Edudojo LogoEdudojo

Our method

Learning is the path between drafts.

A finished essay shows where a student ended up. Edudojo keeps the route: the drafts, and the chat with a coach that asks instead of writing.

Six stops: a draft, a question, a revision, a check, the replay and the teacher's decision.

The route, stop by stop

Student’s lane

The student works with the coach.

  1. 01, Draft 1

    The student writes first.

    In their own words, before any help. The draft is saved.

    Critical Thinking 101 · Essay

    Draft 1 · first attempt

    …which proves that phones are the main cause of falling exam results across the country.

  2. 02, Coach asks

    The coach asks. The student writes.

    It does not rewrite the sentence. It asks what the evidence measures.

    Why we keep students in the thinking

    The coach marks the words “proves that phones are the main cause”.

    Coach asks

    What does the study actually measure: teachers’ opinions, students’ attention, or exam results?

  3. 03, Student revises

    Every draft is kept.

    The student changes the claim. Both drafts and the chat stay in the record.

    See the Replay

    The student strikes the claim and writes: …which suggests phones may affect attention, but does not show that they cause lower exam results.

  4. 04, Red Team check

    Checking AI is a skill to practise.

    Now and then the coach plants a small weak step on purpose, without saying which. Try one.

    How the Red Team works

    Try it

    One step in this coach reply is weak on purpose.

    Tap the step you doubt.

    Does the study show that phones cause falling exam results?

  5. Submitted

    Teacher’s lane

    After submission, the teacher reads the record.

    05, Teacher reads the replay

    Read the drafts, not only the final text.

    An essay takes seconds to generate, so the final text says little. The teacher reads how it changed.

    Write an AI-use policy

    The finished essay

    …which suggests phones may affect attention, but does not show that they cause lower exam results.

    What a teacher has today

    Nothing else.

    No drafts, no questions, no record of how the text was made.

    Who did the thinking?

    With Edudojo, the path to it

    1. Draft 1

      Removed: …which proves that phones are the main cause of falling exam results across the country.

    2. Coach asks

      What does the study actually measure: teachers’ opinions, students’ attention, or exam results?

    3. Draft 2 · the finished text above

    One finished essay paragraph, then two views under it: what a teacher has today (the final text alone), and the path that led to it (the first draft and the coach's question).

    Submitted work · Critical Thinking 101 · Read-only

    Draft 1: …which proves that phones are the main cause of falling exam results across the country.

    The coach asked: What does the study actually measure: teachers’ opinions, students’ attention, or exam results?

    Draft 2: …which suggests phones may affect attention, but does not show that they cause lower exam results.

    Illustration of the teacher's read-only replay of one essay paragraph.

  6. 06, Teacher decides

    The teacher decides.

    Edudojo puts the record in front of the assigned teacher and stops there.

    A practical assessment guide

    Submitted work · Critical Thinking 101 · Read-only

    • 2 drafts
    • 1 coach question
    • 1 How I worked declaration

    Teacher's judgment

    Edudojo leaves this empty. The teacher reads the record and decides.

    Illustration: the record of one essay in front of the teacher, with the judgment left to the teacher.

Research on AI and learning, not Edudojo results

Why the coach asks instead of answering.

Study 1

Standard GPT-4 helped in practice, then hurt on the exam.

Grades indexed to 100 for students who never used AI. While practising with AI: standard GPT-4 148, an increase of 48%; the tutor with guardrails 227, an increase of 127%. On the exam without AI: standard GPT-4 83, 17% lower; the tutor with guardrails about the same as no AI, as the harm was largely removed, as reported.

Bastani et al., “Generative AI without guardrails can harm learning: Evidence from high school mathematics”, PNAS 122(26), 2025 · Field experiment, nearly 1,000 high-school maths students

Read the paper by Bastani et al. (opens in a new tab)

Study 2

Heavier AI use went with weaker critical thinking.

Reported association (correlational)

  1. Frequent AI tool use
  2. Cognitive offloading
  3. Lower critical-thinking scores

Frequent AI tool use correlated with lower critical-thinking scores, mediated by cognitive offloading.

Correlation, not cause

Gerlich, “AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking”, Societies 15(1), 2025 · 666 people

Read the paper by Gerlich (opens in a new tab)

Practice gains from given answers did not last. A tutor that made students think largely avoided that loss.

What we do not claim.

Not a measure
Recorded activity does not measure thinking, integrity or mastery.
No grades
Edudojo does not grade. The teacher judges the work.
No proven effect
Process over product is a teaching approach, not a demonstrated effect of Edudojo.
Beta
In beta, with demo accounts. The Jaipur pilot checks usability, reliability and running cost, not learning.

Next stop: your class

Start with one college, in English and Hindi.

Our first planned pilot is with Parishkar College in Jaipur.

Write to us
See the pilot plan