Slack Productivity Agents

Inspiration

Modern workplaces rely heavily on Slack for communication, but employees still switch between multiple tools for HR information, documentation, and engineering support. We wanted to simplify this experience by bringing AI-powered assistance directly into Slack through specialized agents that can answer questions, perform actions, and retrieve company knowledge.

What We Built

Slack Productivity Agents is a multi-agent AI platform consisting of:

• Personal Agent – Helps employees with attendance management, Slack messaging, and general AI assistance.
• Knowledge Agent – Uses Hybrid Search and Retrieval-Augmented Generation (RAG) to answer questions from company documents and policies.
• Engineering Agent – Assists developers by monitoring engineering systems, analyzing MLflow and API logs, and helping navigate the codebase.

Together, these agents provide employees with a single AI-powered interface for productivity, knowledge discovery, and engineering support.

How We Built It

The platform is built using Python with Slack Bolt for Slack integration and FastAPI as the backend. Large Language Models power the conversational experience, while tool calling enables the Personal Agent to perform workplace actions such as attendance management and Slack messaging. The Knowledge Agent uses a vector database with Hybrid Search and RAG to retrieve relevant company documents before generating grounded responses.

Challenges We Faced

One of the biggest challenges was designing clear responsibilities for each agent. We wanted to avoid a single monolithic assistant that tried to do everything. Instead, we separated personal productivity, company knowledge, and engineering operations into specialized agents. Another challenge was integrating Slack interactions with external tools while ensuring the agents could naturally decide when to answer directly and when to invoke a tool.

What We Learned

This project deepened our understanding of multi-agent architectures, Retrieval-Augmented Generation, vector search, Slack integrations, and tool-based AI workflows. We also learned the importance of separating responsibilities between agents to improve scalability, maintainability, and user experience.

Built With

• ai
• langchain