The End of Human-Heavy Forwarding: Why AI Will Hollow Out the Middle
If warehouses can run without humans, freight forwarding companies are next. The question is not if, but how badly companies will botch the transition.
The automation is coming. Anyone denying it is delusional.
If the industry managed to build human-free warehouses, then what stops it from building human-light forwarding companies? The answer is nothing. The technology exists. The economics make sense. The only question is whether companies will implement it intelligently or turn it into another cost-cutting disaster.
Anthony Miller at Wiser Logtech recently discussed this online and his conclusion was blunt. Anyone who thinks this is not coming is in for a shock.
The right approach should be replacing outsourcing, enhancing in-house teams, automating everything possible, and having people work from exceptions. What the industry will likely see instead is immediate cost cutting, big decisions made without proper planning, higher risk, and apologies when it does not work out.
The pattern is predictable. The consequences will be severe.
The Warehouse Precedent
Brian Newman, who works in supply chain transformation, stated the obvious. “It’s coming. If you can automate a warehouse you can automate the chain.”
Miller agreed. “You can automate any process that is highly repetitive and data driven, whether the data is deterministic or not.”
The warehouse automation wave proved the concept. Facilities now operate with minimal human intervention. Robots pick, pack, and move inventory. Systems route orders and manage exceptions. Humans handle edge cases and maintenance.
Freight forwarding involves similar work patterns. Repetitive data processing. Document verification. Status updates. Exception management. Rate quotes. Booking confirmations. Most of this work follows predictable patterns that AI can handle.
The technology is ready. The question is how companies will deploy it.
The Big Tech Model
Miller noted what has happened to big tech companies and their headcount. It feels like just a matter of time before sizeable forwarders start doing the same.
The parallel is direct. Tech companies overhired during growth periods. When efficiency became the priority, they cut deeply. Meta eliminated thousands of positions. Amazon reduced headcount. Google streamlined operations. Microsoft trimmed teams.
These companies had strong technology foundations. They still cut aggressively. Freight forwarders with weaker technology will face more pressure to reduce costs through automation.
Paul Claydon at a supply chain technology firm agreed with the assessment. Miller responded with what everyone is thinking. “I’m just waiting to see how the big 3PLs justify huge headcount cuts.”
The justification will be simple. Automation enables efficiency. Market conditions require cost reduction. Shareholders demand profitability. The cuts will come.
The Human-Light Reality
Mark Woolnough at a freight operations and recruitment firm made an important distinction. “Human free and human light are very different. Also human light is probably the accurate terminology.”
He pointed to Notion’s new agent updates as an eye opener. “Obviously not industry specific but if used by someone internally with the right mindset...”
The future is not human-free. The future is human-light. Small teams managing automated systems. People handling exceptions rather than routine work. Expertise concentrated in complex problem-solving rather than distributed across transactional tasks.
Peter Creeden, a global supply chain executive, emphasized that the shift is not just about automation but redesigning the forwarding business model around people and technology, not just cutting headcount.
He challenged the sector to rethink resilience. “It’s not just risk management but building capability, connectivity, and accountability. Human-light shouldn’t hollow out the workforce but empower skilled people to work with smarter systems and AI, improving exception handling, compliance, and trust, not cutting costs.”
Miller agreed but added a dose of realism. “But we both already know how this is going to play out. Hard lessons will be learned.”
The Implementation Problem
Stewart B. at Ziegler Group referenced a post about using AI to remove problems in supply chain as opposed to using it to figure out how to do workarounds quicker. “Both need deep planning, a progressive and adaptable strategy and excellent execution rooted in change management, which relies heavily on experience of the business. There’s lots of AI stuff flooding the industry, lots looks a bit clunky and gimmicky, but in the mid or long term, someone is going to start getting this stuff right.”
Miller’s response captured the industry’s frustration. “How to do the workarounds quicker. I swear, the amount of times I heard workarounds from forwarders using CargoWise. Hate it.”
He continued: “This comment needs to be turned into a standalone post. It is 100% true. But making the AI work in the right way will be hard. We’ve seen RTTVP fail. We’ve seen Blockchain fail. We’ve seen digital freight forwarding fail. I believe that we cannot afford to see AI fail, so it will succeed. I just hope that it will succeed in the right way, and not become another round of workarounds.”
The history of technology implementation in logistics is littered with failures. Real-time transit visibility promised transformation but delivered marginal value. Blockchain was supposed to revolutionize supply chain transparency but became a solution searching for a problem. Digital freight forwarding disrupted nothing.
AI cannot afford to follow the same path. The industry needs it to work.
The Foundation Problem
Francine Nielander at Lean Six Sigma consulting identified the critical prerequisite. “If you have your processes and data structured and you can automate or even optimize using AI, great. But those are not the companies that are in trouble and are likely to go for a quick AI fix. It’s not going to help you if you don’t have your ducks in a row. It will make processes even less transparent. Why was this order placed? I don’t know, the AI did it. Good luck in making sense of that.”
Miller agreed. “There is no quick fix in our industry and anyone suggesting otherwise needs to go do something else.”
Samil Shah raised the question directly. “Traditionally, getting value out of software required a good data foundation which in turn relied on good, consistent processes. With Agentic AI, can businesses skip the foundational work and jump straight to value?”
Miller’s answer was one word. “No.”
