Episode 11

Automating ugly freight in an evolving digital world | Cate Hull, Freight Exchange

Freight Exchange founder and CEO Cate Hull takes us behind the scenes of "ugly freight"  — the process of moving goods that are difficult to transport due to size, weight, or the need for specialized equipment.  

She explains how her company leverages machine learning and other tools to overcome the many challenges of working at the intersection of the digital and physical worlds, in a field highly exposed to human error, fraud and theft.  

They also discuss

  • The unique data challenges that arise at the point where digital technologies interact with real-world systems like freight.  
  • How to manage risk in a system defined by human input and with a high likelihood of human error, theft, and fraud.  
  • The roles artificial intelligence and machine learning can play in automating shipping management.  
  • How the space needs to evolve in line with technological advances such as artificial intelligence.

Resources

🎧 FILED 06: Managing misinformation at the speed of AI | John Croll, Truescope

🎧 FILED S02:E4: Companies must focus on reducing risk, not just improving compliance | Dr Miles Ashcroft, RecordPoint

📨 FILED Newsletter: Generative AI will offer up identifiable data if you ask nicely

Transcript

Anthony Woodward

Welcome to FILED, a monthly conversation with those at the convergence of data privacy, data security, data regulations, records management, and governance. I'm Anthony Woodward, CEO of RecordPoint. And with me today is my cohost, Kris Brown, RecordPoint's VP of product management. How are you, Kris  

Kris Brown

I'm good, mate.

Kris Brown

How are you? It's a lovely morning here in Brisbane today. How's it where you are?  

Anthony Woodward

I'm in Seattle and it's a nice day in Seattle, in Seattle terms. And you were only here a few weeks ago. Anyway, you've been on the road.  

Kris Brown

Yeah. And it was very Seattle-ey when I was there too. We had a hot and cold and wet and windy and all of the things, but lovely town.

Kris Brown

I really enjoyed the time there and we went and saw some customers and said hi.  

Anthony Woodward

Today we've actually got a really interesting episode of FILED. There's quite a different direction and I've invited Cate Hull, who's the CEO of Freight Exchange, to talk about the real world. We in the convergence of data privacy and data governance, think a lot in bits and, you know, occasionally we talk about paper, but that's the extent of it.

Anthony Woodward

Cate has an awesome business called Freight Exchange, I'd love to introduce Cate, and introduce yourself.  

Cate Hull

Thanks, Anthony. Thanks, Kris, great to see you both. Yeah, my name is Cate, I'm the CEO and founder of Freight Exchange. Freight Exchange is a software enabled marketplace for what we would decide what we term ugly freight.

Cate Hull

I did not make up the term ugly freight. It generally refers to physical goods that need specialized equipment to move. So, if you think about explosives, you're in Seattle. So, I think guns, all of these things, cold chain, flat pack furniture, they need specialized equipment to move around. And so we are for businesses that trade in those goods and we make it easy to move those goods around.

Anthony Woodward

Really interesting industry. And there's a lot of parallels, I think, to what we talk about every day. I'd love to understand when you think about that supply chain, what the data is that's flowing around, what are the security elements, you know, there must be a whole bunch of stuff happening in Freight Exchange

Cate Hull

Absolutely. It is very much a... The data sources are really diverse and generally don't fit comfortably into what we would call an old school data format. So, for example, paper, there's a lot of paper, there are a lot of labels, there are a lot of, obviously, physical goods but at the same time, in terms, when you talk about data, it specifically maps to things like free text, unstructured data, as well as all of the usual trade documentation, so what is it, who is it going from and to, and so on.

Cate Hull

So, yeah, the mapping between online and offline is, is a really interesting one in our business.  

Anthony Woodward

And do you handle from point to point, like what level do you go to crossing borders, all the kinds of issues where what's the scale of this?  

Cate Hull

In the industry, it is enormous because you are dealing with basically moving everything everywhere.

Cate Hull

And so, you think about multiple languages, you think about currencies, you think about customs regulations, border controls, and then trade terms. So, all of those things come into the story. For us as a platform, we're not doing Uzbekistan at this point and trade terms there, but that's sort of the scale of the problem in our industry that we're tackling.

Anthony Woodward

As you think about all of these goods that you're thinking about moving and the amount of stuff that's going on, what's the level of expected error? How often does it go wrong and how do you kind of deal with that and govern those kind of mistakes and problems?  

