Sarah Operations Lead

• 

  Sarah Thumbnail

The Story of Sarah Operations Lead

Inspiration

In the fast-paced world of e-commerce, operational pipeline and supply chain disruptions happen in real time. Operations leads routinely lose valuable hours manually downloading, opening, and parsing complex multi-tab CSV data matrices just to find out why a tracking node flatlined. We wanted to build a native, zero-friction solution that lives exactly where the team communicates, converting cold, raw data layers into intelligent, human-readable forensic audits instantly.

How We Built It

Sarah Operations Lead is engineered on a TypeScript (server.ts) backend architecture running an event-driven listener loop for Slack's App Mentions, securely deployed via Render. Decoupled Architecture: We implemented an architecture inspired by the Model Context Protocol (MCP) framework, completely separating the core engine logic from data tool ingestion. This lets Sarah fetch, stream, and evaluate secure database files dynamically instead of storing massive data structures in the LLM runtime. Native Interface: We entirely rejected raw Markdown text walls in Slack. Sarah communicates exclusively using Slack Block Kit elements, utilizing structured Section layouts, Context blocks for operational flags, and visually prioritized data hierarchies.

Challenges Faced

The biggest hurdle was designing a robust parsing pattern that could reliably catch micro-anomalies across hundreds of rows of inconsistent e-commerce data stamps (such as our active Ruokamill supply chain data model) without triggering system timeouts. We overcame this by building an isolated data engineering step that strips out file noise before presenting the clean extraction payload to the agent layer.

What We Learned

Building Sarah proved that Slack isn't just a messaging platform; it is the ultimate interactive canvas for real-time operations. We learned how to deeply customize Block Kit payloads to make complex financial and logistical data immediately scannable for enterprise teams under high-pressure scenarios.

Built With

• api
• gemini
• mcp
• node.js
• render
• typescript