The Role of IT in a 3PL with Eze Peralta
Episode Transcript
DD Spotify DD Apple Podcast

In this episode of Everything is Logistics, Blythe Brumleve interviews Ezequiel (Eze) Peralta, the VP of Technology at SPI Logistics, about the role of IT in 3PL.

Ezey talks about his early background in technology and databases, including creating software for his own small businesses. He then goes on to describe his work with SPI Logistics on developing data warehouses and analytics solutions.

The discussion highlights the importance of investing in IT infrastructure for 3PLs and the significant role it plays in optimizing their operations.


  • Connect with Eze via email at



At SPI Logistics they have industry-leading technology, systems, and back-office support to help you succeed. Learn more about SPI’s freight agent program here. Make sure to let them know we sent you!

Digital Dispatch helps you speak confidently about ROI with a website built for your customers, prospects, and employees. With plans starting as low as $90/month, learn how you can take your website from good to great by visiting Digital Dispatch.

Show Transcript

See full episode transcriptTranscript is autogenerated by AI

Unknown: 0:00

LinkedIn presents

Blythe Brumleve: 0:10

welcome into another episode of everything is logistics, a podcast for the thinkers in freight. I'm your host, Blythe Brumleve. And I'm happy to welcome in as a Peralta. He is the VP of technology at SPI logistics. And we're gonna be talking about the role that it plays out of three PL so as a welcome into the show.

Ezequiel Peralta: 0:28

Yeah, thank you very much.

Blythe Brumleve: 0:30

Absolutely not. Now, for folks who may not be familiar with you and your background, kind of give us an idea of how you first started off with technology. When did you know that technology was or IT infrastructure was going to be your career path?

Ezequiel Peralta: 0:46

Well, I think at a very, very early age, I started using computers when I was age for the first computer, my dad was into learning, graphic design and this kind of thing. So he got a an old Mac. And that was kind of like love at first sight, started on the graphic design, more like Illustrator, Photoshop, that kind of stuff, then I, but always really interested in how things really work and the mechanisms behind it, I always knew that I was going to get into coding one day. And then yeah, I used my first municipal databases with with some custom build databases that I did for a small business that I had with. I used to do screen printing. So I created the software for that business and then create a software also for, for other, like small businesses like that. And then my, my first, like bigger job opportunity was the online travel agency where we were with with with a friend, we were developing the data warehouse and all their analytics solution. So that was my first like bigger project. And that was kind of like what led me to SPI the first project that I that I did for SPI with, with my friend Valentin, who is now also working with SPI was the data warehousing tool. And that was our first experienced then I got the offer to, to come on board. And yeah, thanks to the trust of offer our leadership team, Mitch and and Joe Chandler. That that's when all started in logistics for me. And I have to say, I'm really thankful that of all the learning process that I had in SPI about the freight industry, because when I just first started, it was very new to me, I had an idea of the brokerage model from online travel agencies. But I learned so much from from, from everyone here. Yeah.

Blythe Brumleve: 3:04

So can I nerd out with you for a little bit? How do you build a database?

Unknown: 3:09

What more than Well, building a database or, or an implementation of a database models for for a company? Generally there is a lot of research on the on the on the data model on the design the data model, right? What is the scope and the grain of the data that you want to store? It's like you want to store things by date is your business related to timing issues is related to quantities is related to you know, more a very general level than, for example, for a freight brokerage, or these are very complex data models, because, for example, people usually tend to oversimplify, but it's a lot about thinking about more how you model reality into into an abstraction, or in this case, a database. But it's a lot about talking with the actual business experts in order to know how they think about things and try to translate that into a mechanism that that will fulfill those requirements. Something that we that I noticed and analyzing data models for for TMS systems is many times, there's like a glitch or an error or an issue on the on finding that granularity for example, okay, you model a load, right? But then what happens if that load has many line items, and each one of those line items go into a different truck, and then that truck has maybe a team right driving so then you have the driver so then you're doing tracking on the driver but then that load changes driver then, but it's all in the same order right and the customer will see will want to see that as one right. But as one order, but maybe on the carrier side they will see a piece of that order. and you don't want also that data to leak right or? So it's a lot of about that, like designing databases is one of my main. I don't know, I really like it, data modeling aspect of it. But yeah, that's usually

Blythe Brumleve: 5:16

because from what I understand, and I'm very like loose sort of, Oh, I like to consider myself as tech savvy, but I don't you know, that walking into a server room at a, you know, at 3pm was my big like, eye opening moment, like, wow, this is you know, how a lot of the functionality of how things work, but then you're talking about sort of the inside guts of that that server itself and how you're programming different texts, and different I guess TMS, you know, the information that you're reading within a TMS, you're coding that a certain way, so the database understands it. So if you're looking to pull records or sort records, then it will be pulled correctly is that a kind of a general overview?