Cate Hull

Yeah, it's such a good question because basically it is all about managing by exception.

Cate Hull

The whole industry is managing by exception. It's somewhat miraculous that it works as well as it does, given the level of digitization and automation that's in the industry. It's certainly quite far behind a lot of other industries. Yeah, it is challenging. And in terms of the risks, you know, you have things like, because everything is handled by humans, you've got human truck drivers, you've got human people packing the trucks, you've got humans booking the freight, you've got humans pricing everything, it becomes very challenging to manage those risks.

Cate Hull

So, the physical risks might include theft or fraud, or error is usually the big one. Obviously, you can drop things and break them and all of those things. Managing those by exception is really the name of the game.  

Anthony Woodward

It's really interesting in our world. There's a lot of talk about creating data exchanges.

Anthony Woodward

So being able to bring different subsets of data that different companies can share, or different public sector agencies can share it is kind of notion of a data mart. And I think there's a real analogy to the types of problems that you see in the real world of delivering goods between different suppliers and different expectations to these data marts.

Anthony Woodward

I mean, how do you handle the contracts and the governance of those processes so that there is an equilibrium in them? Like, what does that look like for you?  

Cate Hull

It looks like hard work for us, but as a software platform, that's what we're designed to do. So, those contracts are very dense and they're down to the minutiae as well.

Cate Hull

So, for example, generally you'll see contract terms that will be, if your item is more than 1.2 meters in length, we will charge you an extra 75 or if my driver stays, you know, has to unload for more than 15 minutes, we'll charge you 60 in increments of 60, for example. And so, yeah, it becomes that the contract terms are, that's basically why we exist is to decode those contract terms, make them fairly intuitive to understand, mainly for people in the industry, not generally for the lay person, but we do really design it as best we can for the lay person, because we figure if the layperson can understand it, then pretty much anyone can.

Anthony Woodward

In your platform, are you using some machine learning and AI to predict when these types of risks are happening and start to manage that within the supply chain?  

Cate Hull

So, the machine learning applications for the data in our platform or for our industry are numerous. So just a couple of examples.

Cate Hull

Is it going to arrive on time or in actual fact, do we think this is off track? That's the most basic prediction that we make in the platform. So, we said it was going to be there in three days. It's now two days, but actually, you know, we think that something's gone and gone awry. And so, it's about decoding the data signals that we get from the both the carriers and the supply.

Cate Hull

We call them carriers and shippers. It's about decoding those pieces of information and then using those to proactively alert people that there might be something going wrong. We don't detect, well, whether the reason for it going wrong, although certainly we get good signals in terms of why it's going well, why things might be going awry, but we can certainly tell with a fair degree of accuracy when we need to look into something.

Anthony Woodward

There's been a lot of talk in the technology world about trying to combine things like blockchain and other processes to look at two sided contracts and those things. Do you expand your technology into that realm, or have you not gone that far?  

Cate Hull

I think there is very much a use case for blockchain and particularly in relation to the contracts element of trade.

Cate Hull

In the industry, there's this concept where you basically don't. When goods change hands, money should change hands. And so, there's a fairly well defined use case for blockchain in that regard, where you're basically saying, I can, I know that the goods have changed hands because Freight Exchange has given me the data and the signatures to prove that it has, they've codified the contracts.

mm

Cate Hull

There is a really great use case for that. We don't do it at the present moment, frankly, as an industry, just simply digitizing is a really good first step.  

Anthony Woodward

The reason I asked that question is, again, really strong parallels with our world. Kris and I often speak to people about blockchain and when blockchain is going to be used around storing data and how it would describe data and even the data mark contracts of data, but it's the same problem.

Anthony Woodward

In fact, a lot of the data is not digitized to the point where it's well described. So, it may well be “born digital”, but it's not well described enough to put into a blockchain and then act on it. I mean, that's common theme, isn't it, Kris Yeah, absolutely.  

Kris Brown

I think the interesting thing for me was to hear Cate's response there and that it's, you know, we're all still trucking along.

Kris Brown

See my pun there. Very, very good one, but we're all still moving towards this goal of better described data, better control of that data. And in your world, the better control of that data leads to better control of the ugly packages. Now, again, a term that I learned today.

Anthony Woodward

Do you see the machine learning coming to a point where it's able to predict the... as you know, I'm sitting here in Seattle and where I was going with that question was Amazon right now caught up with a friend of mine for lunch predicting what needs to be in the warehouse.

