You added an AI email parser six months ago. Then a spot rate lookup tool. Then an AI-assisted document checker. Your team is now running four different tools — and your operations staff are still manually re-entering the same shipment data between every one of them.
This is exactly where most forwarding operations sit right now. A 2024 global survey of logistics technology adoption found that 96% of freight forwarders have adopted at least one AI-powered tool. Only 17% describe their operations as fully automated. That 79-point gap is the real story. Not AI adoption — integration. If your operation sits somewhere in that 83%, the problem is almost certainly not the tools you've chosen. It's the manual handoffs between them.
The fastest way to reduce productivity in a forwarding office is to add a new tool that doesn't connect to the others. And that's precisely how most AI adoption happens in logistics: one tool for document extraction, another for rate comparison, a third for parsing booking instructions — each solving a narrow problem and creating a new data island.
Operations staff end up doing the integration manually. They read the output from the AI document tool and type it into the TMS. They pull a rate from the lookup tool and paste it into the quotation template. The AI saves a few minutes at the individual task level, but the workflow itself hasn't changed — it's still a sequence of manual handoffs stitched together by people.
This is the tool trap. You have AI. You don't have automation.
Full automation in a forwarding operation doesn't mean no humans. It means that when a job instruction arrives — by email, EDI, or customer portal — the system handles the data movement. It:
Your people make decisions — which carrier, which routing, how to handle an exception. The platform handles data movement, document generation, and status communication. That distinction matters enormously at scale.
In most forwarding offices, the job file lives in a TMS (or a spreadsheet) and the invoice gets raised in a separate accounting system. Someone manually transfers the charges — buy rates, sell rates, local charges, disbursements — from operations into finance. That transfer is where margin leakage and billing errors happen. A job with twelve line items might have two or three keying errors by the time the invoice goes out, and nobody knows until the DSO review at month-end.
Fully automated operations run on a unified platform where the buy and sell rates entered at quotation stage flow automatically into the job P&L, and the invoice is generated from the same data — no re-entry, no reconciliation cycle.
For forwarders who handle clearance, the gap between the forwarding job and the BE filing is typically a manual handoff. Someone exports data from the TMS, formats it, and re-enters it into the customs filing system. This adds hours to each job and creates version-control problems when shipment details change after the BE is submitted.
Integrated operations don't have this problem. When the master bill details update, the customs data updates with it. When the duty assessment comes back, it posts automatically to the job's cost sheet.
Document AI that extracts data from a BL is genuinely useful. But if the extracted data lands in a separate application and someone still has to manually push it into the forwarding system, you've automated one step in a ten-step process. The compounding value of eliminating handoffs across the entire workflow is where the real productivity gain lives — not in any single tool running in isolation.
Consider a mid-size air freight forwarder handling 300 shipments a month out of Mumbai. They adopted an AI document extraction tool for AWBs and a separate spot rate comparison tool for buying. Both work well individually.
But the extracted AWB data is still manually keyed into the TMS. Rates from the comparison tool are pasted into a quotation template. The confirmed quotation gets manually re-entered as a job instruction when the booking is confirmed. The finance team receives a daily Excel dump from operations and manually posts charges into Tally. The customs team gets job details on WhatsApp.
At 300 shipments a month, this operation employs seven operations staff and two finance staff for work that could run with four and one. The limiting factor isn't the AI tools — it's the human integrations connecting them. The AI reduced effort at the edges; the manual handoffs in the middle absorbed all the savings.
Forwarders who have achieved genuine automation share a recognisable pattern. It's less about which tools they use and more about how they're connected.
| Characteristic | The 83% (Tool Users) | The 17% (Fully Automated) |
|---|---|---|
| Operations & finance | Separate systems, manual charge transfer | Unified data model, automatic P&L posting |
| Customs integration | Manual export/re-entry into filing system | Job file feeds BE data automatically |
| Job intake | Freeform email, manual interpretation | Structured portal or standardised parser |
| Milestone triggers | Manual notifications, task reminders | System-triggered on job status change |
| P&L visibility | Visible at month-end, post-invoicing | Real-time per job, from opening |
None of this requires exotic technology. It requires a platform built for freight forwarding end-to-end — not a collection of point solutions. Freight forwarding software built on a unified data model, where operations, finance, and customs share the same job record from opening to closure, is where most of these gains come from.
You don't have to automate everything at once. Most forwarders who close the gap do it in stages, and the sequence matters:
The 17% aren't using more sophisticated AI than you are. They're running on platforms where the data doesn't stop at the edge of each tool and wait for a human to carry it to the next step. If you want to see how that works across forwarding, finance, and customs in a single job from opening to invoice, book a demo and we'll walk through it with your team's actual workflows in mind.