Unknown: 5:58

Yeah, and for example, the main challenge that was one of the main tasks that I had when I started with the data warehousing project here is that we had at that time, three different TMS systems in production. So there was a group of agents using a legacy system that was a client server system, then there was a group of agents working with a web version of that same TMS system. And then there was a subset of agents working with the newer TMS system that we were migrating to, eventually, we'll get everyone into the newer TMS system. But in the meantime, we needed to have our KPIs every day on how the business is doing. Now, one system will say, okay, a load has these fields, right, all these columns, but the other system would say, well, a load for us is these other columns, right? So we needed to create basically all the architecture and the database modeling to pull all that into a common model that will be able to tell us no matter where the load is coming from, no matter the TMS can build that SPI owned abstraction over the business objects that we are dealing with. Because what is a load? Or what is an order for SPI, it might not be the same of what is an order for ABC, TMS or XYZ, TMS? So yeah, that's, that was that was actually one of the of the main challenges. That's how I learned a lot, a lot a lot about the freight industry itself by just seeing the how information is stored in this system. Right.

Blythe Brumleve: 7:38

And so what is I guess, for folks who may not be aware, what is the I guess, positive outcome or maybe dangerous outcome of not having a database that's properly maintained and organized? What could be the I guess, the downstream effects of that not being the case of not being properly optimized?

Unknown: 8:00

I think you lose power on the inside, right, that you can get from your data. I think everyone is familiar with the term garbage in garbage out with shared data management, right? And then, I think more important than the data that is stored in each call each table on the database, then you have the aspect of the relationships, right? How does relate this? A load? How does relate a load with a stop? How does it relate a stop with a tractor? How does this relate with, and not having that model properly, gives you a lot gives you a lot of issues that you don't know that you're having, because it's hard to know, or how to see an issue where you don't know that there's an issue there because it was never there is not something that it was done, and then it broke, it was never there. So basically, I think, yeah, the main point would be losing possibility of insight in the future over your own your own data and operation.

Blythe Brumleve: 9:05

That makes a lot of sense. It's almost like for, you know, the marketers out there, you know, or the sales leaders out there, it's, you know, having a lot of crap leads, maybe in your CRM, like your looks good to have 20,000 leads in that CRM, but if it's crap, or if they're fake email addresses or something, then you know, you're not going to find out until you actually go to start sending them messages.

Unknown: 9:26

Yeah, or you end up you know, because maybe from a lead, you cannot convert that into the model, the proper follow up objects of okay, a lead becomes an opportunity and opportunity becomes an account. And maybe if you're missing some of those tables, where to store information or you don't have those relationships there in the data. Then you will end up basically putting all that into another system or in a spreadsheet or kind of like, you know, building something around it and that's when things start go sideways, because all the purpose of implementing software is to fulfill do business need right? And then kind of adapting your, your processes to the software instead of the opposite and think? Yeah, that's that's not a good path.

Blythe Brumleve: 10:13

This episode is brought to you by SPI logistics the premier freight agent and logistics network in North America. Are you currently building your freight brokerage his book of business and feel that your capabilities are being eliminated due to lack of support and access to adequate technology? At SPI logistics, we have the technology, the systems and the back office support to help you will succeed. If you're looking to take control of your financial future and build your own business with the backing of one of the most successful logistics firms in North America, visit SPI three to learn more. And so when you so how long ago? Did you join? SBI?

Unknown: 10:52

I started with SBI in 2018. Oh, okay.

Blythe Brumleve: 10:56

So that I guess organizing the databases was a big part of your job initially. So what does your sort of the day to day look like for you now? Is it is it managing that database now?

Unknown: 11:08

More than database, I think I'm more focused on the business processes and how trying to see the overall picture besides the operational side of the freight brokerage portion. When I started here, one of my main goals was to talk with the team who is actually using the software, the client care team has been so helpful. AP and AR team has been so helpful in teaching me how they use the software and they came up with all these would it is, is there a way how we could do this, instead of you know, entering the spreadsheet data into the system? Is there a way we could do it better? And then we started talking about it, like, we came up? In some cases, we realized that we didn't need that spreadsheet to start with. Because there were there was another way where we so instead of running into the implementation of a software solution to do that, we started questioning, do we need that spreadsheet? To start with? Do we need that process overall, to start with, or maybe doing a couple of changes in the process will basically define all your problems out of out of scope, and you know, get rid of all these problems without even implementing any software? Because for me it, I prefer to write less code. And I prefer to maintain less infrastructure, less software and less. And the the code that never fails is the code that never is never written. Right. So I feel happy when I have I get to delete code, right? Because I'm decreasing complexity. Basically, in many cases, sometimes you need to, you know, add, but I think that's, like an overall rule for for everyone into the software world is no goal after implementation right away, it's like, Okay, why don't you want to implementing isn't necessary at all, you know,

Blythe Brumleve: 13:08

and I think that you, I love that you said that you actually talk to employees, before you started implementing a lot of different tech solutions, or, or trying to essentially put a BandAid on a broken leg, I think that that's what a lot of organizations do is that they just see the shiny new piece of software, this shiny new piece of tech, and they just go ahead and buy it. And then they you know, throw it on the team that's actually going to use it and the team hates it. So I think that that I love that you answered that question that way where you're talking to the team, before you actually make those tech investments is that as I would imagine, that's proven successful for you guys to to actually talk to the team members first, right?

