at age ten.
The answer is not that children cannot learn. It is that the early years of schooling — from Nursery and KG through the first years of primary — are systematically overburdened, under-resourced, and deprioritised. This is where the crisis begins. And this is where it must be addressed.
Pakistan has roughly 20–21 million children aged 3–5 who should be in early childhood education. Only 31% are.3 That means 13–14 million children enter primary school without any structured early learning — among the worst ratios in South Asia.
The consequences compound. In Punjab — Pakistan's most resourced province — 64% of children aged 7–14 lack foundational reading skills and 88% lack foundational numeracy.5 Children with multiple risk factors score 1.49 SD below their peers on school readiness. Pakistan also holds the world's second-highest number of out-of-school children: 25.1 million aged 5–16.6 The ECE gap feeds directly into this.
95% of schools have no dedicated ECE teacher.7 The National Curriculum 2022–23 mandates 2-year specialised ECE training — implementation has barely started. Most Katchi classes share space and teachers with primary grades. No province has a functioning quality monitoring system.
Pakistan spends 1.87% of GDP on all education — less than half of UNESCO's 4–6% minimum.8 ECE receives just 5.3% of that. Universal ECE by 2035 would require ~$1.85B/year. Current spend is approximately $300M/year.
There is a neurological reality that policy often treats as abstract: the window for phonological awareness, decoding fluency, and oral language acquisition is wide in the early years — and progressively narrows. By Grade 4, a child who cannot decode is not just behind in reading. She cannot access any other subject. Her mathematics problems are word problems she cannot parse. Her science textbook is a wall of text she moves her eyes across but does not read. The classroom teacher sees a child who is "slow" or "not trying." Neither diagnosis is correct — but by this point, neither is easy to fix.
The cost of that gap compounds beyond the classroom. A child who leaves primary school without foundational literacy is significantly more likely to drop out, less likely to enter the formal workforce, and more likely to remain in intergenerational poverty. Pakistan's 78% learning poverty rate is not a school problem. It is an economic problem, a labour market problem, and a human development problem — all of which trace back to what did not happen between ages three and eight.
Burst programmes can produce short-term gains in this context. But they stop working the moment funding ends — particularly when they sit outside the regular curriculum and have not changed what happens in the classroom every day. If the teacher has not changed, the gains evaporate. This is not a criticism of burst programmes in principle. It is an observation about what we have seen in practice, and it shapes why Taleemabad's model is built the way it is.
Governments and donors who fund remediation at scale are, in effect, paying twice for an outcome they could have funded once. Every year of delay increases the intervention cost and reduces the achievable ceiling. Storybooks and textbooks designed to address literacy gaps make a meaningful difference — but high-quality print material at national scale is expensive, and without a teacher who knows how to use them, they sit in a cupboard. The more durable investment is in the teacher.
The fiscal argument is sharper than it first appears. Research on learning poverty consistently finds that failing to reach foundational literacy reduces lifetime earnings by 10% or more per affected child.11 At 78% learning poverty and a working-age population of 70 million, the macro cost of the early years gap is not a development statistic. It is a structural drag on the economy. Early investment in foundational learning is not charity. It is the highest-return human capital expenditure available to a government that cannot afford to keep paying the remediation bill.
Pakistan's National Curriculum 2022–23 is, in many respects, a well-intentioned document. It mandates two years of specialised ECE training for early years teachers. It acknowledges foundational literacy and numeracy as priorities. It recognises play-based learning. The problem is not what it says. The problem is what happens — or rather, does not happen — between the policy document and the classroom.
Teachers receive training. They attend workshops, sometimes well-designed ones. They return to school on Monday. By Friday, the lesson looks the same as it did before they left. This is not a teacher failure. It is a system design failure. Behaviour change in teaching — as in any skilled profession — requires not a single event but a sustained feedback loop: observe, reflect, adjust, try again. A two-day workshop cannot create that. A quarterly school visit cannot sustain it. What is needed is daily, contextualised coaching — feedback that is specific to what happened in that classroom, with those children, this week.
This is precisely what AI makes possible at scale. Taleemabad's model combines structured teacher coaching, lesson-level content, and a diagnostic platform — making Teaching at the Right Level achievable through the existing teacher pipeline. The goal is not a child who passes a Grade 3 assessment. It is a child who, by adolescence, can think critically and independently in a world restructured by AI.
AI can do what human systems cannot afford to do at scale: deliver daily, personalised coaching to every teacher, flag which child is falling behind before the term ends, and adapt instructional content to where a learner actually is. At $2–4 per child per year, this is the most cost-effective lever for sustained behaviour change in teaching that currently exists.
For a child who has not yet built foundational literacy, AI becomes the fastest shortcut through work she doesn't understand. It accelerates the habit of covering content without comprehending it. If the foundation isn't there, AI doesn't close the gap. It makes the gap invisible for longer — and harder to fix when it surfaces.
