The trial that should have worked

A few weeks ago, one of our companies was running a field trial for a new product. The lab testing had gone well, the engineering team was confident in what we had built and, based on everything we had seen internally, the product was ready.

Then it went into a US field deployment and stopped working.

Not everywhere, and not consistently. In some ways, that made the problem harder to diagnose because we weren’t dealing with a straightforward failure that could immediately be reproduced. Something about the operating environment was exposing a problem we hadn’t encountered during testing.

That’s one of the realities of building technology for the environments our customers operate in. You can test extensively before deployment, but the field will sometimes introduce conditions you didn’t anticipate.

Finding the problem was only half the job

After three days of investigation, the team traced the failure to a network edge case. It was specific enough that we hadn’t encountered it during pre-launch testing, but significant enough to affect the product in the field.

Finding the likely cause didn’t solve the whole problem, though.

We couldn’t properly verify the fix remotely.

We could push a firmware update and ask the customer to test it for us, but that would effectively make the customer responsible for proving that our solution worked. If we were going to tell them the issue had been resolved, we needed to be confident of that ourselves.

The quickest way forward was to get someone on site who could work directly with the engineering team, run diagnostics and test each iteration under the same conditions causing the failure.

I had a valid US visa and could travel immediately, so I went.

The field showed us what the lab couldn’t

Testing on site changed what we could see.

The network issue was confirmed, but working under the actual operating conditions also surfaced additional issues that hadn’t been visible remotely. That meant the job wasn’t simply to apply the fix we had already identified. The team needed to keep testing, learning and adjusting as new information came back from the field.

Over the next three days, we worked through twelve firmware iterations.

Some changes moved us closer to the answer. Others exposed something else that needed attention. Each iteration was tested on site, with the results going back to the engineering team so that the next version could be worked through.

Eventually, the product was operating reliably.

But an apparent fix and a proven fix are not necessarily the same thing. Testing continued for another three days before we were comfortable telling the customer that the product was ready to go into production.

That additional time mattered. The objective wasn’t to get to the first version that appeared to work. It was to reach the point where we were prepared to stand behind it.

Owning the technology means owning the problem

The experience reinforced something important about the way we’ve chosen to build technology across our group.

Our teams work across the hardware, firmware and software rather than relying on a chain of third parties to make the product work. That gives us greater visibility across the technology and allows our teams to iterate quickly when something needs attention.

But ownership works both ways.

When something goes wrong, there isn’t another vendor somewhere in the chain that we can point to and say, “That’s their problem.”

If the hardware is ours, we own the problem. If the firmware is ours, we own the problem. If the software is ours, we own the problem.

And that responsibility doesn’t belong to engineering alone.

I said to the team afterwards that everyone, including the cleaner, needs to be able to stand behind what we put into the market. That doesn’t mean everyone needs to understand the firmware or know how to diagnose a network issue. It means the standard belongs to the whole company.

Customers don’t buy our organisational structure. They shouldn’t need to understand which team, supplier or component was responsible when something doesn’t work. They bought the product from us, and the responsibility for making it work sits with us.

What standing behind a product actually means

Of course, we’d prefer every product to work perfectly the first time it reaches the field. But building technology for real operating environments means accepting that conditions will sometimes expose something testing didn’t.

The important part is what happens next.

How quickly does the problem reach the people who can solve it? Do they have enough ownership of the technology to investigate it properly? Does the organisation stop at the first plausible fix, or keep testing until the result can be trusted?

In this case, the answer required people across the business working through twelve firmware iterations and continuing to test after the product appeared to be stable.

The specifics will be different next time. The standard shouldn’t be.

When we say we stand behind what we build, that promise only really means something when standing behind it costs us something: time, effort, another iteration, or the willingness to get closer to the problem rather than further away from it.

Because customers don’t need a promise that nothing will ever go wrong.

They need to know what we’ll do when something does.

Three things I always check in a security audit:

1. Is your tech from five years ago still relevant, or is it easy to circumvent with what’s out there now?

2. Are you tracking the vehicle or the cargo?

3. Is your security provider’s IT policy actually solid — ISO 27001 certified, gold standard? Because if it isn’t, your weakest point isn’t the truck. It’s them.

When last did you actually audit your security provider, not just your fleet?

I was at TAPA’s conference in Oberhausen two weeks ago, there in my capacity as South Africa’s regional lead for TAPA EMEA. What I’m seeing on the ground at home lines up with what the room was discussing: attacks on transport are up. Cargo theft is up. Diesel theft is up. And the methods being used are getting more advanced, not less.

The discussions weren’t about whether transport crime is changing; they were about how quickly criminal methods are evolving, and whether operators are adapting at the same pace.

What struck me most wasn’t simply that incidents are increasing. It was how often the conversation came back to the same point: many operators are still relying on security strategies designed for a very different threat landscape.