Anthony Woodward

Around two to three days before the buyer’s acquiring it. Are we seeing other companies starting to invest in the same processes where they're putting things into a depo, you know, in somewhere like Australia, it's a long way from everywhere else, knowing that within four or five days, it's going to be consumed and picked up by a company like yours.

Anthony Woodward

And are you offering some of that capability to people?  

Cate HullYeah, definitely. In terms of the supply chain, the concept of bottlenecks and just in time supply chain, which is hopefully a fairly well understood concept is what you're referring to there is called demand sensing, and the better you can get at predicting what needs to be at its final destination and when.

Cate Hull

Then the better you can plan everything upstream and absolutely it's really an important part of what we do because if we don't do that well you see what happens in COVID when suddenly no one can get certain things because we just hadn't sensed that the items weren't where they needed to be at the right time and in the right place if that makes sense.

Kris Brown

The interesting thing there for me was, I was living in the UK just as Brexit hit. Being a part of the European Union, the just-in-time supply chains were, you know, incredibly complex and all of those border elements were removed. And of course, there were, you know, car factories that were relying on, as they're building vehicles, literally relying on just-in-time deployment of parts to that supply chain line.

Kris Brown

And then the day after Brexit, all of a sudden, all the trucks were stuck. At the Channel waiting to be checked, whereas previously they would just drive on through and a whole series of things fell apart behind that. How does that affect your platform? So, you know, obviously there's demand sensing element of that, but regulations change because, you know, you are moving across those borders.

Kris Brown

How is that built in?  

Cate Hull

Probably a good example of how. We are affected by some of those unexpected bottlenecks or issues in the supply chain. For example, in the last couple of years in Australia, we've had enormous disruption from a weather perspective, severe bushfires or wildfires and then severe flooding.

Cate Hull

And what happens when there are severe bushfires or floods is roads close and nothing can get through. And... That is a common experience in our world. And so, the way we deal with that is, you know, we are able to ingest information. You know, we know that there's, there's disruptions and we can, you know, essentially automate that.

Cate Hull

So basically, saying either this isn't going to get through, but we predict that it will in X days or X weeks. It's essentially again, just getting smarter and smarter again, using machine learning and risk management techniques to work out how best to mitigate some of those issues. Sometimes all you can do is say, it's not getting there.

Cate Hull

You need to dispatch another one. And that's never, never a good news story for anyone. But basically, it's great to know what's going to happen, but what you do as a result of that is the, is probably the hard part.  

Anthony Woodward

I was catching up with another friend who works for Cruise, who are the self-driving cars here in the U.S., and they've just had a little bit of a regulatory change, having lost their license in California, but that's broken other markets. Clearly, with the whole notion of self-driving and the ability for autonomous delivery is coming, how far off are we? Because we're starting to see some real challenges in that space, aren't we?

Cate Hull

Oh, enormously so. And it's not just self-driving, it's robots everywhere. So, the automation of manufacturing, the automation of warehouse management, and then the automation of transportation is basically, in five to 10 years, I can see there will be a much bigger uptake of completely automated supply chains.

Cate Hull

Regulation is certainly one part of that. That is the future, and it is going to happen. Ironically, when everything's automated, it's much easier to manage disruption.  

Anthony Woodward

Which actually was my next question was, you've now got a whole bunch of information that's flowing through the machine. How's that actually handled in the industry?

Anthony Woodward

What are the thoughts around that? How do you deal with the data privacy of, of what's flowing through? There's some really deep information going out through those systems. That's pretty public.  

Cate Hull

Absolutely. It's really challenging today. There's a lot of very public information and again, we deal with this every day is if, you know, if something... We don't have contracts with Apple, but as an example, if you were shipping something sensitive like drugs or high value goods, it's really important to try and mitigate the risks of the physical movement in the industry. And that's, you know, that's by managing people and processes. And again, of course, managing, locking down as much information and physical, you know, and trucks and warehouses as you can. Yeah. In the future, it is going to get more and more complex as more and more devices coming into contact with those things.

Cate Hull

It's really, oh, my goodness. It's a, it is a huge data problem.  

Kris Brown

How are those challenges being addressed from an industry perspective? And I know I'm probably moving outside of, you know, Freight Exchange itself, but what are the challenges you see there as it relates to that transition for the human workers to a much more digital environment?

