A transport company earns its money in a chain of planning, execution and accountability. Planning routes, allocating vehicles and drivers, monitoring loading times, keeping track & trace up to date, handling invoices and customs paperwork, resolving complaints about delays. A large part of those hours does not go to the driving itself but to organizing it: who drives where and when, what happens if something goes wrong, and how that is recorded for the customer and shipper.
The outcome of a deal was long driven by circumstances related to staffing: how many planners there were, how experienced they were, how many drivers were available on a peak day. Whoever had more people on the phone and behind the planning board could promise more. That is the layer that is now shifting.
Route planning, load factor optimization and the initial triage of delays are tasks where software already takes over a large part of the calculation work today. Not everywhere: at one company a planning algorithm already runs independently through to the final schedule, at another a comparable system is in place but every deviating route is still assessed by a planner and only released after approval. The difference is not in the sector but in how far a company has progressed in transferring decisions and in the oversight of that.
Once that calculation work has largely fallen away, delivery time is no longer the result of how much staff is thrown at it. Two companies with an equal number of vehicles can then show a different delivery reliability figure, purely because one planning layer automatically replans further when there is a disruption and the other still does so manually. The comparison on "we are faster" thereby becomes a comparison on who absorbs the disruption fastest automatically, not on who has the largest planning department.
The same applies to track & trace and customer communication in case of delay. A large part of that can be handled by a system itself: automatic notification, recalculated ETA, a compensation proposal. What remains as human work is the exception: the cargo that is damaged, the customer who calls angrily, the situation in which a human judgment call is needed about what is reasonable. Companies that have sharply organized that distinction place their people exactly where it makes a difference, and not on routinely passing on a delay the system already knew about.
As long as delivery time was a staffing question, the party that could demonstrate it had enough people and vehicles won. Now that part of the planning runs independently, the argument shifts to something else: how robust the system is against disruption, how quickly a deviation is noticed and corrected, and who assesses the cases that are not resolved automatically. That is a different kind of proof than "we have a lot of equipment." It is proof about decision-making under disruption, not about scale.
This also changes what a shipper asks for in a tender. Where previously the question was about fleet size and staffing, the question now is about recovery time in case of disruption, how delays are communicated, who assesses the exceptions. A claim like "we deliver on time" can now only be substantiated with figures on how the system behaves in case of deviation, not with the number of drivers on the payroll.
This pattern of a shifting winning argument as soon as AI takes over a task does not only play out in transport. The same logic can be recognized in what now tips the balance in professional services now that AI takes over case-file work, in how the winning argument in retail shifts now that inventory and pricing work is being automated, and in what changes in education about the comparison between providers now that AI takes over assessment work. It is always the same question: which task shifts from staffing to oversight, and what becomes the new proof as a result.
Whether a company actually has a planning function or part of customer communication taken over is a decision for that company itself, just like the question of what that means for the people who currently do that work. If that question touches on personnel decisions, its own legal requirements apply; those are not addressed here and no substantiation for such decisions is provided here either. What matters here is narrower and more factual: which task in the chain is already largely done by software today, which task runs with human oversight, and which task remains human work, and what that means for the competitive position.
The underlying question of which work in your own planning, execution and customer contact can actually be taken over by AI is answered per task by the work scan from FTE TO AI. For the comparison with your peer group, you first need to know on which dimensions you think you are winning, and whether that can be substantiated. How to assess that yourself for your own situation is described in the explanation of how to find out what your company actually wins on, and how often that assessment needs to be redone once AI takes over more work is described in the explanation of how often a competitive analysis needs to be repeated.
The free dimension check is the first step: you identify what you think you are winning on, and see which of those claims can be defended with evidence. The full benchmark, with scores for your peer group and an evidence matrix per dimension, is under construction.