Two forces, one storm

I’m not going to pretend I can solve South Africa’s unemployment problem in a Substack article. That’s a political and economic issue, not one for people in fleet management.

But it would also be naïve to ignore the environment we’re operating in. According to Statistics South Africa, youth unemployment remains among the highest in the world. That doesn’t excuse criminal activity, but it does form part of the reality transport operators have to manage.

Combine that pressure with criminals who are constantly adapting their methods, and you have a threat landscape that looks very different from the one many businesses planned for five years ago. The problem isn’t that operators stopped investing in security. It’s that the threat continued evolving after those investments were made.

The question nobody wants to answer honestly

Here’s the uncomfortable part. Most of the security infrastructure I encounter in this industry — across fleet operators, logistics companies, fuel distributors — was designed for the threat landscape of three to five years ago. It was good when it went in. Some of it was excellent.

But criminals don’t stand still, and too many operators do.

One thing I’ve learned over the years is that assumptions are expensive. The best operators I work with don’t ignore security; they challenge it. They assume that if the threat changes, their security strategy needs to be questioned as well. Most operators don’t ignore security; they simply assume that the systems which protected them a few years ago are still protecting them today. That’s an assumption worth testing.

If your last serious security review predates the last meaningful change in criminal methods, you’re not protected; you’re simply relying on assumptions that haven’t been tested. There’s a difference, and you usually only discover which one you were on the night it matters.

When did you last run a technical audit of your security systems — not a policy sign-off, not a checklist review, but an actual assessment of whether your hardware and protocols can stop what criminals are doing today?

If the honest answer is “I’m not sure” or “a few years ago,” you already know what your next move needs to be.

What this means practically

There isn’t a single checklist that fits every operation, but there are a few questions every operator should be asking.

• Audit against current methods, not past ones. Ask your security provider what’s changed in criminal tactics in the last 18 months, and whether your current setup accounts for it.

• Don’t confuse “installed” with “current.” A system that was cutting-edge five years ago is not automatically cutting-edge today. Technology ages, and so does its effectiveness against people actively trying to defeat it.

• Treat this as ongoing, not once-off. The criminals adapting to your defences aren’t doing it once and stopping. Neither should your review process.

South Africa’s broader socioeconomic challenges aren’t something any of us in logistics are going to solve on our own. But making sure our security measures reflect today’s risks instead of yesterday’s assumptions- that’s something every operator can control.

The biggest vulnerabilities I come across are rarely the result of businesses doing nothing. More often, they’re the result of good businesses assuming the controls they put in place years ago are still enough today.

So, I’ll leave you with one question.

When did you last check?

The most surprising thing about AI adoption in 2026 isn’t the technology.

It’s who’s leading it.

It’s not the 25-year-olds. It’s the 45- to 55-year-old fleet owners, the same people who once asked me what a browser was, now demanding why their AI driver behaviour scoring isn’t integrated with their fuel dashboard yet.

Something fundamental has changed.

In 1999, I was 18, running my first company selling networks, internet connections, and cell phones. Our customers were 40- and 50-year-old business owners. Selling them a shared printer felt like a philosophical debate. The internet was alien. Email? “Why would I type what I can say on the phone?”

That resistance wasn’t ignorance. It was rational. They had no mental model for trusting the invisible.

Fast forward to 2026, and AI is spreading faster than the internet ever did.

Why? Because today’s decision-makers have been trained by every technology wave since.

They moved from Nokia bricks to smartphones. From “I’ll never bank online” to doing it in N1 traffic. From doubting GPS tracking to making it non-negotiable. Each shift taught the same lesson: Uncomfortable at first → normal → can’t operate without it.

That lived experience has removed most of the psychological friction.

On the ground in South African logistics, the conversation has flipped.

Ten years ago, talking GPS tracking: “How do I know the data is accurate? What if it fails?”

Today, talking AI route optimization, predictive fuel theft detection, or driver behaviour scoring:

“What’s the implementation timeline and ROI?”

The education phase is over. The pre-conditioning is done.

Subscribe now

The uncomfortable truth:

The real constraint is no longer fear; it’s execution speed. The gap between early adopters and the rest is widening faster than in any previous cycle. In AI-driven logistics, that gap will be measured in margin, not just efficiency.

The operators pulling ahead aren’t necessarily the biggest. They’re the ones learning by doing, finding hidden 18% fuel variances, cutting insurance exposure, and seeing their numbers in real time instead of month-end surprises.

Ask yourself honestly: How long did you wait before trusting internet banking?

Do you regret it?

If you’re a fleet operator, supplier, or technology buyer in African logistics, the message is clear: the psychological barrier is gone. What’s left is a simple choice: move early and compound your advantage, or spend years catching up.

The future moves fast. It doesn’t wait for certainty.