Unknown: 13:51

Yes, for sure. And there has to be a balance because sometimes people will want to, of course, stick to the old ways and the way they're doing things the way they know. So there's, you need to also kind of, like, encourage them to where they're gonna see the improvement. And I think sometimes there's some sort of arrogance from the, from people from the software, where it's like, Oh, why people are hating this so much, if it's so much better than what they already do? Well, they are the ones that let's listen to what they have to say because they might have good points and maybe there's little tweaks that we can do in order to you know, fit those needs. So yeah, it's it's proven successful and I think seeing trying to see see that's an overall approach besides okay operations and you know, TMS and then back office, but because let's say you get a new TMS system, right and then that, then you increase your order count by an order of magnitude, right? You have three times more loads, right? Okay, now you have to pay them. Now you have to build them Now, if your back office is working on spreadsheets and printing checks in your office, then suddenly you're going to find out, you're you need a bigger printer, or you need more, what people do is okay, you have all these maybe super fancy software doing the order management. But then on the billing side, you have basically an office with a bunch of people doing very repetitive work. And then during whatever you're winning on one side, you are basically, you know, losing on the other side. So you need to grow all together. I think that has been my idea from the beginning, like, putting a lot of focus on the back office has been has proven also successful.

Blythe Brumleve: 15:43

So give us I guess, a sense of the the tech stack that SPI if that SPI sort of, you know, I would imagine that that the TMS is sort of the backbone of the entire organization. And then what do those additional tools sort of look like? Are you, you know, investing in third party software and integrating that into the TMS? Or what does the tech stack look like, for a typical freight agent coming on board?

Unknown: 16:10

Yeah, so the tech landscape. At the center, I would say, we're trying to move the TMS from the center of the tech stack. And, of course, it's always going to be the center of the stack, because it's reputation management software, and it's built for it, it's where most, most of the domain lifts, right, the all the objects that have the naming of what we deal with a load stop driver carrier, that's lives in the TMS. But that's not everything to the business. So we have the CRM, which is how we get our leads for new agents. And then we have our AR and AP, right? So we used to pay for processing payments, we use them also for doing OCR on the pod that we receive, so all that process is automated. So we link we built an integration with with them, so we can basically automatically submit the payments and audit if a payment. If a payable is ready to be sent, then we send it but if not, it will go into a different queue where we willing to evaluate, okay, there's no mismatch on the there's an exception on the mismatch on the on the on the amount or whatever it is. Then we have all the load boards, right? The T truckstop. Then we're onboarding now with parade to fit in digital matching. We also have punishments for EDI connectivity connecting with all the EDI, valueadded networks and FTP providers. Then we have where's the actual freight for LTL. And then we have we have some bots initiatives also for for those sites where there's no API to call. We also have, we create our own visibility sharing platform, I would say it's all in the background. There's no user interface. But we basically created this reusable components to be able to share location data with any of our shippers no matter the system they use. So we know forecasts before before. So we're feeding their systems with our location data that we get from the tracking providers. We have microphone p 44. And trucker tools, we like to always offer like a few options because you know, features that we don't have in one we can get from another one. And when I was talking about trying to not have the TMS at the center is that we built and we are continuing in the process of continuing building an integration layer or what we call SPI data services. Where for any integration that we do, we integrate the TMS and the data sources with our SPI data services platform. From there, we can feed data to that warehouse and also feed any other data source. So we created some custom libraries are code libraries to to read the light in our own integrations. We deploy this in our in our own infrastructure in the cloud, trying to follow all the you know best practices regarding security, observability governance and all these things. And that has proven also a good approach that you know owning your own integrations. I think that's the key because you cannot be waiting for your software provider to do To integrate a system because maybe that that's not something they want to do, because maybe it's not good for their customer base. But in our case, we cannot be waiting for, you know, another vendor to say, oh, yeah, we're gonna integrate this, and maybe they integrate it, but maybe they don't integrate the data points that you need. And then you waited all that time. Right? Oh, and it's like, well, it doesn't work for me now. So yeah, we built we built quite a quite a nice, I think platform to allow us get the best features from every vendor, but give them a context and a process and a way of using it and documentation and to make it work for us.

Blythe Brumleve: 20:44

And it almost sounds like it's like an apple, like an Apple store with a bunch of apps that you can kind of choose pick and choose on which ones that you want to use as a freight agent. Is that a safe bet? Or is it kind of, you know, this is the suggested tech stack that you should use for X, Y and Z?

Unknown: 21:01

Yeah, actually, actually. One point it's SPI is another alright. So, each office is different. And SPI embraces this difference, instead of trying to equalize and lean or trying to say no, this is the way how you should do things. Of course, we have proposed best practice of course, we have, you know, how we think things should be done. But also we understand there are as many ways of working as as dispatchers and brokers are. And even in that best practices, path, complexity and differences, right? So we try to embrace that. So if someone wants to use EDI for their tenders, but then they want to use load boards in certain way, then we set them up for that right. But then someone else says Oh, actually I want I want to feed my loads from a separate system, well, then we will evaluate if that the complexity of the project and then if we can do it, then we will we will just integrate it. So we tried to make it so it's evolvable. So because maybe tomorrow you will want to incorporate the new two okay. So we just plug it in right. And we adapt to the complexity and requirements of each agent basically. So some agents, we had the case for example, with LTL. The the the TMS we use, it has a good LTL integrations. But for people who is focused on LTL, they need maybe an LTL focus TMS or software, right? So what we did is we created an integration where they booked their loads in, in an LTL oriented software, which is our CO freight. And then we sync those that data with our main TMS. Team, they don't even know where that payable or billing came from. So if it came from TMS one or TMS two, it's a billing and it's a payable, you should you should audit it, you should pay it, or you should, you know, build it to the customer. So that's kind of the approach that we want to have. Where if you come to SPI, and it's like, oh, actually, I want to use this combo of features these combos things. Well, we can make it happen. Yeah.

