How Freight Brokers Use AI To Work Faster | Episode 350

AI is quickly becoming a real advantage in freight brokerage. We break down practical ways brokers can use it to clean and analyze TMS data, catch accounting errors, improve reporting, summarize claims, support prospecting, search contracts and training materials, and reduce software costs—plus where AI-driven TMS and CRM tools are headed next.

Support Our Sponsors:

Togo: Click Here

OperFi: Click Here

QuikSkope - Get a Free Trial: Click Here

DAT One - Brokers & Carriers: Click Here to get 10% off your first year!

DAT Outgo Factoring for Carriers: Click Here

AscendTMS: Click Here and use promo code RA-freight360! to get AscendTMS FREE for 90 days!

Recommended Products: Click Here
Freight Broker Basics Course: Click Here
Join Our Facebook Group: Click Here
Check out all of our content online: Click Here

Transcript

AI In 2026 And Why It Matters

SPEAKER_01 0:00

We're halfway through twenty twenty six and AI has come a long way. Ben, we just wrapped up a great conversation all about how we're using AI today versus years ago when it first started um rolling out mainstream, but you know, how you can use it in brokerage, whether it's for pricing or, you know, sifting through large amounts of data. It really is a game changer. And if you're not using any sort of AI tools yet, I really think that um, you know, give this sh this episode a listen all the way to the end so you're not missing any of the ideas that we've, you know, put out there for you guys. What are your thoughts?

SPEAKER_00 0:36

I thought it was a really interesting conversation. You and I haven't had a chance to like catch up on the topic. And just in the past few weeks, I found so many more use cases. You and I have found so many more things that have saved like literally hours and in some cases days of time using this tool. And also, like, not only is it saving us tons of time and we're able to get more done, but it's also just way more accurate than it was the last time we had this conversation. So, like, I think this is a really good episode and will be super helpful. Like, literally, no matter what your job is, like, there is something in this episode that should be able to save you time, effort, and headaches. Absolutely.

SPEAKER_01 1:14

All right, let's hop into it.

Summer Talk Plus Daylight Saving Debate

SPEAKER_01 1:16

All right, welcome back. We're up to episode 350 here, a bit of a milestone. Ben, how's things going in your neck of the woods?

SPEAKER_00 1:24

Well, it's hot. It's summer, so it's like I was telling Alex the other day. I'm like, I was telling him about you. I was we were talking about it, and I'm just like, dude, like I feel like summer down here is like winter when I grew up. Like, you don't go out during the middle of the day. You can only go out in the morning or the very end of the day. I could just eat more. I feel like I just get fat and lazy in the summer. When I lived up north, it's like the exact opposite because it's just like 107, 108 real feel every day, and it's just like ridiculous.

SPEAKER_01 1:56

Yeah, quite the opposite. I mean, it's this is this is like the time of year where like I get jealous when I see people outside. Like my my desk faces a window, and I can see the street, and I see people walking around, walking their dogs, running, all that stuff. And I'm you know, I work during the day, right? So it's like I try to take as, you know, whenever I can get take a break or whatnot and just get outside and enjoy it. Because by the time I'm done working, it's like, you know, still warm, but all the fun, sunny, good part of the day is gone.

SPEAKER_00 2:27

Hey, speaking of days being over while you're at work, I heard this morning they might be rolling back daylight savings time. So like that may not exist at a federal level. I think that's what they were talking about on the radio.

SPEAKER_01 2:40

Basically said, like daylight savings was was originally implemented, yeah, to try and like increase the amount of daylight for the agricultural industry. And just with like modern improvements in general, it's kind of an irrelevant thing. Like Arizona just doesn't do it at all. They're just like, yeah, you guys do whatever you want, or we're gonna not change our clocks. And a lot, I think a lot of the is there other parts? I think there's like certain native lands in the states that also don't follow. But either way, yeah. I I feel I feel like I hear that every year. They talk like there's always a bill. Someone in Congress is like, we should get rid of daylight savings time. And they're like, Yeah, we should, and then nothing ever happens. So I don't know. Let's see here. Sports. The uh baseball had the home run derby. I actually didn't watch it this year, but that just happened.

SPEAKER_00 3:32

I heard it was like terrible to watch. I there's a couple of people I talked to said I think Will Farrell did the broadcast, and they said it was just like very hard to listen to.