— Pieter Smits

China banned companies from replacing employees with AI. The first major commercial order of self-driving trucks just landed in the United States. Ukraine is using autonomous vehicles to supply frontlines without putting human lives at risk.

The pace of change is real. And the debate that follows it is always the same.

Are all our jobs going to disappear?

I’ve been sitting with that question for a while. And I think the honest answer, the one that’s actually useful to leaders, workers, and policymakers, is more nuanced than either side of the debate usually admits.

What I saw at Walmart

Earlier this year, I was lucky enough to visit one of Walmart’s main distribution centres near Los Angeles. One of the most automated supply chain operations I’ve seen anywhere in the world. Conveyor belts, robotics, and advanced automated picking, the scale was genuinely extraordinary.

And then I asked a simple question.

“How do you load the packages into the trucks?”

The answer stopped me.

“Some people are really, really good at Tetris.”

Even there, at the cutting edge of logistics automation, in one of the most technologically advanced retail supply chains on the planet, human creativity, spatial reasoning, and real-time judgment were still doing work that automation hadn’t fully solved.

Not because the technology doesn’t exist in theory. Because getting automation right for 80% of cases is one problem. Getting it right for 100%, every edge case, every exception, every real-world variable that humans handle on instinct, is a fundamentally different problem. And often far more costly to solve than it’s worth.

That gap between 80% and 100% is where humans still live. And it’s larger than most people realise.

The jobs that will disappear, and the ones that won't

Let me be honest about something first. Some jobs will phase out. Pretending otherwise doesn’t serve anyone.

Roles built almost entirely around repetitive, predictable, high-volume tasks are genuinely vulnerable. If a job is essentially a set of rules applied consistently to consistent inputs, automation will get there. It’s a matter of when, not if.

But that is not most jobs. And it is not the most valuable part of most jobs.

What automation consistently struggles with is the layer of work that requires contextual judgment decisions that depend on reading a situation rather than following a rule. Creative problem-solving under conditions that weren’t anticipated when the system was designed. Human-to-human moments that require empathy, trust, and relationship. The exceptions that fall outside every defined process.

These aren’t niche skills reserved for a small elite. They show up on the floor of a distribution centre as the person who instinctively knows how to fit one more box into a truck. They show up in logistics as the driver who knows this particular client’s depot runs late on Fridays and adjusts without being told. They show up in a development team as the engineer who understands not just what the code should do but why the business needs it to behave that way.

The jobs that survive automation aren’t the ones hiding from it. They’re the ones doing what automation still can’t.

The 80% problem nobody talks about

There’s a version of the automation conversation that treats the remaining 20% as a rounding error. A temporary inconvenience on the way to full automation.

I think that’s wrong. And the Walmart example illustrates why.

Getting a distribution centre to 80% automation is an extraordinary engineering achievement. It requires capital, infrastructure, systems integration, and years of operational refinement. The companies that have done it are genuinely world-class.

But that last 20%, the loading, the exceptions, the edge cases, the judgment calls, often require a disproportionate amount of additional investment to automate. The returns diminish sharply. And in many cases, the human solution is faster, cheaper, and more adaptable than the automated one.

This is not a failure of technology. It’s an honest reflection of where human capability genuinely excels. We are extraordinarily good at navigating ambiguity, handling novelty, and applying judgment in conditions we’ve never encountered before. Those are not skills that replicate easily.

The 80% that gets automated is rarely the most valuable 80%. The hardest parts become more valuable, not less, as everything around them gets faster.

What this means for Africa specifically

I’ve spoken before about why the AI conversation looks different in Africa than it does in the US or Europe. The stakes around employment here are not abstract. Youth unemployment is a structural crisis. The anxiety about automation isn’t theoretical; it’s immediate and personal for millions of people.

Which is exactly why the framing matters so much.

If we position AI purely as a replacement technology, something that comes for jobs, we create fear, resistance, and a defensive posture that ultimately slows adoption and leaves people behind anyway.

If we position AI as an augmentation technology, something that makes people significantly more capable at the work that matters most, we create a completely different outcome. One where Africa’s large, young, growing workforce pairs with tools that amplify what they can do rather than rendering it obsolete.

The Walmart story isn’t a story about automation winning. It’s a story about human creativity surviving and thriving inside one of the most automated environments in the world.

That’s the story Africa needs to be telling. And building toward.

The leader's job in this moment

For those of us leading organisations through this period, the question is specific.

It is not: how do we protect our people from automation? That framing leads to delay, defensiveness, and ultimately a worse outcome for everyone.

It is: how do we help our people find the spots where they will genuinely excel in an AI-augmented environment, and how do we build the culture, the training, and the tools to support that transition?

That means being honest about which roles will change and how. It means investing in the skills that sit in the 20%, judgment, creativity, contextual thinking, rather than only the technical skills that AI is most likely to absorb. It means keeping people motivated through a transformation that, handled badly, breeds fear, and handled well, breeds capability.

