erm.ai

AI Slack agent that watches chats and adds timely insights, auto-summarising progress into n8n for LinkedIn posts. Expandable with domain-tuned agents for law, tech, finance.

Comment

Inspiration

We wanted to make teamwork smoother by giving Slack groups an AI teammate - one that listens, supports, and turns group progress into something actionable and shareable. In addition to this, we wanted to be able to provide multi-agent support with fine-tuned models so that we can create a suite supporting a variety of purposes.

What it does

erm.ai is an AI Slack agent that monitors conversations, adds helpful insights, and automatically generates n8n-powered LinkedIn-ready updates. It can also be expanded with domain-specific agents.

How we built it

We integrated Slack’s MCP with our custom AI agent, added message filtering and context handling, and connected everything to an n8n workflow for automated LinkedIn content generation, which gets outputted to the Slack account. We also experimented with fine-tuning for specialised agents.

Challenges we ran into

Connecting the n8n workflow with the web application due to frequent collisions with the agent running in the background.

Managing noisy data, ensuring that it doesn't interfere with our valid data.

Accomplishments that we're proud of

Working well as a team despite not being familiar with people

What we learned

How to manage conversational context at scale, orchestrate multi-service integrations, design better agent behaviors, and build AI that supports teams without overwhelming them.

What's next for erm.ai

Expanding our library of domain-specialized agents, improving context detection, adding voice and meeting integrations, and creating a polished dashboard for workflow management.

Share this project:

Updates