SPEAKER_01 3:42

He's getting old. It's wild because I like you think about all the hit movies and stuff that he's been in, and he's just aging now. But if you follow the World Cup at all, France got knocked out this week. Spain is I believe yeah, Spain was it two to zero? They knocked out France, so he got the finals coming up this weekend. So I originally had picked like France and Argentina as who I thought would have went, but it appears that is not the case.

SPEAKER_00 4:12

So Yannick Center won Wimbledon.

SPEAKER_01 4:15

That was is that over with now? All men's and women's both? Yep. Okay. Yeah. I I mean, I would I throw I've been throwing whatever's on. I'm in my office. I just kind of just hear it in the background, don't really watch it, but that's good. Yeah, so you got uh England and Argentina at the time of this recording. We don't know that that game is actually gonna happen today. It's Wednesday. So Argentina, the the uh favorite in that one. So news. Let's see here.

MC Numbers For Sale And Risk

SPEAKER_01 4:48

I found this really cool article, and I put it in our newsletter yesterday. There's this data company called Alpha Labs. Did you hear about this? They did like an analysis of all these MC numbers being bought and sold. The guy's name is Matt Fleming. He released it a couple weeks ago. And Alpha Loops is the is the name of the, I guess, the data firm or company. I don't know why it says Alpha Labs in our article, but basically he analyzed like 10,000 Facebook posts, 80,000 telegram messages, and listings from a newly created online marketplace that's just created for selling MC numbers, and there was like almost 4,000 in there. And he did all the math on like what the average price is, but basically it ended up being let's see if I can get the prices here. Yeah, so the the I'll give you the median sale sale price. This is the asking sale price. Obviously, they're gonna negotiate Facebook $20,000, Telegram, $14,000, and marketplace $11,500. So these are literally just the authority. No trucks, no nothing, like no actual hard assets. It's just they're just buying and selling MC numbers. So kind of like shocking. I didn't realize, and I'm I'm also, I mean, I think he kind of broke down the reasoning why the pricing varied. I think it was kind of the the quality of the average buyer was higher on Facebook versus Telegram or this new marketplace. And marketplace, not to be confused with Facebook marketplace. This is like an independent new, new this year, like MC trading selling marketplace. But yeah, 20 grand. And it's had had the the it kind of like how DAT has their high and low threshold. It's that middle 50%. Facebook went as high as 30,000. So 15 to 30,000 was kind of their 25th to 75th percentile on sale price. So people are paying a lot of money because they know how much they can steal with you know, you know, one good, one good hit there. So you get one good clean authority, you go steal a load of you know high value commodity and resell in a black market, you just 10x your money right there. So it's wild to the whole industry, man. Absolutely insane.

Driver Pay Records And Market Shift

SPEAKER_01 7:18

And other news, Tim Haim from um Ascend TMS shared with us a press release that they put out uh related to driver pay. So driver pay reached its highest level on record last month in June. Drivers are now earning roughly 57% more than they did in January of 2020. So that's huge. So keep in mind, Ascent TMS, big partner of the show. You can get 90 days free of their pro version, absolutely free, no credit card required, but you need a referral code, so use the one in the show notes if you want to check that out. Um but yeah, that was pretty crazy because we've been hearing drivers complain about pay and rates for quite a while, and now you're hearing just the opposite that their pay is at a record high level. LTL, they said LTL demand is is higher than it's ever been. A lot of truckload capacity has tightened, and a lot of shippers have routed what they can through the LTL network. So whether it's a volume LTL or if they're breaking shipments up into traditional LTL, you're seeing a lot more go that way. So port volumes are high. We're just kind of we're like in a literally the opposite of where we were 12 months ago. So wild. Oh, good news for the industry. It is, it is absolutely so there's your news. If you guys aren't subscribed to the newsletter, go to the website freight360.net. You can sign up there and we spend it out usually twice a week. And we know you guys are sharing us with your with your colleagues because every time we have a newsletter go out, we usually get like two or three new signups for the newsletter that same day, which means someone probably got it and told their new hire or coworker, hey, you got to sign up for this one. So check it out. Go right to the website, freight360.net. You'll also find all of our other content in there, including the Freight Broker Basics course if you want a full educational option or training curriculum for your team. But yeah,

Trusting AI For TMS Data Moves

SPEAKER_01 9:10

Ben, I think uh we want to go down the some use cases of AI today, huh?

SPEAKER_00 9:15

Yeah, I've been really impressed with how much better it is, even than like three months ago, four months ago. Like much.

