Sustainable Supply Chain

Streamlining Customs: The Power of AI and ML in Simplifying Supply Chain Processes

February 27, 2023 Tom Raftery / Oscar Morales Season 1 Episode 296
Sustainable Supply Chain
Streamlining Customs: The Power of AI and ML in Simplifying Supply Chain Processes
Digital Supply Chain +
Become a supporter of the show!
Starting at $3/month
Support
Show Notes Transcript

Hey folks, I wanted to share with you today's episode of the Digital Supply Chain podcast, where I had the opportunity to chat with Oscar Morales, the co-founder of Sifty. Sifty is a company that's revolutionizing the logistics industry by using AI to help customs brokers increase their efficiency and save time on operations.

In our conversation, Oscar and I delve into the current state of the logistics industry and where it's headed in the future. With AI making headlines, Oscar shares his thoughts on why the logistics industry should be excited instead of nervous about the advancements in technology.

We also discuss the importance of data in the logistics industry and how Sifty is using AI to curate this data and provide better tools for their customers. Oscar also shares his vision for Sifty to become the operating system of the logistics industry, connecting all participants in the supply chain with a cloud of services.

We also talked about the ethics of AI and how Sifty is making sure their output is accurate and unbiased. Oscar emphasized the need for a human element in the efficiency of AI and how Sifty is working with their customers to achieve this.

Finally, Oscar shared some key takeaways about Sifty and what sets them apart in the industry. Their passion for technology and solving operational bottlenecks with creativity, their commitment to growing with their customers and increasing revenue, and their dedication to keeping the information they receive and provide safe and ethical.

If you want to learn more about Oscar and Sifty, head over to their website at siftyml.com or connect with them on LinkedIn. And if you're interested in staying up-to-date with the latest news and resources from Sifty, sign up for their monthly newsletter.

That's it for today's episode! I hope you enjoyed our conversation with Oscar and learned something new about the logistics industry and AI. Stay tuned for more episodes of the Digital Supply Chain podcast.

Support the show


Podcast supporters
I'd like to sincerely thank this podcast's generous supporters:

  • Lorcan Sheehan
  • Krishna Kumar
  • Christophe Kottelat
  • Olivier Brusle
  • Alicia Farag
  • Joël VANDI
  • Luis Olavarria
  • Alvaro Aguilar

And remember you too can Support the Podcast - it is really easy and hugely important as it will enable me to continue to create more excellent Digital Supply Chain episodes like this one.

Podcast Sponsorship Opportunities:
If you/your organisation is interested in sponsoring this podcast - I have several options available. Let's talk!

Finally
If you have any comments/suggestions or questions for the podcast - feel free to just send me a direct message on Twitter/LinkedIn.

If you liked this show, please don't forget to rate and/or review it. It makes a big difference to help new people discover it.

Thanks for listening.

Oscar Morales:

Customs and customs brokers have not received technology or advanced technology for many years and is such a fragile point in the whole supply chain. That is single mistake in a number, a digit, a phrase, can cause a disruption and alter your efficiencies. So in, in essence, all the efficiencies invested in a supply chain can be lost with a very minor mistake, or an absence of information

Tom Raftery:

Good morning, good afternoon, or good evening, wherever you are in the world. This is the Digital Supply Chain podcast, the number one podcast focusing on the digitization of supply chain, and I'm your host, Tom Raftery. Hi everyone. And welcome to episode 296 of the Digital Supply Chain podcast. My name is Tom Raftery, and I'm thrilled to be here with you today sharing the latest insights and trends in supply chain. Before we dive into today's show. I want to take a moment to express my gratitude to all of our amazing supporters. Your support has been instrumental in keeping this podcast going, and I'm truly grateful for each and every one of you. If you're not already a supporter, I'd like to encourage you to consider joining our community of like-minded individuals who are passionate about supply chain. By becoming supporter, you'll not only be helping me continue delivering high quality content, but you'd also be part of something truly special. Supporting this podcast is easy and affordable with options starting as low as just three euros. That's less than the cost of a cup of coffee and your support will make a huge difference in keeping the show going strong. To become a supporter. You just simply click on the support link in the show notes of this or any of the episodes or just go to tiny url.com/d S C pod. Now without further ado, I'd like to introduce my special guest today, Oscar. Oscar welcome to the podcast. Would you like to introduce yourself?

