Field reps were losing leads from blind spots in availability and follow-up. I built the engine that orchestrated 10k+ clients and 250k+ interactions per month with competitive reassignment under the hood.
Problem
Coordinating leads across a field sales force in real time. Reps were losing leads from blind spots who needed callback, when, on which channel. Everything in each person's head; nothing shared, nothing timed.
Design
Competitive reassignment as the core mechanic: every lead carries a SLA timer; if a rep doesn't respond, the next eligible rep picks it up. Visibility everywhere who's working what, since when, on which channel.
Engineering
100% TDD + Clean Architecture from day one. Use-case classes read like requirements documents the squad of 5 devs onboarded fast and shipped without a QA bottleneck.
Operation
Backend sustaining 10k+ req/s at peak with zero downtime through high-traffic periods. 250k+ interactions per month flowing through the system. Pattern survived years after I left.
- Why a SLA timer + competitive pickup?Self-policing fairnessRound-robin punishes nobody; SLA + pickup punishes slowness. The timer makes responsiveness the currency fairness becomes self-policing without a manager intervening.
- Why Node.js + TypeScript?Node + TSSquad was already TypeScript-fluent. Mature ecosystem for the realtime workload (event loop fits coordination work). TypeScript caught domain bugs at compile time on a system handling money.
+20% measured revenue. Became Stone's #1 engagement tool and the team's internal quality benchmark.