SPEAKER_01 9:25

You know what I was thinking about? I'm curious on this because when I was we're going through a TMS integration right now, and one of the big tasks I have to do is I've got to take all of our data from one TMS and we've got to you know load it up on a bunch of spreadsheets to be ported into the new TMS, you know, things like lane history, customers, contacts, locations, factoring companies, all the data, right? And I was thinking to myself, like, what if I just used like if I took a raw data spreadsheet from one TMS, threw it into AI, and had it try to format it to another one, then I was like, then I'm like, there's gonna be, you know, possible like errors and you know, it might not be correct and all this stuff. And so then my big question was like, where is AI now in regards to freight brokering versus human level error? So like what do you trust more? A like an entry-level employee versus AI to get a basic data entry job done. Your personal opinion on where it's at right now. What do you what are your thoughts?

SPEAKER_00 11:35

It's interesting because that's a use case that I really had poor results for sure end of last year. I was doing similar things, what you said. So what I was doing, this is end of last year, was I didn't like the dashboard at all on the TMS for this one company I was involved with. And I'm like, there just wasn't good information. So what I was always doing was literally pulling everything out into Excel, just like you said. But I could get to it, manage it, and configure it faster in Excel and see it better. And there literally weren't even reports that I could get some of this information out of, right? But for me, like I did that a lot. Like all of my 20s, I was doing like spreadsheet modeling when I worked at a bank. Like I just lived in Excel. So like I'm pretty quick with it, and it wasn't that much of a headache. However, I kept testing the models, right? Where I would do it myself and then I would throw the data into like the higher models of GPT at the time, and it just wasn't even close. I'm like, oh, like it is just like absolutely so far off, it's not even usable. And then, because then

AI Finds Duplicates Better Than Humans

SPEAKER_00 12:43

back then, like I was even talking with um Tilo at Levity, and I'm like, there's gotta be like a workaround or a workflow that we could create where like I can have a check itself. And that's kind of how we were trying to solve for again, end of last year. I used it this week and last week on things where it was almost a hundred percent accurate. For example, I had to go into this accounting system because in this client we're transitioning TMSs. So both were integrated during the integration, and one was kind of creating duplicates in the accounting system. And the accounting people were like, well, hey, like I found a duplicate from like a month ago. There were a couple here, a couple there. So before I manually did this with the spreadsheet, I exported everything out of the accounting software, put it into the high model for GPT, which is slower but more accurate, it found every duplicate. Then I manually checked it with my spreadsheet because there's a way to just highlight all the duplicates. And I'm like, it was like a hundred percent right. I did three times this week. It caught every single one of them, did not miss any. I was like shocked.

SPEAKER_01 13:46

That's huge. Yeah. I mean, I remember doing so. If I go back like two years ago, I had a handful of agents that wanted a leaderboard that we could, you know, send out. And I was like, all right, our TMS doesn't have a leaderboard internally. And I was like, well, I can just have a report get scheduled each week, and those that want to participate will have Chat GPT read the report, extract only their numbers, and then create a leaderboard. And I was very detailed about the prompt. And again, this is like two years ago, and it never got it right. It would like get people's numbers wrong, it would include the wrong, you know, brokers, it would skip brokers that were supposed to be in there. And I ended up doing all the corrections myself, and I'm like, I could just do this faster myself. And then you fast forward now. And to your point, I do like how a lot of the AI models have the option, like ChatGBT can have the one that takes longer to respond, but it does a lot of analysis. It like talks to itself and it like tells you what it's trying to do. So I had one, this was actually really cool. This is a little bit different, but similar where it's analyzing data. One of my brokers and myself, we had a customer meeting last week. There's a a shipper that is has been around for years and they are spinning off basically a new entity. Um, it's gonna be its own company, and they're kind of starting from scratch on, you know, transportation providers, et cetera. The owner, his whole thing is like we don't want, you know, going with the good old boys or whoever you've been booking your stuff with for years because there's a there's probably pay-to-play in there or favoritism, et cetera. We want to truly analyze our transportation partners on a lot of different metrics. So for us as a brokerage, it came down to like historically, what have you guys done capacity-wise in certain lanes with certain equipment types? Because obviously we all have access to the spot market, but he wanted to know like realistically, what where do you guys have a good amount of carrier presence and you know, carrier relationships as it applies to it? It was like van and flatbed, and there was like three different outbound

