GSNS, freight and logistics
A five-year-old freight platform, modernized and given an AI assistant
GSNS runs its whole freight business on a custom platform, from quotes to invoices. We took it over, brought it up to date, connected it to QuickBooks, and built an AI assistant into it.
14,000+orders handled in the platform
The problem
GSNS quotes, dispatches, tracks, and bills freight from one web app. Drivers use it on their phones. The office uses it all day.
The app worked, but it was running on aging versions of its framework, and every invoice had to be typed a second time into QuickBooks. When a payment landed in QuickBooks, someone marked it paid in the app by hand. Tax was worked out by hand too.
Orders arrived as emails and PDFs, and a dispatcher read each one and keyed it into the order form.
What we built
A full modernization, without a rewrite
We moved the app to Vue 3 and then through two major Laravel versions to Laravel 13, replaced the old build tooling with Vite, and cleared the security issues that came with it. Each step shipped to a beta site first, then to production through automated Docker deployments.
Security that drivers can live with
MFA is required for every account. Trusted devices remember a driver’s phone so they are not asked for a code on every shift.
QuickBooks, synced both ways
Invoices created in the app now appear in QuickBooks Online with their lines, HST, and attachments. When a customer pays in QuickBooks, the invoice is marked paid and locked in the app. Invoice numbers match in both systems, and a sync panel shows the status of every invoice.
Time off, on the calendar
Approved time off in QuickBooks Time flows onto the app’s events calendar on a schedule, so dispatch sees who is away without asking.
An AI assistant inside the app
A side panel answers questions about orders, quotes, and invoices. Drop in a customer email, an Outlook message, or a PDF, and it opens the order form already filled in, with every field it touched marked for review. It never saves on its own: it proposes, and a person clicks.
Taught from real paperwork
Before writing the order-drafting skill, we studied real orders next to the paperwork they came from, field by field, so the assistant reads documents the way the dispatchers do.
How it works day to day
- Each person has their own private assistant memory.
- If something goes wrong, the assistant files the problem report itself, with the conversation attached.
- The assistant has run on Claude, Grok, and OpenAI. We switch models when a better one fits the job.
- Job cost tracking and profitability reports show which work actually makes money.
The GSNS platform was created by Kieran Prince. CloudWise took on its modernization, integrations, and AI work in 2025.