Oscar Morales:

Thank you, Tom. Yes, of course. I'm Mexican, I'm an entrepreneur. I wasn't into tech. I started my entrepreneurship career in logistics actually. And for the past six, seven years, I've been a techie kind of guy. I self-taught my, uh, to, to code and then I learn about machine learning and ai. I did a switch between traditional business logistics now into logistics and technology

Tom Raftery:

cool. And what are you doing in logistics and technology?

Oscar Morales:

Alongside Julio, my co-founder, we maybe five, six years ago, we started with an idea of designing an algorithm to go around to predict the risk of parcels moving across borders, and that was our first, attempt to create a product that was aimed for, the logistics industry. A bit later then we decided to move forward with that product and build it into a larger scale, and to, formally, create a company around it. So we aim of our products and software to make customs brokers, job a bit more efficient in several bottlenecks around their operations.

Tom Raftery:

Okay. Talk to me a little bit more about that, Oscar, because A, we still haven't named the company that you, that you founded and, and B. What kind of problems do customs brokers encounter that you're helping them get around? Because, you know, this is a, a podcast around supply chain. Sure. But we cover many aspects of supply chain. Everything from engineering, manufacturing, warehousing, you know, so for many of the people listening to this podcast, customs brokers might be something they've never heard of before. Or if it is, they have no idea what customs brokers do day to day.

Oscar Morales:

Yes, you're right Tom customs brokers and customs in particular, are a weird black box that many of us don't really understand what happens and, and what they are. I assume a portion of your audience will understand it better than myself even, but just to simplify or to try to put it in a wording that we will understand for the rest of this podcast. Customs as an area it is this part of, of any international trade operation that appears twice. So you have a customs area whenever there's an export. So whenever a thing, it's leaving a country and then you have an import customs area when, when the thing is going into another country. And in this area you have the government understanding what it is that you're trying to send off and then send in and translating all that information into, into things that they understand so that they can be correctly processed. And the processing could be anything from, perhaps this, uh, thing that you're trying to get into a country needs to be taxed or, you need to pay an excise on it, or it is regulated and you need a certain type of permission, or it's not allowed into the country, so it will have to be destroyed or sent back. Any of those. For many years, before the eighties, I will say companies were dealing with that by themselves because customs were not organized in any way, and the private sector will have to deal with these procedures and language that they wouldn't understand. After the creation of the World Trade Organization and, and, and, and several laws around, uh, international laws, around customs and how to operate better. Then a figure of customs brokers appear and, and their main task of these agencies, of these people is to sift through a lot of the data so that they can give to the government whatever is relevant and ease the process for the private sector. So our company Sifty, its aim is to simplify custom processes for customs brokers and in particular, we deal with low cognitive operations that are very repetitive. So one example is they have to transcribe commercial invoices as short or as long as they are into the system to create a customs declaration. And this is a process that they have to do every single time they're sending imports or an export operation. So you ma, you can imagine the volume of transactions and the amount of time required for any type of operation. If, if your invoice has perhaps two items, then it's fine. You might take five to 10 minutes. But if you're shipment, it's a container with 7,000 different items, and each one of these items have different HS codes and they have different values and they have a different certification required to get into a country, then that, task becomes a massive amount of time. And then you have teams of people working into getting, the, the shipment across the border. It is important to note that borders across countries are a natural evolution of international trade to create friction, and they exist for the reason of countries do not want to have a free flow of people and goods without any type of control, even if it's just for statistical analysis, right? We need to know what is crossing, who's crossing, how often it is. Now in overall, the supply chain of international trade for many years has received a ton of technology. Technology that has increased efficiency from the beginning to the end. And the beginning means when a company is planning to manufacturing in an advanced ERP system with, this AI controls to predict, the supply and demand. And towards the end of, the supply chain, you have last mile, delivery services with advanced algorithms as well to plan how best to divide the merchandise to be delivered on time and, and as fast as possible. However, customs and customs brokers have not received technology or advanced technology for many years and is such a fragile point in the whole supply chain. That is single mistake in a number, a digit, a phrase, can cause a disruption and alter your efficiencies. So in, in essence, all the efficiencies invested in a supply chain can be lost with a very minor mistake, or an absence of information.

