How does Ed.ai’s AI work?

Ed.ai’s AI relies on a chain of AI building blocks. Here are the main stages:

• Reading and transcribing the handwritten paper

  • An OCR developed in-house reads the student’s handwriting.

  • It creates a digital twin of the paper (a faithful text version) with no personal data: the paper is de-identified before any processing by external AI models.

• Initial grading

  • The AI analyzes each answer against the grading scale (and the answer key) you provide.

  • It draws on skill frameworks (built from classroom practice, from state standards such as Common Core and TEKS, and from the work of our learning engineers).

  • A workflow system calls on several AI models, chosen according to the need (sign charts, numbers, written explanations…).

• Reteaching

  Based on the difficulties found in the papers, Ed.ai prepares reteaching activities for the students you choose.

  👉 Learn more: Comment fonctionne la personnalisation des exercices en fonction des erreurs des élèves ?

• Ed, the AI assistant

  With “Talk to Ed!”, an assistant looks at your assessment results to write summaries, suggest activities and answer your teaching questions.

  👉 Learn more: Qu’est-ce que « Parler à Ed » ?

At every stage, the AI proposes and you decide: nothing reaches your students until you’ve reviewed it. And your students’ work is never used to train AI models.