Prompting Tricks For Cleaner Load Reports

SPEAKER_01 15:58

lanes that we looked at. It was like uh we had so we had basically the Southern California regional. So like the pickup and deliveries were within the LA area, and that was for I think it was flatbed. And then we had Baltimore outbound to Ohio and Illinois for van and flatbed, and then we had California outbound to up to the Pacific Northwest to like Oregon and and Washington. And I was like, all right, cool, we can definitely pull this from the TMS and we will, you know, hopefully be able to just run easy reports. And then I find out the TMS, you know, you can pull data out, but I couldn't like just talk to the TMS and just say exactly what I just said to you and have you do it, right? It's like I could run a state to a state or a zip code to a zip code, but I couldn't run that narrative I just gave you. So then I'm like, all right, let me export all the potentially relevant data and I throw it into ChatGPT and I'm literally talking to it, and I hit the dictate button, and I'm like, here's a spreadsheet of all of our historical loads in the last year. I want you to filter this out. And in the first case, I said, I only want you to include loads that have picked up and delivered within 150 miles of Los Angeles, California. And I said, column A is blah, blah, blah, column B is the origin, column C is the date, et cetera. And I went through and just talked to it about what it was, and then it thought for like two minutes and it spit me out this wonderfully like curated load report, and excuse me, it was exactly what I wanted. And I went through and I like I actually like spot checked it, and it was a hundred percent accurate. And I did it for the other two requirements, and I was like, this is nuts. Like two years ago, I couldn't get it to read a basic report. I can get data, and now it's doing an analysis and filtering stuff for me.

SPEAKER_00 17:54

So once one small tip, and then there's a story I'm gonna tell you about that that I ran into last week is what I'm showing some of like the managers in our company is I want them to be able to utilize this for like a bunch of use cases. To get better results from any LLM, you have to ask better questions. And one of the cheat codes I found to ask better questions, use another model to write the prompt for the other one. I'll go to Gemini and be like, write the best prompt to achieve this for Chat GPT, this model. And it gives me the prompt and I get better answers. And sometimes I'll go back and forth two or three times until I really get the right prompt to get, like you said, very good information back, which is another way to just get better results because you get a you get back what you put into it, just like in sales. When you ask better questions, you get better answers, you learn more, right? The other use case that is very similar to what you said that I wanted to bring up. So one of the CEOs for one of our customers is a CFA, which is a chartered financial analyst, which is the people that literally do certifications for business valuations. So they run large models against lots of industry standards, weights, very high-level finance to get to business valuations. Those are the guys that like testify in court for the value of a business, right? They're still subjective, so it's not, you know, totally 100%, but it is like the highest level of financial modeling you can do. I was doing this with him on his company last week, and I was like, hey, Brent, I'm like, can you go grab some of your other models? We'll use one to start with, right? Because building the model out takes time. And when you do this, you have other models, right? So we're in Excel. He showed me this. I didn't even know this existed. There's a Claude plugin for

Excel Plug-Ins Speed Up Modeling

SPEAKER_00 20:41

Excel. And we took an old model and basically gave it all the assumptions that I already had. It just did it correctly. We went and spot checked. There was like a few things that we had to adjust, but I mean, it got 98% of it correct. And it we got this thing done in an hour between the two of us. That would have taken probably two days between two of us back and forth to build a model, get it correct. And I'm like, so there's just a plug-in now in Excel, which means building those spreadsheets out and doing those task things is just kind of like not necessary. And to me, like that is awesome because it brings that to a lower level of skill in modeling. Like, in order to actually do that in Excel, you needed a lot of experience to be able to do this and to make sure it was correct. But now you can literally just ask it questions like you did, and I can get it into Excel, I can get it to adjust Excel. I like blew my mind how fast we were able to do this thing.

SPEAKER_01 21:40

That's pretty sweet. I had a guy, I got a couple other use cases. So this is one that one of our brokers brought to me yesterday, and I was like actually pretty impressed because I was always so basically he's using he's using AI to help him write emails to prospects. And I was like, okay. Like I've I've clearly been sent like hundreds of AI generated emails, and I could tell they're AI generated emails, and I usually just delete them. And but what he did, and he's using seamless.ai, which I've actually never used, but this is what he told me. He sent me a text and he goes, I said, here's a prompt that I'm sending to I'm using AI and seamless that I've been using for a few days. And the prompt was basically like write a short and precise email tailored to this company and about how I can you know meet their needs, blah, blah, blah. And it like, but it's I guess it's a it's AI that's built into seamless.ai, and it's helping him generate kind of cold intro emails that sound like him, that are pulling all the relevant information from that specific company. And then it doesn't just fire them off. He still has to go through and like re but it basically gives him like a 90% solution. Then he can go in and say, All right, I like this, or I want to change that, et cetera, and just help him get you know custom tailored emails out that he would have handwritten himself originally, but now it just kind of does a lot of that basic work for him.