Tom Raftery:

Okay, so taking this container example you had with 7,000 items within it all with different categories and sizes and weights and et cetera, et cetera. How does your solution help that?

Oscar Morales:

Sifty works with a person, works with the human. Our systems need this human element of, analysis to make it more efficient. Now, I'm gonna try to make a comparison of what happens without Sifty and what happens with Sifty. So when you don't have Sifty, when a broker doesn't have Sifty, they have to put in play a ton of people working together to get this container through customs. It means analyzing documents, transcribing information, making sure that the merchandise that it's in there, it is what it's on the documents, and then you have to spend many hours on getting that transaction done. Most of that time is a transcription part, and if you think about this container, perhaps it has a periodicity, right? It comes along every week or every two, three days. So every, every time it comes, you have to repeat the operation and again and again. And the transcription. It's a low cognitive task. it's copy and paste right from one system to another. The problem is, that sometimes this information comes on a PDF or an image that it's not very friendly to do a copy paste or sometimes the information is just too long to do a a quick transcription. What Sifty does is putting software and technology into the hands of customs clerks, or customs brokers. To make it more efficient, it's, uh, analyzing the data from the pdf, from an image, from the file, extracting the relevant data that they need to create a customs declaration, process with historical data, that the custom broker has or that we have, do a risk analysis on every single one of the items that are there. Compare it to the most recent past transactions so that we know what information we can rescue from it and prefill that customs declaration. All of that happens within three minutes, and in three minutes the custom clerk or the broker that is dealing with that information already has a document that is perhaps 70 to 80% complete to finish off in their customs declaration software for pre validation Now, if you are thinking of one transaction, perhaps you say, well, this is not very relevant. You're just saving time on one transaction. But the volume customs brokers are companies that make money on volume of transaction because they get paid a fixed fee for any transaction, whether it's one container, 10 containers, or one box. They have fixed fees. It's a very well regulated market. because there are so many hands within the supply chain, you might understand that it's not a lot of money, right? So, they have to work on a lot of volume to be able to make their rent some day. And that means that the more information they can put through Sifty, the more work they can do faster. And the idea of Sifty is not it's not replacing another person to do their job, it's giving them more time to do more and increasing their, revenue.

Tom Raftery:

Okay. So it's making them more efficient so they can process more items in the same amount of time. That's what you're saying

Oscar Morales:

Yeah. Yeah

Tom Raftery:

Okay. Nice, nice, nice. And. What happens then? I mean, is is that it or is there more beyond that? What happens once it has been processed?

Oscar Morales:

Right. So our job finishes there because we, we've worked with customs brokers for quite some time and, I've worked with them for more than 10 years of my life, in my personal life as, as friends, I have friends that are, are customs brokers, and then professionally as well. And I, we are, we're believers in Sifty. We believe that our job, it's more of a consultancy and advisor role. So we're trying to take away all this transactional, operational, low cognitive things from their hands so that they can dedicate into giving advice and consultancy and expert knowledge to their, customers. So once, once they finish using Sifty, they're ready to, finalize their import or export operation and move on to the next task. And that's it. The better news for us is when one of our customers has customs clerk that said to me, I am very happy of using Sifty because I can go home on, on time. I don't have to. And, and it was, great news because the human element for us around the technology that we build, is how can we improve your life, your work life? And honestly, and this my personal belief, any AI machine learning tool that you create and you're trying to sell to a company, if it doesn't reduce the time of a specific process, I don't think it works. It has to be aimed to that. You need to reduce times. Otherwise you're just using like fancy software to do something, but it's not doing anything.

Tom Raftery:

Sure, sure, sure. No, that makes sense. And where is the logistics industry going next? You know, what's the kind of future ai We've seen a lot of news about AI in 2023 is supposed to be the big year of AI with, things like ChatGPT making headlines here and there, schools getting very nervous about exams and things like that. Does the logistics industry need to get nervous? Or, or where do you see that going?

