Inspiration
Modern teams exchange hundreds of Slack messages every day, making it easy to miss important requests, approvals, deadlines, and action items. We wanted to build an AI-powered assistant that automatically detects actionable messages, converts them into trackable tasks, and ensures that important work never gets lost in conversations.
What it does
FlowGuard AI continuously monitors Slack conversations and uses Google's Gemini AI to analyze messages. Whenever an important message is detected, it automatically:
Generates an AI summary Classifies the task type Assigns a priority level Estimates confidence Creates an interactive Slack task card Lets users Assign, Complete, or Dismiss tasks directly from Slack Stores every task inside SQLite for tracking and history
This eliminates manual task creation and improves team productivity.
How we built it
The project was built using:
Python Slack Bolt SDK Slack Socket Mode Google Gemini 2.0 Flash API SQLite Database Slack Block Kit VS Code
The Slack Events API sends incoming messages to our Python backend. Gemini analyzes each message and extracts metadata such as summary, priority, confidence, and task type. The application stores this information in SQLite and instantly creates an interactive Block Kit task card inside Slack.
Challenges we ran into
Some major challenges included:
Designing reliable AI prompt engineering Handling Slack interactive button events Managing task ownership correctly Preventing duplicate task creation Integrating Gemini while preserving the original application workflow Maintaining real-time updates using Slack Socket Mode Accomplishments that we're proud of Successfully integrated Gemini AI into Slack Built automatic AI-powered task extraction Created interactive Slack task cards Implemented ownership, completion, and dismissal workflow Built a lightweight SQLite task management system Delivered a complete working prototype for real-world team collaboration What we learned
This project helped us understand:
Slack App Development Slack Events API Block Kit UI Socket Mode AI Prompt Engineering Gemini API Integration Python backend development Event-driven architecture What's next for FlowGuard AI
Future improvements include:
Multi-workspace support Team analytics dashboard Email and calendar integration Due-date reminders AI-generated task recommendations Workflow automation Advanced reporting and insights
Built With
- ai
- api
- block
- bolt
- code
- gemini
- github
- kit
- machine-learning
- mode
- natural-language-processing
- python
- slack
- socket
- sqlite
- vs
Log in or sign up for Devpost to join the conversation.