SPEAKER_00 23:07

So that was pretty neat. That is a really cool use case. Here's another one that I had a bunch of stuff that I was having to fix quickly in different places, right? Another thing that I found that is really helpful, right? And you can do this with a project in Chat GPT where you can keep working on something that you have to go back to, right? Or something that's changing. So, like there was a claim I had to work through, and I just didn't have time to read through every email thread to remember what was going on to try to get up to speed to help with this. So if you go in, say you're using Outlook, you can save an email thread, you right-click it or you go to print, and instead of printing it, you hit PDF and it'll create a file for that whole email thread. Then I save them all in a folder, put them into a project, into Chat GPT, and I went, Hey, summarize the communication related

AI For Prospecting Claims And Contracts

SPEAKER_00 24:00

to this claim, what happened, what dates, where, and where is it right now? And I just literally Read it all and in 30 seconds told me what all of these email threads had to do. And then the cool part about the project is when it's not resolved, like there's more email threads that are happening every day and every week. I don't have to read all the emails. Like at the end of the week, I just do the same thing. Go to the email thread, save it, add it to it, and go, where's this claim at right now? And I can manage so much more information in one place. And the other place that I found this is super helpful is there were some law, they weren't necessarily lawsuits, they were like maybe lawsuits, but like dealing with attorneys situations. And if you use Google Notebook LM, this is a really cool feature. You can put up to 50 sources in there. So you can use internet links like URLs or documents. And I don't even think there's a limit on size because the one person I was talking to said he'll do this with books. He gets PDFs for all the books he reads, and he can't remember which book he read something in and when he wants to reference it. So he'll do a folder of like 50 books he read this year on one topic, puts them all into Google Notebook LM, and he can ask a question. Hey, I remember I read this one thing last year. What book was that? What chapter was it? And it'll just find it and answer it. So I was doing this with correspondences with legal agreements. Like I could literally put in all of the legal documents for the situation. I put in all the email threads, and I was able to cross-reference what people were saying versus what was contractually there. Like, say you have a shipper agreement, say you have your carrier agreement, you can put them both in there. You can literally take your entire insurance doc, like these hundred-page docs for your insurance, put that into a notebook, and you title it whatever you want, and you can ask it questions against all three of these documents. Hey, is there anything in my insurance that excludes this? Is there anything in this broker carrier agreement that conflicts with this situation? What is the most likely thing the insurance adjuster is going to say in this situation? What would an attorney say in this? And the prompt I found that was super helpful using other AIs was the prompts start with you are a legal expert that specializes in transportation, read these agreements.

SPEAKER_01 26:14

Buttering up your AI assistant.

SPEAKER_00 26:16

Yeah. But no, but when you give it where it's supposed to, the point of view it should answer from, it will give you different answers. Because then I'll go to the same thing and I'll be like, now you're an experienced CEO in the transportation industry that specializes in third-party logistics. What do you think? And it'll give you two different answers. And then I can work through these things so much faster without having to read and find these things in these documents. It's absolutely wild how much I'm able to get done in things that used to take me forever just by organizing information in there. Here's the other use case I found. We're transitioning a TMS, right? So we have all these call recordings where they're demoing or showing us how to do things. And you remember it when you see it, but some of them were like three weeks ago, some like two months ago, right? Someone asked me a question, how do I do this? Where's this feature? And I'm like, I remember them showing us. And I can't watch three hours of calls to find where this is. I put all the call recordings for that TMS in a notebook LM in one folder. And I just go, ask it the question the person asks me. I'm like, hey, how do they build a load? How do they adjust the BOL to make sure the carrier shows up? How do they and it literally will just find it and then answer it? And I'm like, I can manage so much information in such a short period of time now, just by organizing it in Notebook LM.

SPEAKER_01 27:36

Yeah, that's awesome. I gotta check out Notebook LM. I know you were you you were showing it to me last week. It's pretty sweet. I've been kind of stuck on the chat GPT train for years now.