Kris Brown

You will have gone through some of it already, but here's the background. I have a family member that is an owner operator, excavator, for example, very, very talented at what they do and has been very, very successful, but to this day still operates out of a docket book every day for all of their safety checks, for all of their time allowances and everything else.

Kris Brown

And I've recently helped them to move to an app and go, I'm cool. We're going to, we're going to digitize this because you keep forgetting that you've run out of the docket book. And so, you don't get paid for a particular day because you didn't do your safety checks and you had to go and get the docket books produced and whatever else.

Kris Brown

It's not a lack of intelligence. It's just, it's not a generational thing that they've grown up with. And so, this industry, and I specifically now mean the logistics industry, as you said, 21st century.

Kris Brown

What are the challenges there around that? And then obviously the fraud, the security, and again, even their awareness of the data privacy element.  

Cate Hull

Actually, a really interesting and very exciting part of technology. And it's, it is actually something that we think about all the time, because what we're talking about is digitizing an industry, which is ancient and global and is inherently human people.

Cate Hull

People run supply chains. And so, in terms of how, how we evolve the industry, the logistics industry is with technology and data is really about how do we evolve human interfaces? So how do you make it really easy for your friend to move online or, you know, into digital solutions? So, you're seeing now, or even years ago, really large warehouses were deploying drones and visual techniques to do stock take.

Cate Hull

So basically, you've got three or four storeys worth of goods in a warehouse and the drones fly up and down and do the stock take, you know, reading the bar codes and visually sighting everything. You know, the people who are doing the packing of the goods in that warehouse are wearing headphones and goggles and or, you know, even glasses to direct them with all augmented or, virtual reality to, you know, go and find their way around the warehouse really rapidly. And so, it really becomes more of a question about how can we evolve our solutions to be really helpful and useful to the person. To the end user. And right now, I don't know anyone who loves filling in a form. Like no one wants to do that.

Cate Hull

And so, for us, it's really about focusing on how we can actually use, how we can build more human centered technologies into the platform so that we, you know, essentially digitizing and making it just ridiculously easy for the people who, who have to keep, keep us fed and clothed. So.  

Kris Brown

I'll just let it, let it know it's coming.

Anthony Woodward

See, I like filling in forms. So that, that, that's always fun.  

Kris Brown

You are one of three on this call, mate, just so we're clear.  

Anthony Woodward

That's pervasive, not just across this industry, but all industries. I mean, in reality, Kris the, the things that we do at RecordPoint every day is trying to solve that same challenge, just in a slightly different paradigm.

Anthony Woodward

How do we make it more effective and easier to access the data? How do we make it more effective to find that data? Can we do the AI, I think Kris is that possible. Can we get those glasses?  

Kris Brown

Yeah, absolutely. I'm already visualizing a whole minority report thing about, you know, the FOI request has come in and now it just appears around me.

Kris Brown

Did you just approve my expense for a bunch of the new Apple glasses? Is that what I heard also? Let's do it.  

Anthony Woodward

Yeah, let's get it. Yeah. All right, cool. That's what I heard.  

Cate Hull

We'll have them delivered tomorrow.

Anthony Woodward

We'll organize that with Cate. Kris sort of touched on this in the question around regulations. Is there a pace of change of regulation that you're seeing in your industry? One of the things that we really have sweeping through the privacy sector and the data sector is there is constant regulation.

Anthony Woodward

Like for instance, and I imagine it must impact your world somewhat, the White House a couple of days ago put out a new piece of legislation around AI and managing models and thinking about models. How does that cross into your world?  

Cate Hull

Regulation is inherently part of the, our world. So, if you think about regulations around who can drive a vehicle, and for how long, and how heavy can that vehicle be?

Cate Hull

So those regulations impact us significantly. So, every government has, has rules about how they trade. Again, who can operate a heavy vehicle, for example. And so, we have, we build that into the platform wherever, where, when we're aware of it. And so, it's not a new concept for us. You know, trade regulations are again, ancient and global.

Cate Hull

And so that information is part of what we do. What is really improving is not so much, you know, the ability to know what's in there, so what the regulations are, but it's how you respond to them and, and sense them and make sure that you're compliant, is the challenge. There are other really weird challenges in our world.

Cate Hull

So, for example, have you ever heard of stink bugs?  