Blythe Brumleve: 23:38

So I guess when when you're or a new agent is thinking about joining SPI do they typically meet with, with you or somebody else from the tech team in order to answer some of those questions? And if so, what did what are some of those common questions that they're asking? Yeah, so

Unknown: 23:57

sometimes they would meet with, with me before onboarding, just just to get to know, you know, I will give a demo of the system. Our client care team can can give them was to but sometimes they want something more specific or I want to know how EDI looks like EDI tender it looks like in your platform, okay. And then we do it or you know, what options you have for tracking how reporting works or so generally Yeah, we would do a demo, they explain what they need. And then we explain what are the what is the toolkit that we have in order, you know, to to fit those needs, and how we can make it work. And again, some sometimes because we are a small team, we are not a software company, so we need to evaluate it. Our resources are also limited. So you know, some it might take more time for us to build something than if software company dedicated software, software company But in reality, because we have already so much already built, it becomes faster. So speaking, we're starting from zero, right? You already have all this toolkit, and it's basically grabbing this piece, this piece this piece is how can make it work for you.

Blythe Brumleve: 25:16

And so I would imagine, for a lot of these folks that are going to that are interested in becoming a freight agent, and they contact you guys, and they're ready to move forward, I would imagine that they would be saving a considerable amount of money by using some of the licenses and the software tools that you guys have already spent a, you know, an incredible amount of not only money, but time as well as investing into these platforms. But what if somebody wants to keep, say they've had, you know, a load board for years? Or they they've had their own, you know, sort of CRM for years? Can they keep that valuable data? Can they migrate it over? Or do they have the option to keep it in still within their, I guess, their own ecosystem?

Unknown: 26:00

I think, yeah, I think they have both options, or three options, they could keep it on their side, of course, whatever data they need to enter as part of SBI will be shared, they can choose to migrate that into our operational system, or they can choose to migrate that into our data warehouse for just to store it, and then start fresh with a new install of the TMS. But that's something that we usually do, for example, people has this big list of locations or customers, so we have processes to just upload all that into our platform. And then. So yeah, they have both options. And even Yeah, if they're using a specific load board, and if that loader has an API, then we will work with them to integrate it. Same for visibility sharing, or this very common requirement that we have is all my shippers are wanting, you know, to get a daily file with the location of all their shipments, we can set that up in a day, you know, and I don't know, how is the process in. But if if a small broker, like a one person operation would want to build that out, no, to send a daily visibility file or real time updates to the shippers they will have not only to buy software, but also work with to adjust all the little knobs and they probably don't have time for that. And, and that's the thing, that's the thing, the timing, that that's something that gets forgotten, people are on their day to day doing stuff and doing they're, you know, taking care of their orders, and they don't want to spend that extra time. And they shouldn't have to. And so

Blythe Brumleve: 27:43

I imagine, you know, after you know, maybe some of those initial questions are answered, you know, a new agent signs, and they're ready to get rolling. What does that I guess, onboarding time look like? Because I know for you know, a lot of folks, it's you know, you got to manage the freight every single day, all day, every day. So any kind of, I guess time away from that is almost seen as a frustration. So how do you sort of mitigate that level of frustration as they're onboarding and learning these new systems? What does the I guess the time period look like? And how do you I guess you make it as easy on them as possible.

Unknown: 28:19

So we have an amazing client care team, that will will give the training, we have a portal where you have videos on the training for the TMS, we can upload the locations and customers and we do all the credit check and all of that before you know they they go live date. So by the time they get their login, people can go in the system and start booking loads. From day one from day one, you can you can you can book your loads. And that including, you know, posting to load boards are boring. We, we put a lot of effort in automating the onboarding process and tracking it. So we have you know, we schedule our training sessions. After those training sessions, if there's there's a need of upload these locations for data, we do that so by the time by the time agreed with the with the new agent, we they can start booking loads, day one. And then of course, like getting used to new software, maybe you are slower at the beginning. And then you know, you start getting up to speed and three more new customers and these kind of things. But

Blythe Brumleve: 29:34

it sounds like to you know, it's one of those things that you don't really know, I think how it's going to impact your daily life until you're in the new system. And so it kind of goes back to your earlier point about defining what those processes look like and then maybe, you know, it's changing a little bit of your processes on your end of things in order to speed up things later on. So they're not only you know, saving money on a lot of these different licenses But they're also getting better and more access to different licenses that they probably didn't have before. And then now with a lot of the different automation that you guys are integrating, then they're saving more time, and probably allows them I would imagine to be able to develop deeper relationships with their customers, what are some of those other? I would imagine that that's a lightbulb moment for them, once they really get into the system. Are there any other sort of lightbulb moments for them, you know, making this transition tech wise?

