Tiny Lab

3-Type System — Final Recommendation

Test Results: 90% accuracy (18/20 scenarios) | 4.25/5 clarity | 100% said "just right"

SPR

SPR · 01

Customer Support Triage Agent

We built an AI agent to automate customer support ticket triage, shipped it live, and measured 40% reduction in response time.

EXP

EXP · 02

Local LLMs vs API Calls

We spent 3 months testing whether local LLMs could replace API calls. Tried 6 different models and documented what we learned.

NOTE

NOTE · 04

Token Limits Break Everything

Quick observation: the model consistently fails when given inputs with special characters. Documented for future reference.

SPR · 01 — Sprint

Customer Support Triage Agent

Status: Shipped ✅ Duration: 90 minutes Tools: OpenAI API, Peekaboo

Problem

Customer support tickets were piling up, and the team was spending hours manually triaging them. We needed a way to automatically categorize and prioritize incoming tickets.

What We Built

An AI agent that reads incoming support tickets, categorizes them by urgency and topic, routes them to the right team member, and suggests automated responses for common issues.

Key Takeaway

The agent reduced response time by 40% and freed up 10 hours per week for the support team. The biggest win was automatic routing—no more manual triage decisions.