Wanted: Logistics planner, 2031. Must be comfortable disagreeing with software, and able to explain why. Will set the rules the network runs on, review what it did overnight, and decide the handful of cases it shouldn’t. Will not be asked to sort two hundred alerts before lunch.
Nobody is posting that ad yet. But we think it’s closer than it looks, and the interesting question isn’t whether the planner’s job changes. It’s what a good version of the change looks like, for the people doing the job as much as for the companies employing them.
Where the hours go today
You see many planning departments. They start the day with a queue of problems from overnight activities (trips not delivered on time, delayed vehicles, changed orders by customers, etc.). The human planner has to then sort through the problems one by one in order to solve them. Calling a carrier, moving a dock appointment, re-tendering a load etc and then updating the customer of the changes made.
Yes, it is tough work requiring considerable skill, but for the most part it is a series of patterns that need to be dealt with in a set way. Many of the decisions can be applied whilst half asleep, as they are repeatedly used for similar types of issues. In fact, all of these types of decisions that can be applied via rules that can be checked at a later time are the core of a system, and we expect that by 5 years most of the work currently spent dealing with the queue will have been transferred to software, albeit at very different speeds on different functions.
Better visibility built that queue in the first place. Seeing every slipped trip was the first step, and the next is how quickly someone, or something, acts on it, which we explored in Visibility, which was step one.
Four chores shrink first.
Routine exceptions go early. A late vehicle on a flexible delivery, or a load a carrier rejects inside an agreed rate card, doesn’t need a human once the rule is written down.
Tend to the status chasing of the planner as well. Much of the planner’s phone calls are spent locating a truck. With all systems reading the same trip information, there is no need for these calls.
Re-keying decreases too. A number being moved from one screen to another is one of the clearest indications that 2 systems are not integrated.
The default choice for repeat lanes is the fourth rule. If two carriers win the same lane every week for the same reasons, then these choices should be made by a rule and not by the planner’s time.
Of course, not all of this will disappear overnight, and some things will remain. But the proportion will decrease, and so the planner’s working day will change too.
In a 2018 Harvard Business Review piece, H. James Wilson and Paul Daugherty described the new human roles that grow up around AI as trainers, explainers and sustainers. The labels map onto planning well.
Trainer: Each of the planners in the system learns how to run your business. In a typical planner there are a bunch of rules that have been tacitly shared by people in the business, like that the carrier is awful on hills in the rain or that the customer never does split delivery. It’s a new skill to write down all those rules that people have been running with. Also, overrides on a planner are to teach as well. So, for example, where a planner has overridden a recommendation of the system, that information is also used by the system to improve. A system that does not use overrides continues to repeat mistakes that wear out the planner’s patience.
Explainer: The planner can read back what the system has done and justify it. It’s usually impractical to have someone read every decision, but it’s normal for planners to read a sample of decisions and all the cases that were flagged for some reason. They can also answer a customer’s inquiry as to why a particular shipment went through a particular hub.
Sustainer: Rules go stale. A festival season, a new carrier or a fuel price spike can quietly change the conditions a rule was written for. The planner has to watch for drift and can step a decision back to manual review while the rule is repaired if the system’s choices start to no longer match good judgment.
Of course, no label fits the fourth part of the job: the odd strategic customer request, an unknown lane, or even an unforeseen disruption that the rules can’t cover. In many cases these are exactly the reasons one became a planner in the first place.
A hypothetical, to make this concrete.
8:00. The planner opens an overnight summary and doesn’t face an alert queue. The system handled a few hundred routine decisions. It lists the ones it escalated and the ones that fell outside its confidence, along with the reason for each.
8:30. Two escalations need attention. One is a delayed vehicle carrying a time-critical load for a large customer. The planner calls the account owner, agrees to a revised window, and tells the system to hold that customer’s other loads on the same lane.
10:00. A review of a sample of yesterday’s automated decisions turns up a pattern: the system keeps picking the same carrier on one lane, and the carrier’s recent record there has slipped. The planner doesn’t override each choice. They change the rule and note why.
11:30. A new lane opens for a customer. Nothing in the system’s history covers it, so the planner keeps it in manual mode for now and puts it in shadow mode so the software can learn alongside them.
2:00. A meeting with a carrier about next quarter. The conversation is about capacity and terms, not about last week’s missed pickups, because the routine problems got handled before they became arguments.
The day is not easier. It is different, and more of it is spent on things a person is best placed to do.
What it asks of people
The World Economic Forum’s Future of Jobs Report 2025 found that employers expect roughly four in ten of the skills workers rely on today to change by 2030. Logistics planning will be no exception, and a few skills look likely to matter more.
Writing a clear rule is harder than it sounds. It is not easy to turn your instincts into clear rules that a system can follow and to realize all the exceptions you usually are unaware of.
When a system is making hundreds of decisions, as a planner, you need to be able to read data with healthy skepticism and understand when you can rely on a number and where the number came from.
Relationships with your carriers and with your customers are now much more valuable. Routine problems cease to consume your time and are left to be negotiated with carriers and trusted to customers and shippers with you making good judgment calls as to whether they will perform.
Domain knowledge is more valuable than ever. You can’t make good rules for monsoon routes around here if you haven’t experienced a monsoon rearranging your weeks.
The need for more focus on one of the consequences of software planning has become very apparent to me. The learning of junior planners today consists of them handling hundreds of very small cases under the watchful eye of a senior planner. Occasionally they get it wrong in small but survivable ways, such as a late truck on a Friday.
Will not have learned how to handle such cases when the software starts handling them as well. A single strange exception is likely to puzzle a planner who has only dealt with routine cases under the guidance of a more experienced colleague.
This is not a problem without solutions. Junior planners can learn by running hundreds of trivial cases under the supervision of a more experienced planner. However, for more trivial cases under software control, the next generation of planners needs to learn from the experience of minor errors on Fridays, and that experience needs to be transferred into the learning path for the junior’s development before that experience dies.
What leaders can do now
We don’t have to wait until 2031 to start getting ready. There are many simple steps that can be taken very quickly in most planning organizations.
Plot the planners’ week in reality and be surprised at how much of it is spent dealing with routine issues on a triage basis.
Sort your decisions first into those that a good planner could explain in a few lines of text and those that depend on more details or context. The first group can and should be automated as soon as possible. The second group are those where the planners’ best judgment is brought to bear.
Design a career path for the new skills: rule-writing and reviewing and analyzing network traffic are hard to do well and deserve to be a real role with appropriate titles and career progression.
Guard the learning path of your new planners and decide how they will acquire necessary judgment when the routine cases are pre-handled.
Where Libera fits
A planner can only work in this way if his tools allow him to work in this way. The decision process must be transparent with clear rules for the planner, a clear decision explanation, and the ability for the planner to log overrides so that they can be used for learning. This is where we are headed with the capacity and route planning engine as well as the Freight Transport Management System at Libera. There we can trace back to the award decision from the events on the road and from the settlement. It would be great to discuss how the role of the planning department could evolve in your case.
The planner of 2031 will spend less of their time reacting to issues and more time designing the system and its goals. This is far more fulfilling than spending all day dealing with hundreds of alerts. We believe companies should build such a planner deliberately. Tools are important, but so are the roles surrounding the planner: the person who trains the system, the person who reviews the system, and the person who can take over in case of an error. I want to build for this planner, not for the poor soul stuck in a never-ending queue of alerts.