Case study · Logistics
An AI agent that handles 4,000 shipment issues a week
Twelve people sorted delay emails by hand. Now an AI agent does most of it.
Results
- manual work
- −62%
- response, down from 2 days
- 3 hrs
- to pay back
- 11 wks
- Client
- Regional logistics operator
- Timeline
- 14 weeks
- Service
- AI Agents & Automation
01The problem
4,000 emails a week about late or damaged shipments, all sorted by hand.
02Why it was hard
Every answer needed data from three different systems.
03What we built
An AI agent that gathers the facts, fixes routine cases and passes the rest to a person.
04The result
The backlog disappeared in a month. The team now focuses on the hard cases.
Technology
- Claude
- MCP
- Temporal
- Python
Next case study
Loan decisions in 6 hours instead of 4 days
Have a problem like this?
Tell us in a few sentences. We'll tell you honestly how we can help.