AI agents in logistics: what they are and how SMEs can start

project44, the logistics visibility platform that processes data from more than 280,000 carriers, announced on July 14, 2026 that it is splitting into two businesses: the parent company keeps serving large shippers, while LSP44 launches as an independent company built around AI agents for freight forwarders, brokers and 3PLs, according to the company’s own press release. The move, also covered by Supply Chain Digital, confirms something we’re seeing more and more with clients in the sector: AI in logistics has stopped being an isolated pilot and become the feature an operator needs just to keep margin against the competition. You don’t need project44’s scale to apply the same logic to a small or mid-sized transport or warehousing business.

What project44 actually announced

According to the company, LSP44 launches already profitable — unusual for an AI logistics startup — and its agent portfolio covers three parts of an operator’s daily workflow: carrier procurement (with cost reductions the company puts at around 18%), shipment visibility (a claimed drop of over 60% in “where is my shipment” calls) and exception handling, with more than 80% auto-resolved without human intervention. These are project44’s own figures, not independently audited, but they line up with what other vendors in the sector report: AI performs best on repetitive, pattern-based tasks — tracking, rescheduling, rate comparison — and worse on decisions that need negotiation or commercial judgment.

Why it matters even if you’re not a global logistics player

The relevant takeaway for a smaller carrier, warehouse or distribution business isn’t project44’s size — it’s the three fronts its agents target. Those are exactly the processes where a mid-sized company loses the most low-value management hours. An operator running 15-30 vehicles, or a warehouse with average turnover, doesn’t need a full LSP44-style platform; it needs those same three frictions automated at a scale that matches its real shipment volume.

Carrier procurement and comparison

When a company works with several carriers or agencies, comparing rate, lead time and availability by phone or email eats hours every week. An agent connected to your regular carriers’ rates can propose the best option automatically for each shipment, leaving the final call — the same split most consumers want in other AI contexts — to the person managing the route.

Visibility without constant phone calls

Most customer service tickets in logistics are variations on the same question: where is my order. An agent that cross-references carrier tracking with order status in your ERP or TMS can answer that on its own, freeing operations staff for the cases that genuinely need a human.

Exception handling with a clear escalation threshold

Not everything should be automated to an 80% resolution rate. What’s reasonable for an SME is a clear threshold: minor delays and simple rescheduling get resolved by the agent; anything involving an affected customer, damaged goods or meaningful cost always escalates to a person. Designing that threshold is, in practice, the most important piece of consulting work before automating anything.

How to start without a global logistics budget

You don’t need to build an LSP44-style platform from scratch. The sequence we typically recommend on applied AI projects for the logistics sector is:

• Pick one process, not all three at once — usually shipment tracking, since it carries the least risk if something goes wrong early on.
• Connect the agent to the systems you already use — TMS, ERP, route planning — instead of replacing them; the value is in automating the query and the response, not migrating platforms.
• Set the escalation-to-human threshold before turning on any exception automation, tested against real cases from your operation.
• Measure in weeks, not months: hours saved and calls avoided are metrics you can see quickly if you picked the right process.

It’s the same weekly-sprint approach, with a working demo by week two, that we apply on any custom software project: validate with one real process before committing budget to anything broader.

The takeaway

project44’s move confirms an underlying trend: AI agents in logistics are moving from nice-to-have to competitive standard, echoed across transport and distribution industry commentary. An SME doesn’t need to replicate a global provider’s infrastructure to benefit from the same logic — procurement, visibility and exception handling are the three places where a well-scoped automation starts giving back management hours from week one.

Want to know where to start with your transport or warehouse operation? Get in touch: the first consultation is free and we reply within 24 hours.