PM Brain

Inspiration

While AI has made a lot of progress, Project Management still is quite manual, which makes it inefficient, troublesome and sometime tiring. AI exists in Project Management, but it does only bits and pieces. For example, it can help write stories, provide insights on a project, write briefs etc. But It still cannot manage the project and cannot track important decisions - decisions that are buried under hundereds of zoom calls, thousands of mails, and uncountable slack messages.

What it does

PMBrain - listens, reads and tracks all the artifacts for a project and identifies important decisions, requirements and drift in the project. It is like a small LLM, specific to a project that knows what is going on in the project and what could impact the progress in the project. PMBrain leverages the advanced multimodal capabilities of Gemini to bridge the critical "Alignment Gap" between high-level project strategy and daily execution. By using Gemini as an automated, persistent auditor, the application transforms unstructured project artifacts into a structured, immutable truth.

The integration is central to three core workflows:

Multimodal Workspace Onboarding: Gemini analyze project briefs across diverse formats—including text, document images (OCR), and native audio recordings. It distills core objectives, functional requirements, and strategic context, identifying "missing information" that would otherwise lead to project failure.

Autonomous Decision Discovery: By reasoning over requirements and domain context (e.g., Fintech vs. E-commerce), Gemini proactively suggests critical decisions (Technical, Budget, Scope) that stakeholders must track to maintain alignment. Real-Time Drift Detection: This is the application’s "Killer Feature." As evidence is ingested—ranging from meeting transcripts to Slack exports—Gemini semantically compares new "observed" commitments against the "Golden Source" project baseline. It detects contradictions (e.g., a developer promising a feature that violates a "Privacy First" requirement) and flags these as Drifts, providing severity scores and suggested fixes. By utilizing Gemini’s native audio processing and reasoning capabilities, PMBrain ensures that project strategy and ground-level execution remain perfectly synchronized.

How we built it

I used Antigravity to build this prototype, and leveraging Gemini for backend analysis.

Challenges we ran into

Main challenge I ran into was that Antigravity couldn't get some of the functionality right after multiple attempts. If I were building manually, I would spend quite sometime on designing the functionality, but in some cases it appeared this was missing. While most of the UI and functionality was handled really well by Antigravity, but occasionally, it needed hand-holding.

Accomplishments that we're proud of

I am proud that I was able to create an MVP that works and achieves some of the functionality that I was looking for.

What we learned

I learnt how to use Antigravity and Gemini for fast development, but also learnt that they have limitations.

What's next for PM Brain

I would like to evolve this as a SaaS that can connect to Zoom, Slack, Jira and other project management related apps. The app in its current state requires overhaul in terms of ensuring that all the LLM prompts are accompanied with correct context.

Built With

• antigravity
• gemini
• llm
• typescript