AI-powered grading tools are reshaping how teachers spend their time in classrooms, with winning solutions from Estonia's President's Education Hackathon cutting annual grading hours from 378 to under 100 for a single educator. Rather than automating teachers out of the equation, these tools are designed to preserve the human judgment that matters most while eliminating repetitive manual work that burns out educators.

What Problem Are These AI Tools Actually Solving?

The starting point for one winning team was brutally simple: a high school math teacher spent 378 hours per year grading student work. That's roughly equivalent to nine full work weeks devoted entirely to marking papers and tests. The team behind Punane Pastakas, or "Red Pen," set out to bring that figure below 100 hours annually.

The hackathon, initiated by President Alar Karis and organized by the Office of the President with support from organizations including OpenAI, brought together 135 teachers, students, developers, and education experts across 30 teams. Three grand prize winners each received €6,000 in funding, with the total prize pool reaching €20,000.

Teacher burnout and workload pressures are not unique to Estonia. Across education systems globally, grading consumes time that could be spent on lesson planning, student interaction, or professional development. The hackathon positioned AI solutions as a direct response to staffing shortages and retention challenges in schools.

"I thank everyone who contributed to reducing teacher burnout, making students' lessons more interesting, and making schools' educational decisions more data-based, thus improving Estonian education. New solutions will certainly also help bring new teachers to school," said President Alar Karis.

President Alar Karis, President of Estonia

How Do These AI Tools Keep Teachers in Control?

The design philosophy behind the winning tools reveals a critical insight: the most effective AI in education doesn't replace teacher judgment; it removes friction from workflows teachers already use. Punane Pastakas was built around a specific constraint: teachers should not have to learn new software or change their existing habits.

The tool works with tablets and styluses that teachers already adopted during the pandemic. Rather than forcing educators into a new interface, the AI handles the grading while teachers review and approve the results. The system uses what's called a "master agent" that coordinates smaller AI agents in parallel, including one that searches math textbook content and another that identifies errors and checks whether student work remains logically consistent.

"The trick was not making teachers learn anything new. They already love their tablets, most got one during covid. So we kept the stylus, kept the tablet, and just took over the boring part: the AI grades and the teacher reviews," explained Elias Teikari, AI Engineer at Rapidata.ai and Computer Science student at the University of Tartu.

Elias Teikari, AI Engineer at Rapidata.ai

This human-in-the-loop approach addresses a fundamental concern about AI in education: the risk that automated systems might miss nuance or context that a trained educator would catch. By keeping teachers as reviewers rather than replacing them, the tool preserves professional judgment while eliminating hours of repetitive work.

What Makes These Solutions Different From Generic AI Chatbots?

The three grand prize winners focused on distinct problems within school operations, each designed to address specific pain points rather than offering generic AI assistance:

• Punane Pastakas (Red Pen): Grades handwritten math work and provides detailed feedback while tracking learning gaps across a class knowledge map, allowing teachers to identify which students need additional support.
• AITA (Curriculum-Based Intelligent Lesson Assistant): Helps teachers create and adapt Estonian-as-a-foreign-language teaching materials while ensuring alignment with Estonia's curriculum and local educational frameworks.
• Integrated Workstations: Enables teachers to design cross-curricular assignments and learning stations aligned with national curriculum standards, allowing students to complete tasks at their own level with step-by-step guidance.

Two additional award winners addressed school leadership challenges. DIQU won a €2,000 prize for an analytics tool that consolidates school data from different systems into a single view for administrators. MATx received a special award for its platform that identifies student learning gaps and recommends targeted exercises.

The common thread across all winners: they solve operational problems that consume time without adding educational value, freeing educators to focus on what only humans can do.

How Are Schools Planning to Adopt These Tools?

The hackathon is not the endpoint; it's a launchpad. The three grand prize winners will present their ideas at Latitude59, an international business ideas festival, where they will seek early investment and pathways into schools. The team behind Punane Pastakas has already set a timeline: they aim to bring the tool to Estonian math and science teachers by fall 2025.

"It is now important that these solutions do not remain in a drawer, but find their way into schools. If we want Estonian education not to lag behind in technological development, we must be prepared to invest more in educational technology," stated Markus Villig, CEO and Founder of Bolt, who chaired the hackathon jury.

Markus Villig, CEO and Founder of Bolt

Beyond Estonia, similar models are emerging globally. In the UK, Medly AI won the inaugural ETIH Innovation Awards 2026 for Best AI Tutor or Personalized Learning Agent, recognized for its Socratic tutoring approach that guides students toward understanding rather than handing over answers. The platform now sees between 100,000 and 200,000 tutoring interactions daily, suggesting significant adoption among students seeking exam preparation support.

What Are the Broader Implications for Teacher Roles?

The hackathon framed AI not as a threat to teaching but as a tool for transformation. President Karis connected the winning ideas directly to the challenge of attracting and retaining educators in a competitive labor market. By reducing the administrative burden of grading, schools can help teachers focus on the aspects of their work that require human expertise: mentoring, feedback, and understanding individual student needs.

One team member, Andrius Matšenas, a Product Designer and former high school math teacher, noted that grading took twice as long as actual classroom instruction, with the most difficult 20 percent of work consuming the majority of time. His experience shaped the problem the team chose to solve.

"We applied the magical handwriting-detection capabilities of frontier language models to understand students answers and give teachers insights to conceptual understanding of their students based on level of education, not only right-wrong final answers," noted Andrius Matšenas, Product Designer on the winning team.

Andrius Matšenas, Product Designer

This distinction matters. The tool doesn't just mark right or wrong; it analyzes student reasoning and provides teachers with insights into conceptual understanding. That kind of analysis requires both AI capability and human interpretation.

The hackathon also signals a shift in how governments and institutions are approaching AI in education. Rather than top-down mandates or one-size-fits-all solutions, Estonia's model brought together teachers, developers, and education experts to identify real problems and co-create solutions. The involvement of teachers in the design process ensured that winning tools addressed genuine pain points rather than theoretical possibilities.

As AI tools continue to enter classrooms globally, the Estonian approach offers a template: focus on augmentation over replacement, preserve human judgment in decisions that matter, and involve educators in the design process from the start. The goal is not to make teaching faster or cheaper, but to make it more sustainable and focused on what teachers do best.