Oscar Morales:

Well, Well, I, I, I guess instead of nervous, I'll say excitement. Nervous, if your company or the sector you're working on does not really accept new tech, then yeah, you should get nervous because at some point this is, just going to happen. It's, it's a matter of time. I mean, taking back to your example about ChatGPT, it was just a matter of time for the, for the ChatGPT tool to be available. So it, it will happen the same, but at the core of tech surrounding logistics, there's data. If you think of an import export transaction from anywhere in the world, there are a certain amount of variables containing the transaction that are shared And these variables, this information about, it's a description of the product and who's sending it, who's receiving it, the terms of payment, the insurance pay, the freight. I mean, these variables do not change what they are, but what they contain, of course they do because the different transactions, right. The more data we can share about the nature of these transactions into a place, let's say a distributed ledger or a database. I recently heard podcast about how the Port of Rotterdam is creating this, digital ecosystem to share the data that crosses the Port of Rotterdam with anyone interested in creating something. But a so that bulk of information is what provides better ai. So coming back to ChatGPT, because perhaps right now a lot of people know about it, ChatGPT, what makes it different? It's a massive, really, a massive amount of information trained into very sophisticated computers. The more information you put into it, the better the outcome it is. So this is something that we have to understand. Sharing data, it's important. I've heard sometimes a bit of jealousy of people holding their data on, old spreadsheets, in their computers from the nineties, the eighties, saying, oh, this is pure gold. And I'm like, well, it looks more to me like dirt, and it might have some gold in it you need a tool to sift through all of that data so that you can extract the gold. And that's what Sifty does. Our company dedicates to curate the data automatically, to extract the relevant information and to give you an output so that the, company that is consuming our tools, knows what to do with and can action. There topic of security around sharing the data, and there are many ways, many things that are appearing specifically in web three with blockchains and smart contracts. There is a safe future, but I think, or I believe what it's needed now, it's understanding that sharing data is important. Sharing data as a broker or as a company will help you get better tools or because companies like, like mine, we need more information to create better tools for them. So it is the virtuous cycle, cycle, sorry, of creating better things for, for them.

Tom Raftery:

Mm-hmm. Okay. Okay. And. Where to next for Sifty? I mean, we know now where you are and what you're doing, but what are your plans for the next 2, 3, 4, 5 years? I mean, AI is, growing in leaps and bounds in its capabilities. Are you taking that on board and, planning accordingly?

Oscar Morales:

Yeah, the more we work with brokers, the more we understand, the specific needs and bottlenecks. We're able to create better products. So there's a roadmap. We have an internal roadmap of the solutions to have. The longer version of our roadmap is, Sifty to become the operating system of the logistics industry. That means a cloud of services on top of the industry, giving connectivity to anyone that would like to use any specific product or as all of them, or the whole supply chain or whatever need. Ideally, what we see in the future is this, like I mentioned before, the ecosystem of sharing data, which is very important, not only with Sifty, but with many other participants so that we can all connect to, to retrieve that information, to curate it, to create better products. And our customers will benefit from a set of interconnected, AI solutions to be able to increase their efficiency in their process. There's the thing that I want to make clear here. Um, we're not talking about huge leaps on efficiency for every process. We're not talking about go from whatever your efficiency is now to 80%. Perhaps it might be 30 or 40% initially, because that's how everything starts. But it is important to understand that if you are at zero and immediately you go to 30 or 40%, that's a huge jump. And if you take that into consideration that you're doing the same for several of your processes, perhaps the overall efficiency, it translates into hours or days. It's not only minutes or seconds. This is something we're looking at with one of our customers that, on his operation, he's able, this customs worker in particular, he's able to reduce the time that he processed parcels from, used to take him minutes to now under 10 seconds per per international parcel. So there is a change on the culture within the company as well. The way the operation used to be without Sifty and the operation is now. The way you distribute your people, the way you, command them. And also the type of conversations that you have with your employees around AI. I think there is a fear of ai, replacing jobs or, or cutting jobs in all industries but we believe that that's not the case. At least now with Sifty. Sifty needs the humans. We, we really need a human element for, for the efficiency to actually pull through.