This is the trap waiting for companies desperate for cost savings. AI built on bad processes and dirty data will amplify problems, not solve them. Companies without solid foundations will implement AI and create expensive chaos.
Jennifer Morris at Ship Happens warned about the all-or-nothing approach. “I think the issue is many companies are just diving in head first and making it all or nothing. I have a feeling some of these legacy companies that were kind of flailing already and have turned to AI so they can make major cuts, not mentioning any names at all, are going to have problems. I think a more strategic approach and implementation will be what truly works. Not slap on AI, they barely understand, and hope for the best.”
Miller identified specific companies to watch. “I’m looking forward to seeing what DSV and Maersk do. I’m concerned about the likes of K+N and Ceva/Bolloré. Also quietly excited about some smaller players like Geodis. We’ll see how it plays out, but yes, those who are diving in head first are in trouble. No point getting first mover advantage as an LSP. It’s about long term vision.”
The Human Element
Nacho Gil de Sagredo, who specializes in digital and sustainable supply chains, raised an important constraint. “I think this would work with human-free BCOs. While clients have people in procurement there’ll be people in sales and operations on the LSP side. I think this is why digital forwarding has not exploded yet. Humans like humans when there’s risk involved.”
The observation points to a fundamental limitation. As long as customers have human procurement teams, they will want human contact on the supplier side. Automation can handle routine transactions, but relationship management still requires people.
This creates the human-light model rather than human-free. Small teams managing customer relationships while AI handles operational execution.
John Vonk at Seeburger Benelux agreed. “Totally agree that human-light logistics is on the horizon. The challenge isn’t if we can do it, but how we do it responsibly. The companies that win will be those who blend automation with human judgment rather than swinging too hard on cost cutting.”
The Execution Details
Luca Conner at Pack’N provided practical guidance. “The sweet spot is automation for the routine and people for the weird stuff. The mistake is cutting seats before you can measure exceptions and promise kept. What’s worked for us is simple flows, clear owners, and tracking exception rate and cycle time before we pull labor out.”
He asked where the cleanest early wins are appearing: quotes, status pings, document checks, or invoice matching?
Miller responded that document checks is a tale as old as time with OCR solutions. It should improve greatly and remove the need for armies of people in low cost of labor countries doing manual validation. He identified customs and compliance as the biggest area to change. “It is a major drag on global supply chains.”
This practical approach, measuring exception rates and cycle times before removing headcount, represents responsible implementation. Most companies will skip this step.
Kenneth West identified the strategic shift. “Legacy tech stacks are hitting their ceiling. The shift from human-heavy to human-light isn’t just about cost, it’s about control. Agentic AI flips the model: instead of outsourcing complexity, companies can internalize intelligence. The ones who build exception-based workflows now will lead the next wave of operational resilience.”
Miller called out the buzzword density but acknowledged the underlying point. Companies that get this right will have operational advantages competitors cannot match.
The Reality Check
Ian Aguilar offered necessary skepticism. “I am a big fan of the current gen AI systems. Very useful, far more so than Google was historically. They do get things wrong frequently, though. Like, on the daily, and can be wrong in any given topic, even if you call them out on it multiple times. I’ve had both ChatGPT and Grok be wrong on simple addition, subtraction, multiplication even.”
His conclusion: “So far they are a nice to have tool, yet very far from a necessity. I absolutely could not trust business activities or decisions to it. Very interested to see where we are in 3-5 years though.”
Miller acknowledged the skepticism is healthy. But he also pointed to the rapid pace of improvement. What seems unreliable today may be dependable in months, not years.
Vlad Nikalayeu at Skypace identified the accountability problem. “You know what’s the main problem? It’s still people who’ll be eventually taking responsibility for what AI does.”
He imagined a future where LinkedIn profiles include AI performance metrics. Companies will want proof that candidates can work effectively with AI systems. The HR function transforms into validating human-AI collaboration capabilities.
The Nearshoring Alternative
Troels Daugaard noted that sizeable forwarders are using nearshoring instead to cut costs. Miller questioned why asset-light companies would want to nearshore.
The observation reveals competing cost-reduction strategies. Nearshoring moves work to lower-cost locations but maintains human headcount. AI reduces headcount but requires technology investment and change management.
Companies will pursue both. Nearshore operations will implement AI. The combination will drive costs down faster than either approach alone.
The Warning
If legacy ERP and SaaS players in logistics and supply chain are not using AI or planning AI functionalities, they are already late. The technology is moving faster than incumbent software providers can adapt.
If you are an LSP or BCO doing your own orchestration and you are already doing diligence or trialing solutions, you are ahead of the curve.
The gap between leaders and laggards will widen quickly. Companies that move now have time to learn and iterate. Companies that wait will face crisis implementations when competitive pressure becomes unbearable.
The human-light forwarding company is not a future possibility. It is a current reality being built by early movers. The question is not whether your company will follow. The question is whether you will move strategically or desperately.
Hard lessons are coming. Some companies will learn them in controlled experiments. Others will learn them through layoffs, customer losses, and operational failures.
The automation is here. The execution will separate winners from casualties.
Join the conversation on AI implementation in logistics, workforce transformation, and operational automation at Chain.NET, where supply chain professionals share real-world AI deployment experiences, debate implementation strategies, and connect at events focused on responsible automation. The future is being built now.
 | A guest post by
|