Anthony Woodward

Oh yeah. All right. We all grew up in the Queensland area,  

right?  

Cate Hull

We all know, we've all had our stink bugs, right? But this is a huge global problem. And so, one of the things that if we talk about biosecurity, so the ability for us to, and for regulators to know that what is actually being transported is safe from a biosecurity problem is also a really, really big challenge in the industry. So again, it's just every part of what we do, there are rules because everything has to be fumigated and tested and checked that there are no stink bugs in that container. However, how do you do that in a way that's, you know, automated, secure, and, you know, basically finds the stink bugs.

Cate Hull

There's so much work that has to go into the technology to make that happen.  

Anthony Woodward

I never thought we'd get to stink bugs on the podcast, but, but there is a first.  

Cate HullYeah. Happy to help.  

Kris Brown

The interesting thing, Cate is to understand, you know, are there any other areas? I know you've touched on a couple, but are there any other areas where ML, you know, you can foresee it being useful?

Kris Brown

In your industry, so obviously those data challenges, you're going to be getting more and more data. As you said, there's going to be these elements where you're going to be trying to predict more and more things. Where are these areas that you think are the next big things in your industry?  

Cate Hull

There's been a lot of talk about large language models, LLMs. I think those are incredibly helpful when it comes to vision and using machine learning for, again, the stink bug example is a really great one. If you've got a camera, it can see a stink bug. Then, you know, there's, you know, so ML is going to become, you know, more and more sophisticated around those types of use cases, physical machine learning, the algorithms around optimization. So, we've talked, we've touched on how much data there is once you start to basically ingest the whole of the real world, and then you've got to predict what's going to happen.

Cate Hull

The challenge isn't so much... because in the end, all you're predicting is how much is it going to cost? When's it going to get there and is anything going to go wrong? And if something's gone wrong, what do we do about it? That's all you're ever trying to predict in our industry, how you do that and how well you do that, that and how well the algorithms can actually do that work is.

Cate Hull

Probably the area that's going to be most challenging at some point you know, the, the computers are going to need to get way faster than even what we're using today.  

Kris Brown

Yeah. I think it's a really interesting use case. Like I've started to think of other data sets that could help to feed in. So, you know, even protests on, you know, there was, I think there was the in Canberra, the rigs coming through town on their annual union jaunt, which means that either be a number of vehicles off road, but there would also be blockages on roads and other things. I know that there's in DC, in the US, they have an annual bike ride that basically blocks Route 66 coming into DC is hundreds of thousands of motorcyclists just pour, they call it rolling thunder. It's an incredible thing to see, but it's, they pour down Route 66 and yeah, the vehicles just can't enter for, for hours at a time. So those sorts of, you know, again, more community-based events that are sort of would help feed into even, you know, and we're talking about edge cases here in terms of the accuracy, but as you say, it's, it's all of those data sources that they can start to feed in and bleed that information out.

Cate Hull

Yep. Weather, traffic, you name it. It's, you know all of the cameras, all of, yeah, there is just an enormous data set and it's how, how well can you use that information to make smarter decisions.  

Anthony Woodward

Be able to step back and have a look at something physical, like freighting, you know, the ugly things around us, as you put it and understanding the flows of data that accompany these real things.

Anthony Woodward

It's quite, I think I may have been saying it to you at some point Cate, over a wine at some over a year ago, but I often feel like our world in this data world is actually a reflection of what you're really doing, what we do is shift data around and put it in safe places and lock it down and make sure that it's doing the things it's supposed to do and protecting the innocent, but you actually do it rather than just the bits.

Anthony Woodward

I really like the notion of what you've talked about there of that analogy of how the freight industry works and how those pieces come together. So, thanks for spending some time with us. I really appreciate it.  

Cate Hull

No worries. Got a bit technical, but I can talk about data and machine learning until the cows come home.

Cate Hull

So, I enjoyed it.  

Anthony Woodward

Thank you all for listening. I'm Anthony.

Kris Brown

And I'm Kris Brown. We'll see you next time on FILED.

Enjoying the podcast?

Subscribe to FILED Newsletter.  
Your monthly round-up of the latest news and views at the intersection of data privacy, data security, and governance.
Subscribe Now

We want to hear from you! 

Do you have a burning topic you'd love to hear discussed?
Submit your topic idea now to help shape the conversation.
Submit your Topic