HackBros

Inspiration

Procrastination isn’t just a willpower problem. In our research and personal experience, people stall because goals feel too big and unclear, distractions break momentum, motivation naturally fades mid-way, and urgency only shows up right before a deadline.

We wanted to build a product that reduces the mental load of getting started and keeps users moving when they’re most likely to drop off.

What it does

Momentum-AI is an AI coaching platform that turns a vague intention into a concrete, time-lined execution plan and then actively supports follow-through.

Key capabilities:

• Interactive Goal Solidifier: the user gives a simple prompt, and the AI generates a realistic roadmap with phases, milestones, sub-tasks, and a timeline; the AI also asks clarifying questions (time available, constraints, skill baseline, deadline) to make the plan actionable
• Plan refinement loop: users chat with the AI to adjust scope, pacing, and task granularity until they confirm a final plan
• Motivation support: a reward system (badges/sound cues/progress feedback) and personalized encouragement messages based on user profile and progress
• Recovery when off-track: if a user misses a sub-task, the AI sends supportive messages and proposes a revised micro-plan rather than “punishing” reminders
• Deadline nudges: milestone-based push messages that restore urgency without feeling like nagging
• AI voice messages: the same encouragement, reward moments, and recovery prompts can be delivered via text-to-speech so users can stay engaged hands-free

How we built it

We designed Momentum-AI as a simple loop: clarify → plan → execute → intervene → adapt.

Implementation highlights:

• LLM-powered planning agent that converts a goal into structured tasks with durations, dependencies, and milestones
• A question-asking module that collects missing constraints before finalizing the plan
• A lightweight user profile store (preferences, tone, schedule constraints, progress signals) to personalize messages
• A scheduler that triggers nudges, milestone check-ins, and missed-task recovery flows
• A rewards engine that fires on completion events and generates short celebratory moments
• Text-to-speech integration to render coaching messages as voice for reward and proactive outreach
• A minimal web/mobile-friendly UI to create goals, view the timeline, and mark sub-tasks complete

Challenges we ran into

Generating plans that are both ambitious and realistic was harder than just producing a long checklist. We had to prevent over-scheduling and ensure tasks stayed concrete enough to start immediately.

We also had to balance helpful nudges with user fatigue, since too many notifications can become another form of distraction. Voice was surprisingly nuanced as well, because tone and timing matter for encouragement to feel supportive instead of robotic.

Accomplishments that we're proud of

We built an end-to-end experience that goes beyond planning and actually supports execution in the moments users typically fail. Our interactive goal clarification makes the plan feel tailored instead of templated.

We also plan to shipped voice-enabled encouragement and reward delivery, which makes the product more human and usable in real-world contexts like commuting or walking.

What we learned

A good procrastination solution is not a single feature; it’s a behavior loop. The highest leverage comes from reducing cognitive load at the start and providing compassionate recovery after a miss, not just sending more reminders.

We also learned that personalization isn’t optional: message tone, frequency, and the “next step” framing strongly influence whether users re-engage.

What's next for Momentum-AI

Next we’ll deepen execution support and measurement:

• Stronger distraction guardrails (focus sessions, quick context restore, “next best step” prompts)
• Integrations with calendars and task tools to reflect real constraints and reduce manual updates
• A/B testing for nudge timing, message tone, and voice styles to optimize adherence
• More adaptive planning (auto-rescope when time availability changes)
• Team/education pilots with aggregated, privacy-safe progress insights and template libraries

Built With

• amazon-web-services
• minimax
• next.js
• python
• supabase