Self-driving trucks are coming, and warehouse robots are already here. Autonomous supply vehicles are protecting lives on frontlines right now.

And somewhere in a distribution centre near Los Angeles, someone is loading a truck in a way no algorithm has figured out yet, doing work that sits quietly in the gap automation hasn’t closed.

Both things are true. And the leaders who hold both of those truths at once, rather than collapsing into either panic or complacency, are the ones who will navigate this well.

The question I keep coming back to

Not: are all our jobs going to disappear?

But: how do we help our people become the best possible version of themselves in a world where AI handles more and more of the repetitive layer?

For Africa, for our organisations, and for us as leaders, getting that answer right is the work.

Where are you seeing humans outperform automation in your own operations? I’d be genuinely interested to hear what’s working on the ground.

My gut reaction was scepticism. It feels like the kind of policy that sounds good in a press release but creates real friction in practice, slowing down innovation, protecting inefficiency, and ultimately putting companies at a competitive disadvantage.

But I kept thinking about it. And when I applied the same ruling to the African context, something shifted.

Africa has some of the highest unemployment rates in the world. Youth unemployment in particular is a structural crisis that no single policy or technology is going to solve overnight. So when AI enters the conversation here, the stakes are different. The anxiety isn’t abstract; it’s immediate and real for millions of people.

Which made me ask a different question entirely.

What if the frame of “replacement vs. no replacement” is the wrong frame altogether?

The architect analogy

In our R&D team at Globaltrack, we’ve been using AI coding tools Claude Code and OpenAI Codex to accelerate our development process. And what I’ve watched happen is genuinely interesting.

A software developer, traditionally, had to wear every hat. Architect, plumber, bricklayer, plasterer. Every layer of a build, from high-level system design down to writing boilerplate code, sat on the same person. That spread is cognitively expensive. It limits how deep any one person can go on any one layer.

With AI handling the execution layers, something changes. The developer can sit almost entirely in the architect’s chair. They’re thinking about structure, about system design, about the problem at a level they rarely had time to before. And here’s what nobody told me would happen: the more you focus purely on architecture, the better architect you become. Removing the execution burden doesn’t make you lazier. It makes you sharper.

That is augmentation. Not replacement.

Why this matters for Africa specifically

If we accept that AI, when implemented thoughtfully, can make people significantly more capable rather than redundant, then the conversation changes completely.

It becomes about access. About training. About which organisations and governments are willing to invest in helping their people grow into an AI-augmented way of working, rather than simply protecting the status quo or, on the other extreme, automating without consideration for the human cost.

Africa’s young workforce is one of its greatest assets. Pairing that workforce with AI tools that amplify capability rather than displace it could be genuinely transformative. Not in a hype-cycle way. In a practical, skills-building, productivity-unlocking way.

China’s ruling may be blunt. But the instinct behind it that technology should serve people, not simply replace them, is worth taking seriously.

The question I keep returning to isn’t “how will AI replace us?”

It’s “how do we use AI to become a better version of ourselves, way faster?”

For Africa, getting that answer right matters more than almost anywhere else.

How is your organisation approaching AI augmentation vs. replacement? I’d be interested to hear what’s actually working on the ground.

The logic is sound. Millions of litres in storage, a price jump days away. That’s not reckless — that’s basic commodity trading.

But something broke in the execution this time.

Farmers are unable to operate. Transporters are grounded. People genuinely cannot get to work. The conversation in the market has shifted from “fuel is expensive” to “fuel might not be there at all” — and that is a fundamentally different crisis.

There is a tightrope between capitalising on a market opportunity and capitalising on your customers’ desperation. I think the industry crossed it.

What makes this harder is the context. The Middle East situation is real, ongoing, and outside anyone’s control here. Southern Africa was always going to feel the pressure. But the internal supply chain conflict — the hoarding, the distrust, the positioning — that part is a choice.

And it’s a choice that can be made differently.

You can still run a profitable business. You can still make money in a tight market. But there is a version of this where the industry works together, rebuilds trust in the supply chain, and comes out of this period with its customer relationships intact rather than fractured.

That version is still available. But the window doesn’t stay open forever.

What’s your read on this? Are fuel suppliers justified — or did the industry miscalculate the cost of lost trust?

There’s a growing debate at the moment around fuel suppliers holding stock in anticipation of price increases, particularly as supply pressures continue to affect parts of the world and Southern Africa more specifically.

On the surface, it’s easy to understand the logic behind it. If a supplier is sitting on millions of litres of diesel, waiting a few days or even a week before releasing that stock can result in significantly higher margins. From a purely commercial perspective, it’s a rational decision. Businesses are built to optimise for profit, and in isolation, this is exactly that.

Where the Logic Starts to Break

The challenge is that supply chains do not operate in isolation.