Unknown: 30:30

I think I think a big part of it is the fact that they don't have to care take care of billing and paying. Right, the back office support, that's a huge thing, because it allows people to focus on, you know, managing freight, not managing, you know, if a cheque didn't go to the, to the right address or things like that. At the same time, I think there's a high benefit on participating in an organization that is oriented to high efficiency. And it's not that we, you know, when we're going to think about it, we're going to work with the agents in order to understand the problem and try to find a solution, not a not not saying, Oh, you're wrong, you should do it this way. Okay, what are you trying to say? Right, like how we can? I think a light bulb moment is when? Yes, so I just need to book freight into when my customers will? Yeah, that's it. And, you know, it's, and they gain a lot of the fact that because of being part of an organization that is focused on high efficiency, they start seeing the benefits of that as well of, okay. I can learn from these business processes, I can learn from other agents as well, I can learn from this network, you know, how, how can I do my, my part better, but knowing that, caterers are gonna get paid, your money is going to be collected, you're going to get your Commission's, you're not going to have you know, any, if you have an issue, then you're going to call client care, and they're going to help you right away. If there's a claim, we're going to take care of that, you know, work with the agent to solve that claim. So I think that light bulb moment is where, oh, now I can, now I can be hyper focused on on, on on my business.

Blythe Brumleve: 32:32

And I think, too, it's your highlighting a good point that for a lot of these folks, I would imagine that you don't know what you don't know. And then you finally get into a new system of process is a new way of thinking. And maybe at first, it's kind of a little worrying, because you're so used to having your hands and everything. And now you don't have to worry about that anymore. So maybe at first, it's kind of jarring a little bit, you know, surprising. I know, I've experienced this in my own business, where I've implemented different automate automation strategies. And it feels weird to not have to micromanage each of those little things. I imagine a lot of that is going on with them, too. And once they make or complete that onboarding process,

Unknown: 33:14

Yes, totally. I think the key to come that is communication and observability. Like if you can have KPIs and reporting on on, you know, how much how many loads you were able to move? Before? How many loads? Are you being able to move now, how much time you were spending on these processes, and that, now you're spending zero? And that takes time, and that, you know, some people will, it depends on the personality. Also, some people will still want to have their hardcopy on paper. And that's okay. Like, they want to have their spreadsheets with, you know, their, what their commission should be. And I think we encourage that. It's like, it adds to the transparency. And if there's an error or an issue with with, you know, data inconsistency, then well, thanks for letting us know, we'll, you know, work on that. But yeah, the key to that, I think is observability, reporting like KPIs and be able to see that and measure.

Blythe Brumleve: 34:25

Do you wish there was a central place to pull in all of your social media posts, recruit employees, and give potential customers a glimpse into how you operate your business? Well, all of this should already be on your website. But too often, we hand that responsibility of building our online home off to a cousin, a neighbor's kid down the street, or a stranger across the world. Digital dispatch believes in building a better website at a fraction of the costs that those big time marketing agencies would charge. Because we've spent years on those digital front lines are Our experienced team focuses on the modern web technologies to bring in all of the places you're already active online, show off those customer success stories, and measure the ROI of it all in one place. With manage website plans, starting at $90 a month, head on over to digital to see how we can build your digital ecosystem on a strong foundation. We've got explainer videos right on the website and the ability to book a demo immediately find it all over at Digital So with when everybody you know, I guess, when COVID happened, you know, a few years ago, everybody's working from home, there was a big sort of outcry, I think in a lot of like the corporate tech world that this is going to open everybody up to, you know, malicious attacks and cybersecurity issues and things like that. But with an agent model, you guys are kind of used to having these remote offices all over the place. How were you, I guess, monitoring or you know, mitigating, you know, potential cybersecurity risks in sort of this environment that most companies find themselves in with remote offices all over the place.

Unknown: 36:12

Yeah, so that there, it would depend on on the, what's called the vector, right, or where the attack is coming from, or there are many ways so for example, email phishing, like, it's, it's really hard to combat because, you know, if it only takes one person to click a link, enter their password, and then, right, so die, we try, we do what everyone does, which is protecting our email server as much as we can. And we use Cloud for that we don't use on premise for emails, or we don't use on premise for, for almost anything, we do have some some internal storage, but we are leveraging cloud infrastructure in order to cover that, but use in a way that, for example, all of our we use multi account setup. So every integration that we do is its own little world. So if you're attacking that integration, basically, we can take it down and spawn it. You know, we can spawn like another one, like, with a different with different, you know, URLs and everything in minutes, or like, you know, because we've been putting a lot of effort on automating our deployment, automating our infrastructure, we're using infrastructures code, I'm going more in the internal parts. But

Blythe Brumleve: 37:43

no, I think that's cool. Whenever you know, cybersecurity issues happen and what you have to do to mitigate those risks. And in a modern day environment. Yeah. And

Unknown: 37:51

we have daily snapshots of our databases and daily snapshots of our we log all the queries and the data warehouse and all of that. And I think, what our I think cloud cloud infrastructure, using the right way helps a lot in and it opens up other possibilities for attacks. But I think, yes, we try to not keep all of our data as if you see what's before people will have this one database that if you lose that database, you're done. So we have our structures more distributed.

Blythe Brumleve: 38:28

Now, I don't know if you watch the office, but there's like that one scene in the that, well, you know, even if you don't watch the office, but there's that one scene that you know, kind of everybody will will use as a meme where it's like, Oh, my God, it's happening. And you know, everybody starts running around the office like crazy. I tend to think that that's what happens when like a cybersecurity attack happens. Is that accurate? are probably not at all your you guys are probably, you know, you know, have so many different protocols that you don't even have to worry about that. But that's what I imagined happening anyways.

