AI Email Automation in Logistics
What Happens When an AI Reads Your Rate Request Emails
I was demonstrating our AI Email Agent to a freight forwarding CEO in Chennai last month. Halfway through, he stopped me.
"Wait," he said. "Show me that again. Slowly."
So I did. And as I walked through exactly what the AI was doing—step by step—I watched his expression shift from skepticism to curiosity to something close to disbelief.
"That's what my operations team does," he said. "Except it takes them 45 minutes."
The AI had done it in under 4 minutes.
I realized then that most people don't actually know what AI does when it processes a rate request. They hear "AI" and imagine either magic or hype. The reality is neither. It's methodical, logical, and surprisingly similar to what a skilled human does—just faster.
Let me show you what I showed him.
The Email
Here's a real rate request (anonymized) that came into one of our customer's inboxes:
"Hi team, need urgent rates for 2x40HC from Mundra to Felixstowe. Cargo is auto parts, around 18 tons per box. Customer wants door delivery to Birmingham. Please include all charges and confirm free days. Need by EOD if possible. Thanks, Rajesh"
Looks simple, right? It's not. There's a lot packed into those four sentences. Let me break down what the AI does with it.
Step 1: Understanding the Request
The AI doesn't just scan for keywords. It reads the email like a human would—understanding context, inferring meaning, and identifying what's explicitly stated versus what's implied.
From that email, the AI extracts:
Origin: Mundra (port)
Destination: Felixstowe (port) + Birmingham (door delivery)
Container type: 40' High Cube
Quantity: 2 containers
Commodity: Auto parts
Weight: ~18 tons per container
Service type: Port-to-door
Special requirements: All-inclusive charges, free days confirmation
Notice what the AI figured out that wasn't explicitly stated: Felixstowe is the discharge port, but the final destination is Birmingham—so this needs inland haulage. "All charges" means the customer wants a single, comprehensive number. "Urgent" and "EOD" signal priority.
This is the translation work that usually lives in your best employee's head.
Step 2: Pulling the Right Rates
Now the AI knows what's needed. Next, it queries your rate engine.
This isn't a simple database lookup. The AI has to:
Find ocean freight rates for Mundra → Felixstowe (40HC)
Check multiple carriers if you have contracts with several
Pull origin charges (terminal handling, documentation, etc.)
Pull destination charges at Felixstowe
Find inland haulage rates from Felixstowe to Birmingham
Verify free days policy for each carrier option
If any rate is missing or outdated, the AI flags it. It doesn't guess. It tells you: "Inland rate for Felixstowe-Birmingham not found. Please update or confirm."
In the demo I showed that CEO, the AI pulled rates from three different carriers, compared them, and selected the one that matched the customer's priority (fastest transit, since the request said "urgent").
Step 3: Applying Your Business Logic
Here's where it gets interesting. Every forwarding company has its own margin rules, customer-specific pricing, and business logic. The AI learns yours.
For this quote, the AI applied:
Standard margin percentage for this trade lane
Customer-specific discount (this was a repeat customer with negotiated rates)
Rounding rules (your company rounds to nearest $5? $10? The AI knows)
The result: a quote that looks exactly like what your experienced pricing team would produce—because it's built on the same logic they use.
Step 4: Generating the Quote
Finally, the AI generates a professional quote document. Not a raw data dump—an actual quote formatted to your company's template.
It includes:
Complete charge breakdown (ocean, origin, destination, inland)
Total per container and total for shipment
Transit time estimate
Free days at destination
Validity period
Terms and conditions
The quote can be sent automatically, or queued for human review—depending on your comfort level and the request complexity.
What the CEO Said Next
After I walked through all four steps, the CEO was quiet for a moment. Then he asked:
"What happens when something goes wrong? When the email is confusing or the rate isn't there?"
Fair question. Here's the answer: the AI doesn't pretend to know what it doesn't know.
If the email is ambiguous ("somewhere in Europe"), it flags for clarification. If a rate is missing, it tells you which one. If the request is unusual (hazardous cargo, oversized containers), it routes to a human.
The AI handles the 80% of requests that are routine so your team can focus on the 20% that actually need human judgment.
I've found that once people see the actual process—not the marketing hype, but the real step-by-step—they get it.
AI isn't replacing your team. It's doing the repetitive, time-consuming work that keeps your team from doing what they're actually good at: solving problems, building relationships, and growing your business.
That CEO in Chennai? He signed up for a pilot the next week. His operations team was skeptical at first. Now they call the AI their "night shift colleague"—because it processes requests that come in after hours so they have quotes ready by morning.
That's not magic. That's just good technology, doing what good technology should do: making work easier.
