TaskFlow AI

Students juggle classes, assignments, activities, and commitments, but current schedulers treat tasks equally. What if AI understood academic priorities and orchestrated time intelligently?

Comment

TaskFlow AI

Inspiration

Students juggle classes, assignments, activities, and personal commitments—but most schedulers treat all tasks the same. We asked: What if an AI truly understood academic workflows, priorities, and dependencies—and intelligently orchestrated time?

What it does

TaskFlow AI is an intelligent scheduling and task manager that acts like a personal executive assistant for students. It goes beyond calendars by offering:

  • Smart Scheduling — Auto-plans study sessions around deadlines, exams, and your energy patterns.
  • Intelligent Task Breakdown — Splits complex projects into realistic subtasks with time estimates.
  • Context-Aware Prioritization — Weighs course load, prerequisites, grading weights, and current standing.
  • Adaptive Learning — Improves recommendations from your habits, completions, and feedback.
  • Natural Language Interface — “I have a CS midterm next week and a paper due Friday” → your week is orchestrated.
  • Proactive Reminders — Sends timely nudges based on importance and current availability/location.

How we built it

  • Frontend: Next.js + TypeScript (app router), accessible UI with fast interactions.
  • Backend & Data: Convex for serverless functions, real-time data, and storage of tasks/schedules/user patterns.
  • AI Integration: OpenAI (GPT-4) for NL understanding, parsing deadlines/priorities/effort and proposing schedules.
  • Calendar: Google Calendar API for two-way sync.
  • Auth: Clerk (OAuth 2.0 under the hood) for secure user management.
  • Learning Loop: Lightweight feedback signals (e.g., “finished early/late,” “too dense”) tune future scheduling.

The AI engine processes natural-language inputs, extracts key entities (deadlines, priorities, estimated effort), and adapts scheduling decisions based on user feedback and completion patterns.

Challenges we ran into

  • Context Understanding — Distinguishing due dates vs. ideal completion dates by assignment type.
  • Time Estimation — Predicting realistic durations across students and task categories.
  • Priority Conflicts — Resolving overlaps among multiple high-stakes tasks.
  • User Trust — Explaining why a plan changed to build confidence and control.

Accomplishments we’re proud of

  • Built a functional prototype that parses complex schedules into optimized study plans.
  • Reached ~85% accuracy on time estimates for common academic tasks after tuning.
  • Designed an interface students can use within minutes.
  • Seamless calendar integration for quick adoption.
  • Privacy-first options with minimal data exposure and local-first design choices where possible.

What we learned

  • Feedback loops are essential for genuine personalization.
  • Balance automation with user overrides and transparent explanations.
  • Academic workflows are complex; explainability increases trust and adherence.

What’s next

  • Collaboration — Group projects with auto-meeting scheduling and task delegation.
  • Learning Analytics — Insights into study patterns and productivity trends.
  • LMS Integrations — Canvas, Blackboard, and others.
  • Predictive Warnings — Early alerts for upcoming conflicts.
  • Voice Assistant — Hands-free capture and replanning.
  • Study Techniques — Built-in Pomodoro, spaced repetition, and review blocks.

Built With

Share this project:

Updates