Momentum AI

From goals to execution, intelligently.

Comment

Project Story

Inspiration

Every project starts with an exciting idea, but somewhere along the journey, the same question always appears:

"What should I work on next?"

Most productivity applications are excellent at storing tasks, creating lists, and tracking progress. However, they rarely help users make the most important decision: determining which task deserves their attention right now.

As someone who frequently balances certifications, personal projects, and software development, I found myself spending more time deciding what to work on than actually making progress. Long task lists and multiple deadlines created decision fatigue instead of clarity.

That inspired Momentum AI.

Rather than becoming another task manager, I wanted to build an intelligent project assistant capable of understanding a project objective, breaking it into meaningful work, and continuously recommending the next highest-impact task. The goal wasn't simply to organize work—it was to help users build momentum.

What it does

Momentum AI is an AI-powered project prioritization assistant built entirely on AWS serverless services.

At its core is M1 (Momentum Intelligence), an AI agent powered by Amazon Bedrock Nova Lite.

Instead of asking users to manually create dozens of tasks, Momentum begins with a project objective.

Users simply provide:

  • Project name
  • Project objective
  • Project deadline

M1 then generates a complete execution plan consisting of actionable tasks, suggested deadlines, logical execution order, and explanations describing why each task contributes toward the project's success.

Once tasks are accepted, M1 performs intelligent prioritization using the Eisenhower Matrix, while also considering:

  • Project objectives
  • Project deadlines
  • Task deadlines
  • Task descriptions
  • Inferred task dependencies
  • Previously completed work

Every recommendation includes:

  • Priority score
  • Eisenhower quadrant
  • AI reasoning
  • Recommended next action
  • Confidence score

To help users stay focused, Momentum presents all prioritized work inside a clean dashboard featuring an AI-generated Eisenhower Matrix, project statistics, project colour coding, and a scrolling deadline ticker highlighting work due within the next three days.

The result is an application that doesn't just organize work—it helps users decide what to do next.

How we built it

Momentum AI was built as a fully serverless application on AWS during a single weekend.

The frontend was developed using React, TypeScript, Tailwind CSS, and Vite, then deployed using AWS Amplify Hosting.

The backend uses Python FastAPI running on AWS Lambda, exposed through Amazon API Gateway.

Projects and tasks are stored in Amazon DynamoDB, while Amazon CloudWatch provides monitoring and logging.

The intelligence of the application comes entirely from Amazon Bedrock Nova Lite.

Rather than implementing complicated prioritization algorithms in Python, I designed the backend to act purely as an orchestrator.

It retrieves project information, constructs structured prompts, invokes Bedrock, validates the returned JSON using Pydantic, and delivers explainable recommendations back to the frontend.

This approach keeps the business logic inside the AI, making it easier to evolve the reasoning process without changing application code.

Challenges we ran into

The biggest technical challenge was making large language model responses predictable enough for production use.

A web application requires structured, reliable data—not conversational text. I solved this by carefully designing prompts that instruct Amazon Bedrock to return validated JSON containing priority scores, Eisenhower classifications, reasoning, recommended actions, dependencies, and confidence scores.

Another challenge was controlling project scope.

There were many features I wanted to include, including authentication, collaboration, calendar integrations, notifications, and scheduling. However, with only a weekend available, maintaining a disciplined MVP was essential. Focusing on one core capability—AI-powered decision support—allowed the application to feel complete rather than overloaded with unfinished features.

Accomplishments that we're proud of

I'm particularly proud that Momentum goes beyond being another task management application.

Instead of requiring users to manually organize everything, M1 actively helps them plan their projects and determine what matters most.

Other accomplishments include:

  • Building a complete serverless application using AWS services.
  • Successfully integrating Amazon Bedrock as the primary reasoning engine.
  • Designing explainable AI recommendations rather than opaque priority scores.
  • Implementing AI-powered task generation from simple project objectives.
  • Building a clean dashboard centered around the Eisenhower Matrix and real-time project visibility.
  • Delivering a polished MVP within a single weekend.

What we learned

This project reinforced that generative AI is far more valuable when used as a reasoning engine than simply as a chatbot.

I learned how to design prompts that consistently return production-ready structured JSON, making AI outputs reliable enough for frontend applications.

I also gained practical experience using Amazon Bedrock to solve decision-support problems, where every recommendation must be justified instead of simply generated.

Finally, building Momentum entirely on AWS serverless services demonstrated how quickly ideas can move from concept to deployment without managing infrastructure, allowing more time to focus on product design and user experience.

What's next for Momentum AI

The current MVP proves that AI can meaningfully reduce decision fatigue during project planning, but this is only the beginning.

The next evolution of Momentum is transforming M1 into a truly autonomous project execution agent.

Future capabilities include:

  • Automatic reprioritization whenever project conditions change.
  • Multi-user workspaces and authentication.
  • GitHub, Jira, Slack, and Google Calendar integrations.
  • Daily AI project briefings and proactive deadline alerts.
  • Voice-driven project planning using Amazon Bedrock.
  • Long-running agent sessions capable of continuously monitoring project health and recommending actions without waiting for user prompts.

The long-term vision is simple:

Momentum AI won't just help users manage projects—it will become an intelligent teammate that continuously reasons, plans, and guides projects toward successful completion.

Built With

Share this project:

Updates

Submission history