SPEAKER_00 27:47

Me too. The thing also that I I had to do that I really liked about, and I just started really using this. I knew about it and I never really had enough use cases to do it, but I just it's funny. Like when you have more work to do, like you find a way to use the tools to get more done. So the other thing I found that was super helpful is I really like to listen to things because I can do other stuff, right? So like podcasts, books, mostly I listen to. I'll read a little bit at night, but usually I just fall asleep. So I needed to get up to speed on private equity for this situation, right? Deal structures, what's common, what are pitfalls, what to avoid. So I started with GPT and I went, hey, what podcasts or books should I read to get up to speed on this? And all the all the books it gave me, I'm like, I went to my Audible and I'm like, like, I have those books. I'm like, I read them. I'm like, I can't read six books in three days to like get up to speed on this. So I had it give me all the references. And then I said, I went to ChatGPT and I used deep research and I went, create a report, an extensive, detailed report on the state of private equity right now in this industry. What is the market like? What are buyers saying? What are investors saying, and what are common mistakes and make it detailed, extensive, and check your work, right? Creates this long report, put it in a notebook LM, and then I had to create a podcast for me that was 30 minutes. So I took like a 55-page report, right? Actually, it was

NotebookLM Turns Files Into Answers

SPEAKER_00 29:11

like a hundred pages, put it in a notebook LM. It created like a 30-minute podcast I listened to, and it was literally exactly every question I was trying to refresh on. Is it read to you? Everything, dude. It does the it does a podcast like me and you. There's literally two people talking, and it literally turns this entire document into a conversation that is engaging, relevant. And I'm like, oh, like I remembered all of it, and it took me a half an hour. Like it was wild. Wow.

SPEAKER_01 29:36

So we had one. I'm actually probably due to do this again, but this is a really good kind of an internal audit use case. So if you're in leadership or you're on the you're in part of the business where you're trying to keep your costs under control and make sure you don't have any frivolous spending. So we do, we, you know, the larger you scale your business, you start to find that you're you increase a lot of costs on software licenses for TMS, for loadboards, for you name it, right? Any any kind of tech product that your employees or brokers at et cetera will be using. And if you're on a per person per month payment structure, this is where I found it was very, very helpful for me to audit, you know, who had access to what when the last time they logged in was, who didn't need a certain license, or who had a license that didn't even work with the company anymore. So I did this a couple months ago. And my goal is to do it like maybe, you know, probably quarterly, at least twice a year, is just to make sure we can do all that. So I took, I got reports from all of our vendors on all of our active licenses, the last time the person logged in, what the cost is, and I ran it through and I was like, hey, you know, check all these out. And then I also I fed it, and here's our active list of brokers, and I said, run through all these, find me the ones that stick out because either the the user doesn't work for the company anymore or they haven't logged in in a long period of time, et cetera. And it helped me create this entire license audit, right? And it I still went in and manually double checked the things that it said, like check, check for this, check for this, check for this. But a lot of it I just knew off the top of my head, I'm like, he has a license, he hasn't worked here in two months. Or this person hasn't logged in in, you know, in some cases has never logged in, right? And I'm like, then we need to talk to them and find out like, do they actually need a license? Or did we accidentally, you know, give them a license? And what I found out is in some cases, we knew we were supposed to close the license out, but we forgot to. Um, in some cases, we put the request in to close out, like truck stop was notorious for this. We had found we were getting invoiced, we we had like $4,000 in charges from them for over the past like year at this point for people that we had requested to shut their account down and

License Audits Siri And The Future

SPEAKER_01 31:56

they just never did it. And we were able to use all the documentation, pull up our past emails, and take it to them and get the charges refunded or credited towards our future bill. And I was like, that's insane. That like they and they tried to fight us on it at first. And I'm like, how can you not hold yourself accountable when we did the proper steps to close an account down and you guys didn't actually do it? But in some cases, right, we forgot to do it and we gotta eat that or own that. But it was super helpful because what we used to do is one of the guys that works with me, he used to manually do this every single quarter. Like where he'd be going scrubbing the list one by one, figuring out do they still work here? And then he would be calling them, like, hey, do you still need this account? It shows you haven't logged in in a while, or hey, does employees still work for you? Blah, blah, blah. And now it's like I did all of it in like 10 minutes, instead of like it would take an entire week for a guy to do it in the past, which is the wild part.