What is often missed in this conversation is the broader impact that these decisions have on the system as a whole. When suppliers begin to hold back product in anticipation of higher prices, the immediate effect is not just a shift in margins, but a shift in perception.

Customers start to question whether their suppliers are still aligned with their interests, or whether they are now operating against them.

That is where the real risk begins.

Trust Is Not a Soft Metric

In logistics and supply chains, trust is not an abstract concept. It is operational. It underpins planning, decision-making, and execution.

When that trust starts to erode, even slightly, the effects ripple quickly across the system.

Transporters become hesitant to commit to routes without certainty of supply. Farmers begin to worry about whether they will have access to the fuel required to keep operations running. Businesses delay decisions because the variables they depend on are no longer stable.

In some cases, it becomes more than just a commercial issue. It affects people’s ability to move, to work, and to maintain a level of normalcy in their daily lives.

What started as a rational pricing decision begins to create systemic friction.

From Market Behaviour to System Risk

This is where the conversation needs to shift.

The question is not whether businesses should make money. They should, and they must. Profitability is what allows companies to operate, invest, and grow.

The real question is how that profitability is achieved in moments where the system itself is under pressure.

Short-term optimisation at one point in the supply chain can destabilise the entire network if it is not balanced with a broader sense of responsibility.

The fuel supply chain is not just a market; it is infrastructure. And infrastructure depends on coordination, not just competition.

The Role of Leadership in Moments Like This

When different parts of the system begin to work against each other, the overall outcome is worse for everyone involved, including those who may have gained in the short term.

This is why moments like these call for a different kind of leadership.

Not the kind that focuses purely on extracting maximum value from the situation, but the kind that recognises the importance of maintaining trust while still operating as a business.

There is always a way to balance both. It may not maximise margins in the immediate term, but it preserves something far more valuable in the long run: the integrity of the system.

What This Really Comes Down To

Ultimately, this is not just a conversation about fuel or pricing.

It is a conversation about how we choose to operate within interconnected systems. Whether we see ourselves as individual players or as part of something that only works when there is alignment.

The real opportunity in moments like this is not just to perform better as individual businesses, but to ensure that the system continues to function as it should.

Because when the system works, everyone wins.

So I did what more people are starting to do instinctively: I took out my phone, took a photo, and ran it through ChatGPT. Within minutes, I had a clear home treatment plan, and within a few days, she was completely fine.

I’ve been thinking about that moment a lot since.

Because of how ordinary it felt, there was no panic on my side, just a practical solution, available immediately, from wherever I happened to be standing.

That, I think, is what most people get wrong about AI. They’re waiting for the sci-fi version. The one that replaces entire departments overnight and makes the news. But the real shift, the one that actually matters, is quieter. It’s the gap between people who reach for these tools and people who don’t.

The Echo of 2020

I want to set some context, because this week has felt a little like early Covid again.

Peace talks are breaking down between Iran and the US. Fuel supply uncertainty is back on the agenda, and LNG prices are moving. Some countries are floating working-from-home conversations again. The world is messy, as it tends to be, and that messiness lands squarely on logistics and supply chain every single time.

We’ve been here before. I remember the scramble in 2020: operations built on physical presence suddenly needing to function remotely, supply chains that assumed certain trade routes suddenly rerouting on the fly, fuel cost modelling that was obsolete within weeks.

Subscribe now

The businesses that got through it weren’t the ones with the biggest fleets or the deepest pockets. They were the ones who adapted the fastest. Who found digital substitutes for physical assumptions? Who asked “what does this situation actually require?” instead of “how do we get back to normal?”

Here we are again with one very significant difference.

The Tools Are Incomparably Better.

During COVID, the big digital unlock was things like Zoom, e-signatures, and digital dispatch. Genuinely useful, but fairly blunt instruments.

What we have now is a different category of capability entirely. OpenAI released another major update just this past week. The pace of improvement isn’t slowing; rather, it’s accelerating. And the practical applications are growing faster than most operators have time to explore.

In logistics specifically, what I’m watching:

Route exceptions are being handled more proactively, not reactively. A truck breaks down, and the system has already suggested alternatives before the dispatcher picks up the phone. The dispatcher’s job doesn’t disappear; they just make the call. What they’re not doing is starting from scratch at 11 pm.

Fuel anomalies are being caught in near real-time. Not on a Friday morning, report telling you what happened on Tuesday. Driver behaviour patterns surface before they become incidents. Fatigue flags, route deviations, and SOP compliance gaps. Not because someone went looking, but because the system flagged it.

AI tools are starting to meaningfully compress the administrative load in cross-border logistics across Africa, not eliminating human judgment, but eliminating the repetitive groundwork that burns hours.

None of this is science fiction, and it’s available now. The question is adoption.

The Fax-To-Email Shift, Again

I’ve used this analogy before, but it keeps being the right one. When email became mainstream, some businesses shifted quickly, some businesses shifted slowly, and others waited until they had no choice. The ones who waited didn’t go out of business overnight, but the gap between them and the early movers compounded year by year until it couldn’t be closed.

