Schools: AI Supports Teachers with Marking and Lesson Planning

A vocational school with 80 teachers and 2,400 students faces the same problem as many schools: teachers spend a disproportionate share of their working hours on administrative and repetitive tasks — marking, formulating feedback, lesson planning, documentation.

Research shows teachers work an average of 47 hours per week, of which only 23 hours are spent in direct contact with students. The rest falls on preparation, marking, meetings, and administration.

Student Data Must Not Enter External AI

Student work, performance data, and personal development reports are specially protected under school law and GDPR. Use of external AI services — even for seemingly anonymised texts — was ruled out following a review by the data protection officer.

Local Teacher Assistant with SoverIQ

The school runs SoverIQ Stack on its own server, disconnected from the internet. Teachers can use it for:

• Marking support: student essays or coursework are automatically analysed for structure, argumentation, language accuracy, and task fulfilment — a structured feedback draft is produced for the teacher to adjust and sign
• Differentiated feedback: suggest different feedback formulations for different performance levels
• Lesson planning: generate a lesson design with phases, methods, and material suggestions from learning objective and class context
• Assessment tasks: create tasks for defined competencies and difficulty levels
• Report comments: suggest individual report statements based on performance data and pedagogical observations

Results

Time spent on marking fell by 38%. Feedback quality — measured through student and parent surveys — improved measurably: feedback was perceived as more concrete and more individual. Teachers report having more time for the actual pedagogical work that brought them into the profession.