Stop Planning.
Start Doing.

Student uploads a PDF/photo of the assignment brief or fetches from Google Classroom.

Data: Assignment file or Classroom link

AI: Parse and Extract

Nudge extracts: task type, deliverable format, word count, and deadline from the document.

Data: Assignment text

AI: Extract: parse structured metadata

AI estimates total hours based on assignment type, complexity, and length. Student can override.

Data: Assignment type, scope, deadline

AI: Predict: estimate realistic hours

ClaudBot asks 2-3 personalised questions to adapt the plan to the student context.

Data: Student context answers

AI: Generate: personalised questions

Nudge generates a Pomodoro plan (25-min work / 5-min break blocks) mapped across available days.

Data: Deadline, calendar availability, answers

AI: Generate and Recommend: build and name sessions

Student reviews the plan, drags blocks to adjust, then confirms. Plan is exported as .ics or pushed to Google Calendar.

Data: Confirmed plan

AI: Output: calendar export

AI Capabilities in Nudge

Four core AI actions power the full workflow

Read and interpret assignment briefs

Estimate realistic completion time

Build personalised Pomodoro plans

Name each session block with sub-goals

Human-in-the-Loop Gate

Before anything is pushed to Google Calendar, the student must review the generated plan and explicitly confirm. No calendar write happens without student approval. This is the hard checkpoint.

Human-on-the-Loop Monitor

After AI generates the time estimate, the student sees the reasoning and can manually override total hours before the plan is built. The student stays informed and in control without making every micro-decision.

Input: "2,000-word marketing essay on brand loyalty, due Friday" -- AI Estimated time: 10 hours -- Student confirmed override: 8 hours