Reselling contracts to building trust

With agents’ abundance all around us. I sense a cycle repeating itself.

Back when digital software became ubiquitous in our lives and all businesses were being reimagined using it, brokerages in logistics sprung up. I see a similar pattern of businesses being revisited with agentic AI as the new platform.

In logistics, we try to stick to the same frame. So, the focus is always to increase efficiency within the same frame. Using analogical crutch, we are on the lookout for a faster horse not an automobile.

Henry Ford would have been frustrated with logistics people, so are many innovators who are trying to pitch novel frames with agents in logistics and transportation.

First, lets go back to the previous era and talk about digital brokerages in the transportation space. None survived in their intended approach and slowly pivoted away or sold themselves. I think digital brokers are coming back into the play because AI agents do labour, working on the buying and selling of freight is now scalable with general intelligence dropping its costs rapidly. So essentially all brokerages are digital by extension.

Yet, the outcome will still be the same, unsustainable businesses. They will gain peak euphoria before the ground shifts underneath them. So, tools in the market looking to displace labour in operations of brokerage with AI are just transient in nature.

In the previous era, the cost invested wasn't worth the squeeze with how little they made in margins. With Agents, unit economics won't be the problem. It would be oblivion due to shifting landscape.

Going back to 3Cs concept is committing to a shipper or cargo giver that they are competent to follow due process while having the capability of finding the right capacity.

Most services in logistics are compound of 3 fundamental components. We can break any type of physical service in to these components.

For a Carrier they are providing access to freight that wouldn't have been available otherwise.

One of the realisations I have had over the past year is that aggregation of carriers is not what a brokerage offers. They offered trust to both counterparts prior to moving freight. And that trust is being tested by recent regulations starting the with Montgomery case in United States and followed by Dallas Court case filed that held the brokerage liable for negligent hiring. From the post;

If you’re a broker, ...........: the more involved you get in directing a driver on a specific load, pushing them to keep going, communicating directly about schedule and health, exercising the kind of control an employer would, the more exposure you’re creating for yourself under a borrowed-employee theory, independent of the underlying negligent-hiring question. Staying a broker, in the legal sense, means staying out of the driver’s seat.

If you’re a shipper, Home Depot is genuinely good news, but only for the passive role ; pick a properly authorized, safety-rated carrier and let it do its job, and you’re on solid ground. That protection doesn’t extend to shippers who get more hands-on with how a load actually moves.

So, if a broker that came into their stride during the last decade. Thanks to digital technology, they were able to mass manage shipments and provide economies of scale via price benefit to shippers. Information asymmetry with working capital helped them become prevalent in US freight. This led to different incentives where brokers were willing to work with any carrier, as long as they were qualified by the industry standards. The price was constant and not the carrier. More reliance on technology that provided vetting and more targeted the fraud activity becomes.

But the rulings are now incentivising different behaviour. Focus on who you work with and be hands-off while executing. Brokers then have to re-invent themselves and prioritise different facets on both the customer and carrier side.

While carriers themselves can leverage agents and AI to be accessible to Shippers. In the 3C's, carriers are constrained by capacity and capability, competency is on hire from agents.

What if it takes 30 documents and 4 verification licenses to work with a shipper. The carrier can task an AI agent to work through the process. On the other hand, shippers can engage 10x counterparts because agents are out there continuously onboarding, negotiating and transacting with carriers. Yes, there are 100 factors to check and match making is not easy. Agents can make this work when factors are defined accurately, also easy to delegate.

Brokers are prevalent because search friction optimisation doesn't benefit the participants in decentralised markets. Now, agents can represent shippers and carriers on their respective behalf and find one another given the access to digital platforms(internet being the biggest open platform). Carriers can build their own track and trace portal that communicates their ETAs to shipper agents who are monitoring without interfering.

Information asymmetry helped the broker maintain a moat. When information is much more accessible because humans are not seeking it, the market is going to shift. I am curious to see when the demand for such platforms, which help all counterparts by being on them, will pick up.

While we can all keep ourselves occupied by trying to build a faster and leaner execution model of brokerages with AI. There is a platform shift in the flanks planning to go mainstream.

And no, brokers will still exist as trust needs to be built between two counterparts. You need a mediator whose best interest is for a transaction to happen. Capability expansion would be an option that brokers should be exploring before they either get bought or written off. Until then consolidation will continue where technology and information asymmetry were the differentiators.

Round up

Moats in Logistics inspired by The Moat Map

This relationship between the differentiation of the supplier base and the degree of externalization of the network effect forms a map of effective moats

This post from Ben Thompson in Stratechery has gotten me to think about moats that we could build in technological companies in logistics and transporation space.

A tweet from Nikita , who ran Twitter's product until recently, posted that technology no longer a moat.

It got me thinking that technology was never a moat in logistics and transportation. Instead, it was the theory behind the product and technology. It led to building a moat with brand and process power, if you follow the 7 powers framework by Hamilton Helmer.

AI doesn't solve the need for theory, it just shortens the duration between origination to building moat through exercising the power.

So, how does one's technology help suppliers or sellers differentiate while buyers benefit due to network effects forms the basis of theory.

Links that resonated

Collecting Favours

I don't want to spoil this post. Please read it and enjoy one of the smallest pleasures of being human.

Cheaters and Whistle blowers in the agent swarm

Deepmind ran an experiment with 100 agents to solve a set of mathematical equations and were able to reproduce similar results to the Hugging face incident. They cite Elinor Ostrom's work as an inspiration to learn from. I concur that her ideas are one of the best representation of preserving common resources among collectives.

Sign off

I keep coming back to the same topic of AI and the unravelling that will or is happening in the spaces I inhabit. The cadence of weekly writing is more challenging because I have 3 or more slightly tangential ideas that I want to write about but end up with the one that was easier to finish.

In an age where AI slop is all around us, my solution is to write more in public. I will be posting my posts on the personal site and try to stitch them up coherently over here.

I hope there is no doubt that all of these opinions are my own and don't represent my employer.

Signing off till next time,

Vivek, back on the road meeting brokerages.