One dashboard for SLA firefighting and distributor quoting

The problem

Two disconnected jobs ran the business day to day. SLA risk lived buried in the help desk's own reporting, so nobody had one ranked view of what was about to breach, which clients were spiking, or whether the day's queue would even fit in the hours available. Resolved tickets with zero hours logged for hourly-billed clients went unnoticed, quietly leaking billable revenue every month. On the procurement side, the distributor had no browsable catalog or product search API at all, only a daily price-and-availability feed, so every quote meant manually hunting for the cheapest entitled price type, checking margin by hand, and hoping nobody had priced it under the manufacturer's minimum.

What we built

We built one platform that runs both jobs. The help-desk half pulls tickets into a live board ranked worst-first by SLA risk, with imminent-breach alerts, per-client volume spikes, and a same-day capacity forecast, backed by an efficient polling endpoint and a wallboard mode for office displays. It also drives ticket claiming, a 0-100 agent performance score across resolution rate, first-response speed and SLA adherence, a monthly review deck exported straight to PowerPoint, an Unbilled screen that surfaces resolved-but-unlogged work for hourly clients, and on-demand CSV exports. The procurement half builds its own catalog from the distributor's daily feed and purchase history, so staff can search parts, sweep every price type for the cheapest entitled rate, solve a sell price for a target margin, flag anything under MAP, and print a quote in the company's own house style. Quotes carry a draft/sent/accepted/declined lifecycle with chase reminders and can link back to the ticket that triggered them, and a deal radar diffs catalog scans to surface price drops and cost increases automatically. An optional AI layer writes narrative commentary on top of the numbers, but every calculation is deterministic plain JavaScript; the model only comments on results that have already been computed.

The outcome

A ticket stops looking neglected the moment a quote goes out against it, so the two halves of the job now reinforce each other instead of living in separate tools. What's about to breach SLA moved from a question someone had to go dig for to something visible on a wallboard, and the Unbilled screen turned silently unbilled hours into something someone actually checks. On procurement, margin protection and MAP compliance stopped depending on someone remembering to check by hand, and the deal radar catches distributor price moves the team would otherwise only notice after the fact. The same pattern, bill for time under SLA, resell hardware at a protected margin, fits any business shaped that way: VARs and hardware resellers, AV/telecom/security integrators, copier dealers and other field-service outfits billing hourly, and internal IT departments buying their own hardware.