Subscribe now

We’re in that transition again.

The people who will feel this most acutely aren’t the big enterprise logistics players with dedicated tech teams. It’s the mid-market operators who are either building a compounding advantage right now or slowly giving ground to competitors who are.

What my daughter’s rash taught me indirectly, from a stoep in the middle of nowhere, is that the utility of these tools doesn’t require proximity to a major city, a dedicated IT department, or a digital transformation budget. It requires the habit of reaching for them. It requires the mental model that says: before I drive three hours, let me try something else first.

The Practical Question

If you’re running a fleet operation and reading this, here’s the honest question I’d ask: where are you still doing manually what a system could be doing for you?

Not the big strategic stuff. The daily grind. The reconciling fuel slips at the month’s end. The rebuilding of routes when something goes wrong. The realisation on Friday that something had happened on Tuesday.

That’s where the leverage is. Not in replacing your operations team, but giving them leverage they don’t currently have.

Share

The disruptions coming, and they are coming, fuel supply and geopolitics don’t care about your quarterly targets, will reward the operations that can adapt quickly. The ones with real-time visibility. The ones where the human in the loop has good information and good tools, not just good intentions.

That’s the shift. It’s already underway. The question is where you are in it.

Q1 is done. And if I am honest, it was one of the most interesting and at times disorienting quarters I can remember in a long time.

Not because of the geopolitical noise, although there was plenty of that. But because of the pace. The sheer speed at which things moved across every company in our group left me reflecting in a way I did not expect.

I want to share some of what I learned.

The speed was real. So were the problems.

Early in Q1, we leaned hard into Claude Code and Codex across our development teams. The promise was real, and it was delivered. Speed went up. Significantly. Things that used to take days started taking hours.

But here is what nobody tells you about that kind of acceleration.

New problems appear that never existed before. Problems you have no playbook for. No institutional memory to draw on. Because they are genuinely new.

Bugs that looked like one thing but were something else entirely. Integration issues that the old way would have caught early, but that the new pace skipped past. Moments where the team was moving so fast that the review process could not keep up.

Q1 was a big lesson in that. And I want to be clear, I am not saying the technology is the problem. The technology is extraordinary. The lesson is about how you integrate it. How do you build the processes around it? How do you keep your teams anchored even as the ground is shifting?

We are making great progress. I am genuinely proud of the developers across the group who embraced this shift. But the learning was real, and I think it is worth being honest about it.

The world is not ending. But you cannot look away.

There is a version of the AI conversation that says everything is about to change overnight. Every job is replaced. Every industry is disrupted. The world will be unrecognisable in three years.

I do not believe that.

And there is another version that says it is all hype. That nothing fundamental is really changing. That the sensible move is to wait and see.

I do not believe that either.

The truth I keep coming back to is that both things are partially right, and holding them in tension is where the real work is.

The world is not changing overnight. But the pace is fast enough that you cannot afford to look away. The companies and leaders who will come out ahead are not the ones going all in on every new tool. They are the ones who have figured out how to move with purpose, adopting what creates real value, building the processes to support it, and keeping people at the centre of it.

That balance, two firm feet on the ground, at least one hand on the technology, is not a compromise. It is a strategy.

The moves we made.

Beyond the technology side, Q1 brought real momentum on other fronts.

Globaltrack’s expansion into Europe is something I am genuinely excited about. We believe we can add strong value there, and the early signals are encouraging. The demand for security of shipments, containers on road, and on rail, is significant. We spent time with a trial customer in the US and saw hundreds of trucks moving through their operation every hour, with our technology implemented and working; that was a moment.

It is one thing to build something. It is another to see it working at scale, in a market you are new to, solving a problem that actually matters. Q1 had several of those moments.

We also had great new people join across the different companies in the group. That matters more than most things. The right people, in the right environment, moving with the right tools, that is where the real progress comes from.

What I am taking into Q2.

A few things I keep coming back to as we head into the next quarter:

First, the technology is not the hard part. Learning how to move with it without losing what already works is the hard part. Do not throw out the baby with the bathwater. But do not pretend the bathwater is fine either.

Second, speed without review is a risk. The faster your team moves, the more important your review and integration processes become. Not to slow things down. To catch what the speed misses.

Third, the leaders who will struggle in the next few years are not the ones who move too fast. They are the ones who are not moving at all. Waiting for a perfect moment that is not coming. Hoping that the tools their competitors are using will somehow not affect them.

The question is not whether to embrace technology. It is where to find the balance between moving with it and staying grounded. That is what Q2 is about for us.

And I think that quest is worth sharing. Not because I have all the answers. But because the questions are real, they are affecting real businesses, and sometimes the most useful thing a leader can do is say honestly, this is where we are, this is what we learned, and this is where we are going.