Unknown: 38:58

Yeah, knock on wood. We didn't have a serious incident. Like, in our, in our systems, like, at least since I'm here, we had like phishing attacks or to emails people trying, you know, to get get to other users, emails and these kind of things. But on the on the, on the TMS side, and on the integration side. We take it very seriously, and we try to be as preventive as possible. And so far, it's been good. And so yeah, it didn't happen. Yeah. There was no, there was no panic so far. But I imagine that would be what what what happens, but that was one of the vectors that were more exposed as we need to really monitor multiple systems, right? But so what happens if one of those systems go down? But the fact of the fact that we have our integration layer means that You can replace that with a spreadsheet. In the meantime, while while you're, you know, you can mitigate that, and your entire operation will continue working, you will probably slow down because you have one component that is not feeding the overall process at the rate it should. It should. But you can still operate, right like even for us, for example, if we don't have access to our TMS but someone wants to get a report on how many loads we did yesterday, you can just go in and get it because that data comes from data warehouse, a different system, different storage, different permissions, different credentials, different everything. So you cannot attack the entire thing. Right like that. I think that's the main point.

Blythe Brumleve: 40:43

So there's different layers of security that that sit on top of it. Yeah. Now, you know, we've talked it we all we've mentioned the phrase a few times during this discussion with automation, but I kind of want to talk a little bit about, you know, AI and chat GPT can, you know, really, like take a step outside without somebody mentioning it? I feel like it definitely in my world, you can't not hear about it multiple times every single day. I'm curious as to you know, what your your early thoughts on using systems like this, on top of the data that you already have, which I think is kind of where Chad GBT like technology is heading.

Unknown: 41:22

Yes, I think we use, we use GPT. And we use also other AI tools for coding, oh, we use GitHub co pilot. And I think these tools are awesome, as long as you know, you know, to how to direct them. Right? So generally, I found that for junior developers, it helps but it's not maybe it's not going to take them to the next level. Right. But in the hands of someone who is more on the software architecture side where, you know, and data modeling, as we were talking before, or, you know, exploring a problem, space, and its solution space. And all of that, I think is interesting, because it can provide a lot of nice insight, and it can speed up a lot of the boilerplate code and, you know, sometimes, but it's not gonna help you define your processes better, because the, the business users are the ones who really know how the process should be, and, and should drive that. And I think, going back to what you're saying about data, right, like, I think there is an under underestimated the power of the processes over the data, because people are focused on the output on the which is the data itself, right? It's, oh, I'm seeing this spreadsheet or report or BI tool, but that that is just a output from a process, right? And what is your process? So how that data gets cleaned up? How it gets collected? How it gets? Is it compliant it the way how you're collecting, it is compliant, or you're getting in legal trouble because of how you're collecting it? Like there are so so much more to that than, you know, seeing the number and also with business process? So I think yeah, to answer your question, Chad TBT, we use it a lot, mostly on that site, but we'll give it a lot of direction on to what we're building. And and yeah, I think it works better, the more you know about what you're doing.

Blythe Brumleve: 43:40

That's interesting. You're the first I've asked, I think in like the last four interviews I've had, I've asked, you know how, you know, folks are using chat GPT like technology, and almost all of them have said that they are I think, you know, about three out of five had said that they are and then two of them are kind of, you know, just a little scared by it. But I agree with you, I think it the outputs are there. And you can see those kind of early lightbulb moments happening. But it's still take somebody that is highly experienced that can see the nuance in it, especially like on the on the marketing side of things, or even on the podcasting side of things. I'm using it, I'm using chat GPT a lot there for that, but I know what to look for and what not to look for, if that makes sense. Like some of the stuff that I'm getting like a you know, they'll it'll so for example, I'm using it in, you know, transcriptions so I'll take a podcast transcript and upload it to chat GPT and tell it to find me the most, you know, five, you know, impactful moments from this conversation. But sometimes it will pull things that we did not even talk about. And so you still have to have that nuance that experience level to fact check it to make sure because it will confidently tell you an answer and it might not be the right answer.

Unknown: 44:58

And also these tools are So buyers that we can actually get them to say what we want him to say. And it becomes this feedback loop loop where you can, you can basically make Chetcuti say that two plus two is five. And so and I think you'd said a key word there that is nuanced, like humans are very highly adapted to understand the nuances of everything we are always, you know, are curious minds are trying to, you know, find all those patterns that don't match. And you know, those details, and I think AI tools are not as good on the adult level. So they're very good at, you know, okay, write a piece of code that does this thing. But then you need to go in and find those nuances and make it if you just use it as is, you're probably going to, you know, suffer that pain in the future when you need to go back in that to that and do it again, because

Blythe Brumleve: 46:04

do you see, I guess, in an environment where like, a chat GPT powered, like chat box is sitting on top of like a TMS, and, you know, maybe for some of your customers, or some of the freight agents that can, you know, use chat GPT to ask, you know, what's the most efficient carrier to handle this lane for me to something like that exists? Or it kind of sounds like, you know, maybe like a parade would be doing that instead of like a chat GPT

Unknown: 46:29

Yeah, I think because GBT is trained for natural language processing, right. And in many cases, to do digital matching of freight, you are looking at other types of, of data is not natural language processing. So chatbots could help mostly on the support site, maybe because they will give you they will make it really, really nice answer about this should be your carrier because of this and this and this. But if they don't have access to a machine learning model that uses the right signals to determine that, how can you know that? It's, it's actually that that carrier is the best? So I think GBT is mostly oriented to natural language processing, other machine learning and AI tools, I think will be a best fit for more like digital matching rates,