Tom Raftery:

Okay. What about the other aspect of AI that scares people, though? The fact that it's a black box. The fact that you have things like, the infamous Microsoft Tay, for example, the, the AI chatbot that Microsoft released on the world, and it went, uh, it went far, right? Fascist, loving. how do you, how do you ensure, the accuracy and I don't think ethics are gonna come into Sifty, but, you know, how do you, how do you ensure the, the accuracy and yeah, just the accuracy of the output Sifty.

Oscar Morales:

I think ethics, it's a major role in training any AI system. I'm not gonna get too technical on this, but we can talk about bias. That the bias included on, on a training set, for example, in a training set is this massive amount of data that we have to go through. So there is a very extensive job of data scientists and machine learning engineers on curated data without a bias. So that means put it in the most natural way so that the computer can understand it. And, and, and balanced. Balance means that you're not giving the computer only good examples or the best examples you need to keep this balance and say, bad examples are this, and this are bad because of a reason. Now, this black box that you mentioned, it is being a black box, I believe because, and I'm gonna quote my co-founder Julio, right now. He talked about recently that machine learning is more about a craft rather than science. And, and this is a myth that we need to start breaking. There is a lot of science behind machine learning. To understand it, you need a fundamental understanding of mathematics, advanced mathematics and statistics, and applied analytics and a bunch of things, programming as well. But at the end, to get the outcome that you need, you need to craft it. You still need to go through an artistic process of choosing the best algorithm and and creating the best data set. So coming back to your question and, and how do we make sure that the AI that you're creating, it gives you what you need? Well, as a company, we need to be ethical with the information that we provide. We need to be ethical with information that we get. That means keeping it safe. I mean, the information that goes through our system. It's really, really sensitive information, right? We have prices, companies, sometimes we even get credit card numbers, phone numbers, addresses, so as any bank will do, we protect it. We, put several layers of, of protection, but also we treat it as, information that we want to get to our customers. With the same importance we, we get, we treat the information that we receive in.

Tom Raftery:

Okay. Okay, cool. We're coming towards the end of the podcast now, Oscar, is there any question that I haven't asked you that you wish I had or any aspect of this we haven't touched on that you think it's important for people to be aware of?

Oscar Morales:

Well, rather than a question, Tom, I would like to make summary of what three main takeaways, for Sifty to your audience. So Sifty is about efficiency. We produce software to decrease the time on certain bottlenecks and, to make them easier for customs brokers. we use a lot of machine learning. We use a lot of advanced technologies, but we work with the humans, so there's a human element to our solutions. We, we partner with a human to get the best out of the technology. And something that perhaps I haven't talked much about is our passion at Sifty and why we, we like what we do. We really understand technology very well. We understand how to apply machine learning, and we like the challenges of operational bottlenecks at Customs specifically for customs brokers. So we have this passion and of curiosity and how to solve things with a lot of creativity. So, our solutions are a bit out of the ordinary. They're not very standard, but they work. and, and our customers are saving time. They're saving time on per operation, means they can get another operation or two or three more within the same time. So it's increasing revenue. We grow with our customers and that is one of the most important, aspects of, of our company.

Tom Raftery:

Cool. Cool. Great. Oscar, if people like to know more about yourself or Sifty or any of the things we discussed in the podcast today, where would you have me direct them?

Oscar Morales:

So So our our website is the main source of information siftyml.com. Over there you can find information about me, about Julio. Our our LinkedIn page is there. Once you get into LinkedIn, you'll see news and resources, written by the team and, and ourselves, the co-founders of the company. So I will, I will invite anyone to just reach out through the, through the website. There are many channels of communication there. If they will like to join our newsletter, I will be more than happy to send them a once in a month curated news outlet of interesting things that we've seen, that we've heard, and a little bit of our products.

Tom Raftery:

Nice. Nice. Cool. Okay, Oscar, that's been really interesting. Thanks a million, for coming on the podcast today.

Oscar Morales:

Thank you Tom. I really appreciate it.

Tom Raftery:

Okay, we've come to the end of the show. Thanks everyone for listening. If you'd like to know more about digital supply chains, simply drop me an email to TomRaftery@outlook.com If you like the show, please don't forget to click Follow on it in your podcast application of choice to be sure to get new episodes as soon as they're published Also, please don't forget to rate and review the podcast. It really does help new people to find a show. Thanks, catch you all next time.

Podcasts we love