A question for you.

How are you finding that balance in your own business or career right now?

Are you moving fast enough? Or are you holding back, waiting for more certainty that is never going to come?

I would genuinely like to know. Drop a comment or reply to this note directly. These conversations are some of the most useful ones I have.

— Pieter

The disruption is real. The oil is not flowing the way it was. Diesel prices, which were already high, have now moved in a direction that is making many transport businesses genuinely nervous.

But here is the part that nobody wants to say out loud.

The fuel price itself is not what is going to break some of these operators. What is going to break them is the fact that their contracts were already signed before any of this happened.

Locked In. Costs of moving. No Exit.

You negotiated your rates. You committed to your customers. You built your margins, carefully, realistically, around a fuel cost that made sense at the time. And now, through no fault of your own, that number has moved significantly in the wrong direction.

Not all contracts have fuel escalation clauses. In fact, most do not. Which means the operator absorbs the difference. Every litre. Every kilometre. Every month until the contract comes up for renewal, if the business is still standing by then.

This is not hypothetical. This is happening right now, to real operators, running real fleets, trying to honour commitments they made in a different cost environment.

Some of them are going to make it through this. Some of them are not.

And the thing that separates those two groups is not the size of their fleet, not the quality of their customer relationships, and not how hard they are working right now. Most of them are working incredibly hard.

The thing that separates them is what they built before the storm arrived.

The $1,800 Number That Changes the Conversation

One of the companies in our group made a decision, not in response to a crisis, but as a straightforward operational improvement, to implement a focused fuel management solution. They started tracking consumption per vehicle, per route, per driver, and systematically fixing the inefficiencies they found.

I’m quite proud of what they achieved. They were speaking with one of their customers recently, and the customer realised that before the current fuel spike, they were already saving approximately $1,800 per truck per month. The ROI on that system was sitting at somewhere between 500 and 600%.

That saving wasn’t created during the crisis. It was already in place when the shock hit. And now it has become a real buffer, the difference between a difficult quarter and a genuinely dangerous one.

The operators who had that kind of visibility running are absorbing something that is breaking others.

Subscribe now

Why “We’ll Get to It Eventually” Is Actually a Strategy, Just a Bad One

I understand why technology adoption gets delayed in logistics. This is a demanding industry. The day-to-day pressure is relentless. There is always a fire to fight, a driver to find, a border delay to manage, and a customer who needs an update right now.

Sitting down to evaluate telematics or fuel management systems can feel like a luxury when you’re managing exceptions from six in the morning until late at night.

But here is the uncomfortable truth: that repeated decision to delay is itself a strategic choice. It is a choice to remain exposed. To keep your margins thinner than they need to be. And to have no buffer when something unexpected hits your cost base.

And in today’s world, unexpected things are hitting cost bases constantly.

Even in the more “old-style” industries like logistics, the companies that adapt to and embrace technology are the ones more able to handle the shock of massive diesel price increases. The ones who don’t move struggle, and eventually, something unexpected can set you back to a level that’s hard to come back from.

The Honest Question

I am not writing this to be alarmist. Logistics is a resilient industry, full of resilient people, and the operators navigating this right now deserve real credit for how they are managing.

But this moment is asking every transport business a question worth answering honestly:

If another cost shock hits your operation in the next six months, and in this environment, the probability is not low that you are in the group that absorbs it, or the group that scrambles.

That answer is not fixed. It is a choice. And the window to make it before the next disruption arrives is right now.

The operators who are not panicking this month made a quiet decision, months or years ago, to invest in visibility, in data, in the tools that give them control over the costs they can actually control.

It is not glamorous. It is not a Silicon Valley story about disruption and transformation.

It is just good operations. Built before you needed them.

If you are navigating this right now, I would love to hear how your operation is holding up and what is working. Drop me a note or leave a comment below.

Leave a comment

Driverless trucks.

Fully automated supply chains.

AI replacing entire teams.

But if you spend time inside real logistics operations, the story looks very different. Automation has actually been entering logistics for a long time already. First, the fax machine replaced phone calls. Then, email replaced fax. Then digital systems began replacing lengthy email chains among dispatchers, customers, and drivers.

Each step reduced manual coordination, but none eliminated people from the system.

Logistics Looks Simple Until You Operate It

From the outside, logistics looks straightforward. It is moving something from point A to point B. But anyone who has worked inside the industry knows that is rarely how it plays out. A truck arrives late, documents go missing at the border, the weather disrupts a delivery window, or customers suddenly change their schedule.

These things happen every day, and they are not unusual events. They are the normal operating environment, and that is where automation runs into reality.

Where Automation Actually Helps

Automation is extremely good at handling the things that repeat over and over again, such as tracking vehicles, processing status updates, or monitoring fuel usage.

