Parcelry: AI Freight Brokerage Software Case Study
Freight brokerage software for a parcel shipping broker: an AI assistant gathers parcel details, requests quotes from carriers, picks the best one, adds Parcelry's margin, and sends the customer a price. Once they confirm, it tracks the shipment and keeps them updated until delivery.
Industry
Logistics & Supply Chain
Timeline
12 weeks
Team
3 engineers, 1 ML engineer, 1 designer
Category
AI Development
Where things stood
Parcelry's brokers quoted every shipment by hand. They emailed customers back and forth for weights and dimensions, then phoned or emailed four or five carriers for rates, worked out the margin in a spreadsheet, and re-typed the price into a quote email. Quotes took most of a day, and once a shipment was booked, customers called in to ask where their parcel was.
How we built AI freight quoting
An AI assistant that collects parcel details
Built a chat assistant on the Parcelry website that asks for whatever is missing (origin, destination, box count, weight, dimensions, contents, and delivery date) and turns the conversation into a structured shipment request. Unclear answers, such as weight without units, get a follow-up question instead of a guess.
Carrier quotes, selection, and margin on autopilot
Each request goes out to Parcelry's carriers at once, through their APIs where they have one and by structured email where they don't, with the AI reading the replies. Quotes are compared on cost, transit time, and fragile cover against the customer's deadline, the best option gets Parcelry's margin rules applied, and the customer receives the price in the same chat. Brokers can review or override any quote before it goes out.
Confirmation and tracking in one place
When the customer confirms and pays, the shipment is booked with the chosen carrier and appears on the ops dashboard. Tracking events from the carrier update the shipment's timeline, and the customer gets an update by SMS, email, and chat at every step through to delivery.
Results from automated quotes and shipment tracking software
Minutes
From first message to a priced quote, down from most of a day
5
Carriers asked for a rate on every request, automatically
Every step
Customers updated automatically from pickup to delivery
Tech stack
Next.js
Powers the customer website, the chat widget, and the ops dashboard.
Claude API
Runs the assistant that collects parcel details and reads carrier quote replies.
Node.js
Sends quote requests to carriers, applies margin rules, and books shipments.
Carrier APIs & webhooks
Fetch rates, create bookings, and stream tracking events from each carrier.
PostgreSQL
Stores requests, quotes, margins, shipments, and the full tracking history.
Stripe
Takes payment when the customer confirms a quote.
Twilio & SendGrid
Send the SMS and email updates at each tracking step.