SPEAKER_00 32:49

So it's great. It's crazy. Like, I don't know if you saw this. I read this yesterday, but the new version of Siri is out, but I don't think it's like publicly available. It's on certain audio about this. Well, the the person who wrote the article was definitely like one of the guys from like TechCrunch or whatever that gets early access. So, and I can't remember, I think it'll be fully released in the fall to like the applicable models, but he was like demoing it to write articles, right? And it was wild because they he, I mean, the headline is basically the Siri that you were promised years ago, like actually is there. So you he's like, I don't even unlock my phone a lot of time. He's like, here's a couple of the examples. Hey, help me find a birthday present for my dad based on all my text messages with my dad. And it went through the whole text history with his dad and anything that his dad ever mentioned that he thought he was interested in or wanted, gave him a list. Then he was like, Hey, can you find where this is the cheapest? Went and did it. He's like, I didn't even have to unlock my phone. Another one, right? He was like, Hey, can you check on the flight that I emailed about for next month? And it went through all of his emails and found it. And for me, the thing that is interesting is like there's a bunch of lawsuits right now between OpenAI and Apple because OpenAI is apparently gonna be building hardware. And I don't think phones will work the way they used to. I don't think applications will exist the way they used to in a few years. Because you don't need to go to a screen anymore to see some of these things. You're just not going to do that. Because for me, the one I want more than anything, and I'm maybe gonna hack my way into a solution to this in the in the interim, is the search functions in emails are terrible. Like Outlook's search function is absolutely atrocious. I can't find anything in there. I have tens of thousands of emails in different email accounts that are all in Outlook. I can't find anything in there ever. I'm like, I try to search by a keyword, but someone sends me an email and doesn't put a good subject line in it, so I can't find it. If you put in the email address in the domain, it finds like 10% of what somebody sent you and then limits it. But all of your emails, if you use Apple, they're literally not saved in Outlook. They're saved in Apple. And I can't remember if it's it's in Spotlight or whatever. So these files are in my computer, and AI is way better at searching your email to find things than the inherent search bar. And I'm gonna be super excited to be able to just ask it questions to find things in emails instead of searching around for 10 to 15 minutes to be able to find the relevant email that I know somebody sent me last month about a certain thing.

SPEAKER_01 35:28

Yeah, that's pretty cool. It is that time of year when Apple starts to release all their because the stuff usually drops in the fall, right? Like the new iPhone and they'll have all their new new stuff. So here's what I think is really cool too with AI in the TMS. And like you said, you know, going through a TMS transition, one of the, you know, one of the first steps you do is like just a ton of demos, right? And I was really impressed because I did the same thing about five years ago when we looked at an upgrade and AI just wasn't a thing at the time. This was back in like 2021 when we first started looking at stuff. And like really AI blew up probably later that year is when it became a mainstream thing. But now, you know, as we started looking last summer and you know, all the way through, we finally, after about, you know, 10 or 11 months, like settled on what our our our solution was going to be. I was really, really impressed with a lot of the TMSs that the AI functionality is not it's not something that's like just a flashword. Like in some cases, yeah, they just say, like, oh yeah, it's AI. It's like, no, that's just automation and like workflows. That's like not actually intelligence. But kind of like what you just outlined with the new Apple thing where it'll read through your text messages and emails, some of the AI bots that I've seen in these TMS platforms now, you can do that. Like, you can interact and say, like, hey, for example, like I kind of like I talked about earlier, like, look at my load history and conversations with this customer. Where are some areas that we can improve on, you know, pricing or you know, uh reducing deadhead mileage on the carriers that we selected, like things like that. That like that I think is going to be where there's a lot of game-changing and time-saving things for you know, the folks in the trenches in brokerages, not having to do the deep digging and the research on your own, but having it kind of served up to you on a platter of like, here's actual, not just raw data, but here's good intelligence that you can make informed decisions on based on all of the data that you have in your system. And a lot of them would have like email plugins available and you know, access to you know, maybe other data, you know, systems that you might have that would have good information for you. And one of the things I found though is that we had a we had different brokers throughout the company that would, you know, sit in on certain demos and give us their feedback. And we had some that were like, hell yeah, I love this. I can't wait to be able to go from what we're using now to an extremely modern AI-driven TMS now. And then on the other spectrum, I had like I had one guy that was like, Yeah, I'm not like a fan of tech and AI and all this. Like, I just don't trust it. And I'm like, what? Like, whether or not you trust it doesn't mean it's you know, your competitors are gonna be using it. So um I think that's one of the big things is the adoption of how to use it, how to feel comfortable with it, right? Versus, you know, doing something yourself.

