Home / Blogs & Case Studies

Case study

How a 3PL cut manual work on spot freight by 80%

Yash Singh · 16 Oct 2026 · 2 min read

• ~80% less manual effort in the workflow
• 95%+ accuracy reading details from emails
• ~100% of incoming inquiries picked up

The company

Our customer is a third-party logistics company in India. One of the things it does for its clients is arrange trucks at short notice. This is called spot freight. We are keeping the company's name private for now.

The problem

Spot requests came in by email, from many clients and many different addresses. Each one started a long chain of work for the operations team.

1. Read the inquiry and work out what the client needs.
2. Send the requirement to transporters and ask for bids.
3. Collect the bids as they come back, and update or reject them as better ones arrive.
4. Award the load to one transporter.
5. Collect the vehicle number and the driver's details.

All of this ran over email. Bids came back in different formats: some in the email body, some as tables, some as attachments. The team spent a large part of the day copying details between emails and sheets, and keeping track of which request was at which stage.

On a busy day, it was easy for an inquiry to sit in the inbox for hours, or for a good bid to get missed.

What we set up

We connected an AI agent to the mailbox where spot inquiries arrive. Nothing else changed. Clients and transporters kept sending emails the way they always had, and the team kept its existing systems.

For every inquiry, the agent:

• reads the email and pulls out the route, vehicle type, load and dates
• sends the requirement to transporters for bids
• reads each bid as it comes back, and keeps track of updates and rejections
• records the award once the load is confirmed
• collects the vehicle and driver details from the transporter

Anything unclear, and any decision that involves money, goes to a person on the team. We also started with just this one workflow, instead of trying to automate everything at once.

The results

About 80% less manual effort. Most of the reading, copying and chasing is now done by the agent. The team spends its time on the requests that need a person, not on typing.

Over 95% accuracy reading emails. Inquiries and bids arrive as free text, tables and attachments. The agent pulls out the right details more than 95% of the time. When it is not sure, it flags the email for a person instead of guessing.

Close to 100% of inquiries picked up. Before, a request could get buried in a busy inbox. Now almost every one is picked up and tracked from the moment it arrives.

Unusual requests still need a person. That is on purpose. But the routine ones, which are most of them, now move with very little manual work.

What comes next

The team has since asked about using the same approach for freight invoice audit, reconciliation and billing. That would connect the rate agreed at the start of a load to the invoice that arrives at the end. It is the full loop we are building: procure, execute, audit, recover. Why invoice checking matters.

Keep reading

• Blog · 16 Oct 2026 · 3 min read

  AI agents for logistics, explained simply

  Everyone is talking about AI agents. Here is what one actually does in a logistics team, without the jargon.

• Blog · 16 Oct 2026 · 3 min read

  Why most freight invoices never get a proper check

  Most finance teams say they audit freight invoices. Very few can check all of them properly. Here is why, and what it costs.

No Horn: the autonomous back office for enterprise logistics. Write to yash@nohorn.in.