Ed.ai, a startup based in Lyon, France, has secured 5 million euros in funding from investors including Bpifrance, La Poste Ventures, and 50 Partners. The company is developing an AI platform designed to assist teachers in grading and providing feedback on student work. The platform analyzes handwritten assignments, identifies errors, and suggests corrective exercises that teachers can adjust before sharing with students. According to Ed.ai, the system has significantly increased the number of assignments it processes, from 100 per week to 2,500 per day. The company plans to demonstrate the effectiveness of its system in France and the United States before expanding further. The platform processes student work as photographs or digitized documents and provides a proposed grade, detailed comments, and exercises aimed at addressing detected errors. Teachers can adjust the grading scale, review the AI's analysis, and decide what feedback to give to students. Ed.ai emphasizes that its system is not an automatic correction tool but an assisted pre-correction system that retains teacher responsibility for final evaluation. The AI analyzes student responses, extracts criteria and skills, and generates a proposed grading scale. Teachers can modify instructions, mastery levels, and points awarded before submitting assignments. The AI recognizes mathematical handwriting with a 95% recognition rate and can track reasoning steps, assign points to partially correct answers, and recognize equations, graphs, or geometric constructions. However, the AI must distinguish between different types of errors, such as sign errors, inappropriate methods, or transcription errors. Ed.ai aggregates results to identify mastered concepts, recurring errors, and reasoning steps that pose problems. At the class level, it can group students facing similar difficulties and highlight areas where teachers need to focus. The platform aims to provide a detailed map of learning progress rather than just a grade. This level of detail helps teachers make informed pedagogical decisions. The company generates personalized remediation activities based on identified errors, which teachers can modify, print, or distribute digitally. This approach addresses the challenge of differentiated teaching by reducing the preparation time required for individualized exercises. Ed.ai's team includes former teachers and individuals with experience in software infrastructure and classroom experience. The startup's platform processes handwritten copies, allowing students to continue working on paper while digitizing after evaluation. This approach reduces adoption friction and addresses the unique challenges of handwritten reasoning, hesitations, and errors that digital platforms may not capture well. Ed.ai ensures that student data is not used to train its models or those of its suppliers. Ed.ai has expanded its offerings to cover ten subjects from middle school to high school in France and has deployed its solution in 500 schools, up from 120 in a year. The company charges institutional clients starting at 6 euros per student, per year, and per subject, with decreasing prices for multiple disciplines. It plans to use part of the funding to conduct studies on student progress and teacher well-being. Ed.ai is accessible through various digital work environments and integrates with platforms such as ÉcoleDirecte and Pronote. The company has also begun its American expansion, adapting its content to the frameworks of fifty states and integrating with platforms like Canvas, Google Classroom, Clever, and Schoology. It plans to open new subjects and launch pilots in higher education, including experiments with the University of Maryland and Francis Marion University. The grading market is already occupied by companies such as Gradescope, CoGrader, Graide, and Pensieve, each with their own focus areas and user bases. Ed.ai's unique positioning combines several rarely combined choices, including starting with paper, covering scientific and literary subjects, keeping the teacher in the validation process, and extending grading with the generation of differentiated activities. The company aims to process two million copies during the 2026-2027 school year and one million per month in the following two years.