AI in logistics and transportation — route optimisation, dock scheduling, shipping documents

European TSL today is hundreds of thousands of companies — from one-truck owner-operators to large 3PLs. Mid-sized businesses of 30–250 people (international hauliers, freight forwarders, mid-sized warehouse operators) face a demanding environment: rising labour costs, the mobility package, ETS2 adding a CO₂ cost, driver shortages, price pressure from forwarders and end customers.

In this setup AI in logistics isn't a choice — it becomes the condition for protecting the margin. Companies that automate order intake, shipping document OCR and route optimisation operate 20–40% more efficiently than competitors still running these processes by hand.

Transport orders arrive in many formats: emails with PDFs, forwarding-portal forms (Trans.eu, Timocom), Excel from regular customers, EDI from large shippers. Each channel today means dispatcher work — read, check, key into TMS, confirm to customer.

An AI order intake agent reads documents, extracts the key data (relation, dates, weight, dimensions, price, conditions), checks profitability, verifies vehicle and driver availability and prepares an order draft in the TMS. For a 50–200 vehicle fleet that's 60–80% less dispatcher work on standard orders.

CMRs, freight invoices, waybills, customs documents — in a typical transport company that's hundreds of documents a day. OCR and IDP recognise fields, extract data (date, load location, unload location, dimensions, signatures), classify the document and feed the TMS or billing system.

For a 50–200 vehicle fleet that means 50–70% less admin work and a shorter time to customer invoicing (from days to hours). That directly improves DSO and cash flow.

AI route optimisation doesn't replace the dispatcher — it supports them by analysing tens of thousands of combinations of orders, vehicles, drivers, dates and mobility-package constraints and suggesting optimal assignments. Combined with a TMS (Marcos, Inelo, Trans-iT or proprietary) it delivers 5–15% reduction in fuel and empty-mile cost.

At fleet scale of 50–200 vehicles that's hundreds of thousands of euros per year. Additionally: less dispatcher time on routine planning, more time for customer contact and non-standard orders.

Dock scheduling in customer warehouses is one of the largest hidden costs of TSL. A driver arrives at a warehouse and waits — 1, 2, 4, sometimes 8 hours — because the booking didn't match the unload slot. That's unproductive labour time, fuel, driver cost, ETS2 cost.

AI in dock scheduling integrates with customer systems (or runs as a middle layer with an AI agent), updates bookings in real time, coordinates with other carriers and alerts on delays. The result: 30–50% less driver time on customer yards.

Most TSL boards see reports in Excel with a 2–7 day delay. Power BI combined with TMS, GPS and the accounting system delivers a near real-time view: profitability per vehicle, per driver, per lane, empty-mile cost, driver time, contractual penalties, net margin.

An AI assistant in Power BI or Microsoft 365 Copilot adds management commentary — answers natural-language questions ("which lanes are least profitable?", "which driver had the most delays this quarter?"). It shifts the decision culture from reactive to proactive.

TSL companies are increasingly subject to NIS2 (as part of critical infrastructure) and security requirements from large customers (retail chains, automotive). Every AI deployment must be designed with network segmentation, access control, audit trail and AI policy in mind.

A vCISO and an advisory partner are becoming standard for companies with 100+ vehicles. This is also a precondition for many tenders — large customers require security audits and compliance evidence.

A practical 6–12 month path. Months 1–2: CMR and freight invoice OCR. Months 3–4: order intake automation for regular customers. Months 5–6: AI route optimisation with TMS integration. Months 7–9: dock scheduling, management reporting, Power BI. Months 10–12: NIS2, AI policy, security audit.

Total programme cost for a 50–200 vehicle fleet is typically EUR 45–95k spread over a year, with ROI in the first 12–18 months. Main sources: admin reduction, shorter invoicing, lower fuel cost, shorter dock waits.