Blythe Brumleve: 47:27

like specialized different activities, where chat, GBT is just not there yet,

Unknown: 47:32

we should not underestimate the complexity of the problems we're dealing with, like having dynamic pricing is not a trivial problem, like matching your carrier with a load is not a trivial problem. And there's a lot of these also, like, of course, on a sales on a sales call, there's going to be you know, this is easy, it's going to be easy, you know, we're going to give you this number that is a magic number that if you put that number in your bids, you're gonna win that bid. It I don't, I think it doesn't work like that. It's, it's just, it's more complex than that it might work for, for a subset of people a subset of cases. But there are so many,

Blythe Brumleve: 48:16

I think everybody wants a blanket approach to this, forgetting that there's other people on the other side of the transaction that have their own needs and specifications. And they're all nuanced that that they want to pursue or prioritize.

Unknown: 48:30

And something if I may add something to that is, I think there's something kind of important to mention about the model itself of software companies and startups. And versus having people in your broker, your your freight company, a freight broker company who understands technology and is able to at least build something, even if you're not going to build a TMS. Because SPI is it's been in the industry for a while, right? And anything that we build is going to stay with the company for a while. We're not trying to build something quick, how to market and then we get some customers and because that's not our business, our businesses freight broker, not selling software, with technology companies, sometimes they just run after a set of features because they got a few, you know, contacts at some, you know, business experts that they tell them, Okay, this way you should build is very bias to that subset of people and then they try to build it as quickly as possible. And in that process, sometimes quality gets an possibility of customization and then all the nuances can get lost. So, I think, yeah, there's something kind of like broken in that model itself. And I think that why we took this path of, okay, we need vendors, we need all this technology because we, we, we cannot build this dynamic pricing tool ourselves in a time, that would make sense. So we're going to use them, but okay, we're going to also own the integration, so we can use them in the way that we need, making sure it worked for us. Also, what happens if that new company that is all new and shiny, doesn't do? Well, it could happen, and SPS still be here. But we we have seen companies, tech companies come up and down, you know, get, you know, or maybe they start sort of focusing on brokers, but then suddenly, they have a big trucking company, customer, they start adding features that are only for acid base. And then we are like, Oh, we are customers, do we want these new features too, but you need to wait, because now we have another customer that is, you know, and, and that's part of their model, because of course, then they have to do that because they need to be profitable. And we need to be profitable, too. So I find the tension there. And I think we a lot of what we built it kind of answering to that tension as well.

Blythe Brumleve: 51:18

And I think that you you said it very well, where it's almost it's it brings a whole conversation, you know, kind of circling back and full circle where you're really analyzing your processes from not only a personnel level, but also from a business level. And you can't be distracted at the business level from the shiny new tech object that takes you away from your overall mission. And that is, you know, providing technology solutions that can you know, help your freight agents now and in the future. So, with all that said, you know, where do you Where does I guess sort of the the tech outlook tech outlook look like for you know, maybe in the months to come? Maybe, you know, in the coming year? What should you know, maybe some freight agents that are on the fence, you know, about joining the company, or maybe some some current freight agents that are that are listening to this, what should they expect, you know, coming down the pipeline from SBI?

Unknown: 52:12

Well, well, we are working on is into, in all continue improving the integration layer, so we can continue making it easier to integrate whatever new software and tools that may come up, make it easy to integrate, then there's some automation in the back office also as well, because we know that the company is growing. So we need to make sure that you know, we we don't have a bottleneck in our back office that there's there are some automation initiatives with for the agent facing, we're integrating digital freight matching. Next to that is going to come the digital booking as well what they call the book now. So we have LTL integration with LTL specific tooling. And all that is seamless and connected with the TMS. So those we are launching a scoring program for for carrier. So we are director of carrier procurement Mark came up with this amazing, amazing logic to to to score carriers. Yeah to offer more tools to agents to to book their freight yet to get access to to improve their carrier relationships to reduce the possibility of claims and fraud. We'll be working also on the dynamic pricing side. But we want to do that all in, you know, we want to do it in a solid way because we rush it Yeah, because I find that there are certain tools are not mature enough yet. That my perception at least is like they're very new. But like you know it? Oh, you should charge $2,000 for this. When I asked how did you get to that? I never got a clear answer. How do you get to that? And they don't want to tell you. So it's trust. So we might we will you will use these tools but okay. How because once you start automating, it's okay. Let's automate bidding, right. Okay, now, are you bidding at the right price? Are you bidding on those that are actually good for you? So all of

Blythe Brumleve: 54:40

that, yeah, that's interesting. It's not just about you know, integrating the tool and then never looking at it. Again. It's the constant monitoring of it of what is the value that this tool is bringing to our platform?