All resources above are currently deployed in government schools. Content spans Urdu, English, Mathematics, Science, and Social Studies — aligned to the National Curriculum 2023.
In December 2025, Taleemabad deployed a 6-week structured FLN curriculum across Balochistan government schools — 191 lesson plans across three subjects, for Grades 3–5, with 76 teachers achieving an 87% activation rate.
Nearly every teacher ran the baseline screener before beginning instruction — meaning teaching started from where each child actually was. Every lesson included differentiated Beginner, Intermediate, and Advanced tracks. An independent RCT (Prof. Jishnu Das, Georgetown) is measuring outcomes — results expected May 2026.
Taleemabad does not work around government — it works inside it. Every school in Taleemabad's network is a government school. Every teacher is a government employee. The model is designed to become part of the system's own infrastructure, not a parallel programme that evaporates when grant funding ends.
Active government contract for 341 schools. Teacher certification is linked to Annual Confidential Reports (ACRs) — the government's own HR system for career advancement and promotions. This makes teaching quality improvement a career incentive, not just a programme requirement. Taleemabad is the leading EdTech in Pakistan operating at this level of government integration. We are advocating for the same ACR linkage in Rawalpindi and Punjab.
The Punjab provincial government is actively evaluating Rawalpindi results to decide on a province-wide expansion covering 6,000 schools. Taleemabad's evidence from 261 Rawalpindi schools is the direct input into that policy decision. AKU-IED certification validates the programme against national ECE standards.
Taleemabad advocated for early years investment in Islamabad Capital Territory and provided evidence to support the case. The government responded by establishing dedicated Early Year Centres in government schools and reducing the early years curriculum burden — creating protected space for foundational skill-building. This is exactly the systemic change this paper calls for, and it is already happening.
1,300 teachers are currently using Rumi on trial — a projected reach of ~65,000 students. The goal is systemic FLN delivery measured by ASER benchmarks — not a parallel project, but a permanent feature of how Sindh's government schools teach reading.
The case for early investment is not ideological — it is fiscal, neurological, and empirical. Pakistan cannot remediate its way to learning outcomes. Every year the system waits to fix the foundation, it compounds the cost and reduces the ceiling of what children can achieve. The teacher is the most durable delivery mechanism in any school system. Building her capability — early, consistently, at scale — is not a supplement to the strategy. It is the strategy.
Taleemabad has run burst programmes. We have seen them produce real gains in a matter of weeks. We have also watched those gains fade when the programme ended and the classroom reverted to its previous state. We have deployed training without sustained coaching and seen teachers unable to transfer new skills to practice. We have written lesson plans that were too complex and watched teachers default to the textbook. These experiences directly shaped the model we run today — and they are why these recommendations are specific rather than general.
The curriculum framework already calls for dedicated early years teachers and foundational skill priorities. The failure is in implementation. Provincial education departments need accountability mechanisms tied to these provisions, not just endorsement of the framework on paper.
Current teacher support programmes count training days and completion rates. These measure inputs, not impact. Loop Closure — was a gap identified, was it acted on, was it verified closed — measures whether teaching actually changed. Fund and evaluate coaching programmes on this basis.
Every burst programme Taleemabad has run has confirmed the same thing: gains are real but fragile without a teacher who can sustain them. In Rawalpindi, we are testing a different approach — FLN skills woven into every lesson plan teachers already use. The skills travel with the teaching, not with the grant.
In ICT Islamabad, teacher certification is tied to ACRs — the government's own promotion and career advancement system. This creates a career incentive that outlasts any project cycle. Punjab and Rawalpindi are next. Every province that adopts ACR-linked certification builds a sustainability mechanism no donor grant can replicate.
In an economy reshaped by AI, the question is not whether children can read a passage. It is whether they can reason through it, question it, and build on it. FLN is not the endpoint. It is the on-ramp. Setting national targets around this framing — rather than around grade-level reading pass rates — would reorient what the system is actually building toward.
We invite policymakers, provincial education departments, and development partners to consider three specific commitments:
1. Protect the early years. Mandate dedicated FLN time in Nursery through Grade 3. Remove curriculum content that competes with foundational skill-building before those skills exist.
2. Move from training to coaching. Replace one-off teacher workshops with continuous, school-embedded coaching — human or AI-assisted. Judge teacher support programmes by Loop Closure, not completion rates.
3. Link teacher accountability to learning outcomes. Expand ACR-linked certification — already operational in ICT Islamabad — to all provinces. Career incentives are the most durable sustainability mechanism education systems have.
Taleemabad is available to brief education departments, share programme data, or support the design of early years FLN frameworks in any province. We are building this infrastructure in Pakistan's government schools today — and we are ready to scale.