SPEAKER_00 38:29

So two one, there's two things I want to make sure I don't forget. One is the CRM, and the other one was the interview I heard with the anthropic CEO. But the CRM, this one I read about this week, is this was an article about basically AI is a threat to like sales force. So smaller and medium-sized companies have been testing and found a lot of value in using an AI model as a CRM. Because if you have it save the memory, you can literally just tell it what you want. Like, hey, I just got off a call with Nate at free 360. He said this, this, and this. I'd like to call him back in like two weeks. This is what I think he needs. So it doesn't have the like scheduling ability, right, to tell you when to call Nate, obviously, but it remembers. So when I go to call you next week, I can be like, hey, what did I want to talk to Nate about and how did the last few calls go? And it will remember and tell you. So as far as like being able to capture information and give it back to you, it's apparently being adopted a lot by smaller medium-sized companies. Interesting. Which I thought was super interesting. Here was the other one you had. The other one was I listened to the Joe Rogan interview with the anthropic CEO. And they were talking about like what does obviously like the the economy look like on the other end of like a lot of these things being adopted in the next few years? And the major point he made is that like we're gonna move to an economy that is mostly favoring people's creativity, right? And curiosity. Because like all of science and all of everything that moves forward are people just asking better questions and being curious and being um creative and how and what you do with it.

SPEAKER_01 40:15

Because real quick, is this the perplexity CEO?

SPEAKER_00 40:19

Sorry, perplexity CEO. Yeah. Are vend Sreninvas? It's a really good interview because what he really talks about is that like if you have the answers to everything at your fingertips, would you really want your kids to be memorizing facts in school for the next 15 years? Like, is that even helpful? And would you want them, would college look at all the way it does now? Because most of it is the regurgitation of information to get a grade that is good or bad. Most of the trust has been eroded in the entire academia, right, over the past few years. Degrees have got more expensive, they become less valuable, they're more diluted. And like, I really been thinking, I mean, because we have kids, and I'm like, yeah, like foundationally, I you need to learn how to read, get better at that. You need to be able to do math, you need to be able to understand the basics. Like, I would, you know, those things I think are still relevant, but I think everything else in school that you're memorizing is honestly just not going to be a thing anymore in 25 years. And I don't even think it's like a helpful skill set. How many people do you talk to that graduate college and you ask them, What did you learn? They're like, I don't remember. Like they it you don't even retain information well in the way it's taught and the way people are tested against it, right? And the entire school system was basically created a hundred years ago so that they could create better workers to go work at big companies that would follow rules. And like that's not really a valuable skill set, and those jobs aren't even around anymore. So, like, I just don't think a lot of how we do everything is gonna exist the way it does when our kids are even in their 20s.

SPEAKER_01 41:56

Yeah. Yeah, it's wild. It's so crazy because I think about like when when did Chat GPT get released?

SPEAKER_00 42:07

I think it was released.

SPEAKER_01 42:08

No, when did when did like the 3.0 or whatever? Like the when like the first 21. Yeah, so like we're talking like less than five years ago, yeah, right. And it is like completely changed the way that people are doing business now. So we'll see. We'll see. That's AI. Let us know if you guys are using. I'm actually curious if you guys leave a comment or send us a message how you're using AI in brokerage, whether it's for you know email management, for rate analysis, or any kind of time-saving automation. Let us know. I'm very curious to see how you guys are using it in addition to the kind of the use cases that we shared today. So anything else on AI?

SPEAKER_00 42:49

I'm sure there is, and I'm sure I'll remember after this. But we covered quite a bit. And again, like it's I am like shocked at how much I'm able to get done on things that used to take me so much longer. Like it was crazy. Like, literally yesterday, somebody was like, Oh, I think there's a bunch of duplicates in this system. And they're like, I'll go when I get an hour or two, I'll go look through it. I was like, wait a minute. I could anything I can export into a file, it took me, I walk, I like literally, it took me less time to walk to my kitchen and back to export this in there, find these duplicates and message them back. And the person was like, wait a minute, how did you find them that quick? Like, it took me like three and a half minutes. And I'm like, that would have taken a human being an hour of arduous, boring work that nobody wants to do. And it was like so much faster. I'm like, oh my god, like the amount of things I'm able to get done at this point is just blowing my mind.

SPEAKER_01 43:41

Yeah. Super crazy.

SPEAKER_00 43:43

Yeah.

SPEAKER_01 43:44

Good stuff, good conversation. Final thoughts, Ben.

SPEAKER_00 43:47

Whether you believe you can or believe you can't, you're right. And until next time, go bills.