Unknown: 54:50

Yeah, exactly. Because it's okay this tool is telling me that carrier A is really good for you is very close to your load. Does that means that you Good for your load is telling me that it's closed by. Now, if I what we did is okay, we get that information from a vendor that this case parade gives us the carrier matches by location. But maybe we find that doesn't in our right rating, that's an F carrier. So maybe, you know, you don't want to book with that carrier, because he already had with SPI, so isn't that those are the things that we're working on right now, you know, in all these integrations that we're building, adding all this, and also, our agents don't need to log in into a separate system to get all this, Oh, that's good. We got, you know, oh, we have a new tool, okay, here's your login, okay, then, you know, now you have to go to another 72 tabs, open that data, and then you need to copy it back somewhere else. So we don't like that. So all integration, we will we try to bring into the TMS. And, but keeping the option that if we need to the other way, we can, we also have the option.

Blythe Brumleve: 56:12

So it sounds like a you know, an agile and flexible solution, especially for a lot of these agents that are out here that they might not, they don't know what they don't know. And so being able to still have some of the tools that they're comfortable with, but also being exposed to learning about these new systems and processes and ways of thinking that I think just goes back to sort of the the root of all, you know, tech purchases that you should be thinking about before you make that purchase is what do your processes look like? And what do you wish that you didn't have to do? So you can do more of what you want to do? And you don't, you should never make a tech purchase? If you don't know what those processes are in the first place.

Unknown: 56:55

It totally, totally, totally, and you should know what you need. And if there's no need, then there's nothing to reoffer, right, like do we offer so yeah, something that I appreciate from from SBI that we have a lot of good communication with agents and agents can call me, they don't have to go to over like, you know, 20 people to get to the IT department to look at their requirements. And, of course, we have a list of priorities. We have our roadmap and everything, but everyone can just, you know, Hey, so what can we make these happen, you know, this customer has a very specific need, they need a file with this specific format, and they want it and this is my you know, this is if I land this account, will be great for my business. And then so we work closely with agent to make it happen. And if he's very, very, very custom, we don't say, Well, I don't know is to custom is like, we knew he was going to be custom. We are ready for it.

Blythe Brumleve: 58:00

I love that I loved I think that that's a perfect place to sort of land on for a lot of, you know, it prospective agents, you know, incurred agents, you know, just thinking about maybe making the switch or or, you know, maybe they're on the fence right now that, you know, it doesn't have to be a painful process. I think, you know, for most of us that have worked in freight long enough, you've experienced the pain that is a terrible, you know, tech stack onboarding, I once worked for a company that we had a TMS and we integrate it, we spent two years adding it to our system, and it still didn't have an accounting functionality to it. So it was kind of a hot mess. And the company ended up going out of business anyways. So I mean, that was probably you know, the wrong tech investment can be detrimental. And so at least you know, from that angle from a freight agent perspective, you know, your your sounds like everything that SPI is doing is thoughtful, and then also customizable.

Unknown: 58:59

Yes, totally. All right, well, great description. Oh, thank you.

Blythe Brumleve: 59:03

Well, well, where can folks you know, if they're interested if they are one of those folks that are kind of on the fence and you know, thinking about becoming a freight agent, where can folks you know, find out more information I imagine it you know, right there on the website, you know, SPI Any other resources that you think that that would be helpful for a potential freight agent to know

Unknown: 59:25

Yeah, our website SPI three PL look at the numbers. Then I would encourage them to talk with, they can email me personally ie Peralta at SPI Feel free to chat and say hey, I'm interested and you know, we want to have them off the system for sure. Let's set it up. If they want to reach out to Mike Mike Michalik. He will be also happy to so I guess just reach out to us and like you know, our executive team is it He's here to talk with you, we're not going to be hiding or you know. Yeah, we'll I encourage them to email us directly to me or Mike or any on the on the team, I can I can share the emails of Iran,

Blythe Brumleve: 1:00:17

for sure. And I think I'll make it easy for folks, I'll put a put your email address in the show notes. So hopefully, you won't get you know, hit with any kind of, you know, too much spam or anything like that. But just to make it easy for folks out there, I'll put your email in the show notes, as well as the SBI, three website, which should be already in all of our show notes. So any episode that you listened to, when everything is logistics, you'll see them they are a proud sponsor, or I am a I guess a proud podcaster that has the support of SBI. So these shows are definitely brought to you by SBI. But all of the conversation that you heard, and this is as honest and authentic as they come we know we're kind of you know, no no BS around here. So I say I appreciate your insight and kind of entertaining a lot of the different you know, questions that I threw your way this was really insightful conversation and and appreciate your time.

Unknown: 1:01:12

I appreciate it, too. I enjoyed it. And yeah, thank you very much for having me.

Blythe Brumleve: 1:01:16

Absolutely. I hope you enjoy this episode of everything is logistics, a podcast for the thinkers in freight, telling the stories behind how your favorite stuff and people get from point A to B. If you liked this episode, do me a favor and sign up for our newsletter. I know what you're probably thinking, oh God, another newsletter. But it's the easiest way to stay updated when new episodes are released. Plus, we dropped a lot of gems in that email to help the one person marketing team and folks like yourself who are probably wearing a lot of hats at work in order to help you navigate this digital world a little bit easier. You could find that email signup link along with our socials in past episodes. Over at everything is And until next time, I'm Blythe and go Jags

About the Author

Blythe Brumleve
Blythe Brumleve
Creative entrepreneur in freight. Founder of Digital Dispatch and host of Everything is Logistics. Co-Founder at Jax Podcasters Unite. Board member of Transportation Marketing and Sales Association. Freightwaves on-air personality. Annoying Jaguars fan. test

To read more about Blythe, check out her full bio here.