These tasks happen thousands of times across fleets every single day. Machines are very good at this kind of work. They do not get tired, and they do not miss patterns easily. When systems begin to take over these layers of coordination, something interesting happens: routine work begins to disappear.

Real-world examples show clear returns, companies implementing telematics, route optimization, and RPA in freight processing often see 20-40% reductions in operational costs, faster cycle times (up to 50% in some cases), and significant labor savings on repetitive tasks, according to reports from McKinsey and industry cases like optimized routing at UPS or AI-driven efficiencies in Asian logistics providers.

What Rises to the Surface

Once routine coordination is automated, the remaining work becomes much more visible and tends to look different.

Unexpected delays.

Customer communication.

Operational trade-offs that require judgement.

In other words, the parts of logistics where context matters. This is where people remain essential. Logistics rarely breaks in predictable ways. When something goes wrong, it usually happens in a messy combination of circumstances that a system cannot fully anticipate. Someone still needs to step in and decide what to do next.

Automating the Bulk

The most effective way to think about automation in logistics is actually quite simple.

Let the systems handle the predictable bulk work. Let people handle the exceptions.

At one of our companies, Globaltrack, that balance is something we think about often. The goal is not to remove operators from the process. The goal is to remove the repetitive coordination that consumes their day. When that happens, operators spend less time watching dashboards and more time solving real problems. That tends to make both the system and the job itself better.

To integrate without major disruption, the key is phased implementation: start with pilots in high-impact areas (like tracking or exception alerts), ensure tight integration with existing systems (WMS, telematics), map processes first to avoid silos, and provide training so teams shift from routine monitoring to higher-value judgment and problem-solving. This redeployment often boosts job satisfaction and retention, as staff members move toward strategic roles rather than being displaced.

Why This Matters Even More in Emerging Markets

In many emerging logistics environments, like much of Africa, infrastructure variability, border complexity, poor road quality, and operational unpredictability introduce far more exceptions than highly automated systems (think heavy robotics) are usually designed for. Here, lighter, more flexible solutions shine: real-time telematics for visibility, digital platforms matching loads to carriers, predictive ETAs factoring in local weather/traffic, and mobile apps for drivers. These improve coordination and reduce costs without over-relying on rigid infrastructure. Experienced operators remain critical when conditions change, and in emerging markets, conditions change often. This is why the conversation around automation sometimes sounds very different depending on where you are operating.

Risks and Pitfalls to Watch

Scaling automation isn’t risk-free. Common pitfalls include over-automation (automating exceptions poorly, leading to bad decisions or more manual fixes), integration failures with legacy systems causing downtime, cybersecurity vulnerabilities from increased data flows, and underestimating the need for clean data/upskilling. The biggest risk? Rushing big-bang implementations instead of gradual layering, better to build incrementally, test rigorously, and keep humans in the loop for judgment.

Automation Will Continue, Just Not All At Once

Automation will absolutely reshape logistics, but it will not happen through one dramatic breakthrough. It will happen the way it always has, gradually.

A new sensor here, a better decision-support system there, more integration between vehicles, operations, and data. Layer by layer, the routine work becomes automated.

The important work moves closer to the people who understand the system best.

Curious how others in the industry are seeing this play out. Where has automation genuinely improved logistics operations? And where has it made things more complicated?

To order food, I had to scan a QR code on the table.

Order on my phone.

Pick up the food myself.

And then pay through the same QR code.

At one point, I was even prompted to tip 18–20% before any human interaction had happened.

Everything was automated. Efficient. Optimised.

From the business side, that makes complete sense: lower labour costs, faster throughput, better margins.

But something struck me.

The process had been automated, but the experience hadn’t been redesigned.

Subscribe now

In the past, a waiter might notice you look exhausted after hours of travel, ask where you’re flying, and suggest something lighter and energising before a long-haul flight. A small moment, but surprisingly helpful when you’re running on fumes.

That context just disappeared.

Automation removed the human layer, but nothing replaced the value that layer used to create: empathy, intuition, and genuine personalisation.

And that’s where I think many organisations are right now.

We automate the existing process instead of asking the harder question:

If automation and AI are available, what should the experience actually look like now?

Imagine the same restaurant simply scanning a boarding pass (or linking to flight info via app) and instantly knowing:

• I’m about to take a long-haul flight

• I’ve probably been travelling for hours

• Something light and energising might be better than a heavy meal

Automation could enable better personalisation, not just fewer staff. It could surface recommendations more tailored than most humans could on the spot, while freeing people for higher-touch moments.

Automation itself isn’t the issue.

Automating a process without redesigning the experience is

We end up with efficient but soulless interactions that are profitable in the short term but loyalty-eroding in the long term.

Curious where others have seen this play out, automation improving efficiency but quietly making the experience worse (or, better yet, examples where redesign turned it around)? Drop your stories below, I read every one.

Leave a comment

#Automation #AI #CustomerExperience #Leadership #Travel