Case study: AI agents across the operations of a company with its own production and couriers
A company with its own production, its own couriers and a large staff ran many business processes through people and departments. We built an AI service that brought those processes together. About 40 agents now work autonomously and do work that several departments used to do.
The client and the starting point
We don't name the client. The business, the work and the results are real.
The company has its own production, its own courier service and a large staff. Many of its business processes ran through people and departments, each holding its own piece of the work.
The aim was to bring those processes together in one service and hand work to agents, so that people keep the decisions that need a person.
What we built
One AI service across the company
An AI service that brought the company's processes together in one place.
Process optimization
Many of the company's business processes were reworked and optimized.
About 40 autonomous agents
The agents run autonomously. Together they do work that several departments used to do.
How it works now
The system does
- Do work that several departments used to do
- Run autonomously
- Work inside one service that connects the company's processes
People do
- Keep the decisions that need a person
Before and after
| Area | Before | After |
|---|---|---|
| Business processes | Run through people and separate departments | Brought together in one AI service |
| Work of several departments | Done by staff | Done by about 40 autonomous agents |
Result
The company's processes run through one service. About 40 agents work autonomously and cover work that several departments used to do, and work across many business processes was optimized. The company reduced part of its staff. The team keeps the decisions that need a person.
Results in numbers
- Around 25
- Around 1,500 a day
- Roughly 1,500 hours a month
- Down by roughly 20% over the first year
Project snapshot
- Scope
- AI service, AI agents, process automation
- Client size
- Well over 100 staff, its own production and a courier team of around 30
- Timeline
- About 6 months to the core service; agents added in stages after that
- Period
- Started in 2025; still expanding
- Team
- Two senior developers: architecture and agents, integrations and data
- Stack
- Node.js (TypeScript), PostgreSQL, n8n, LLM API, Telegram Bot API
- Work the agents took over
- 5 areas: order intake, courier dispatch and delivery updates, production scheduling, supply requests, daily reporting
- Decisions that stay with people
- Prices and discounts, hiring, supplier contracts, complaints that need judgment
Next step
Describe the situation in a few lines. One of the founders replies and says whether we can help and what the first step would be.
Discuss your project