Building an AI-Native Freight Forwarder from Zero to Trust

Behind the code.

Behind the code. [music and singing]

Behind the code. [music and singing] That's where the future grows.

That's where the future grows. Welcome to Behind the Code brought to

Welcome to Behind the Code brought to you by Atomic Build. My name is Emit and

you by Atomic Build. My name is Emit and with me today I have Pierre from Agent

with me today I have Pierre from Agent Astra. Thanks for joining me today,

Astra. Thanks for joining me today, Pierre. Super excited about our

Pierre. Super excited about our conversation.

conversation. Thank you for having me.

Thank you for having me. And if you don't mind, just if you can

And if you don't mind, just if you can give a little intro about yourself,

give a little intro about yourself, that'd be great.

that'd be great. Yeah, sure. Well, um, I'm Pierre. I

Yeah, sure. Well, um, I'm Pierre. I think the name gives away that I'm

think the name gives away that I'm French. So, I was born and raised in

French. So, I was born and raised in France. uh currently based out of New

France. uh currently based out of New York and I'm working on Agent Astro

York and I'm working on Agent Astro which is uh an AI native logistics group

which is uh an AI native logistics group starting with uh air freight.

starting with uh air freight. Awesome. And um if I can ask which part

Awesome. And um if I can ask which part of France are you from?

of France are you from? I'm from Paris. I was I was born I grew

I'm from Paris. I was I was born I grew up and went to school in like a onem

up and went to school in like a onem radius for like 17 years. So very very

radius for like 17 years. So very very local Parisian. Yeah.

local Parisian. Yeah. So the biggest you went from the biggest

So the biggest you went from the biggest city in France to the biggest city in

city in France to the biggest city in the United States.

the United States. Yeah, pretty much. I also lived in uh in

Yeah, pretty much. I also lived in uh in London. Uh that's where I studied. And

London. Uh that's where I studied. And then before moving to the US, I was

then before moving to the US, I was actually living in Mexico. Uh so I lived

actually living in Mexico. Uh so I lived in Mexico City for about 5 years.

in Mexico City for about 5 years. Wow. So Spanish, French, English cover

Wow. So Spanish, French, English cover them all?

them all? Yes, correct. Yeah. And and my kid who's

Yes, correct. Yeah. And and my kid who's uh 3 years old speaks well at least

uh 3 years old speaks well at least understands all three. She speaks mostly

understands all three. She speaks mostly Spanish though.

Spanish though. Wow. That's I barely speak English so

Wow. That's I barely speak English so she's doing very well. [laughter]

she's doing very well. [laughter] Your English is better than mine.

Your English is better than mine. So, so um before we get into a lot of

So, so um before we get into a lot of technical depth, if you could had to

technical depth, if you could had to explain agent Astra to a freight

explain agent Astra to a freight forwarder who has never used an AI

forwarder who has never used an AI agent, um what is that like what does

agent, um what is that like what does that explanation look like?

that explanation look like? Yeah, I mean it's it's fairly

Yeah, I mean it's it's fairly straightforward. The way I talk to

straightforward. The way I talk to people in the industry about what we're

people in the industry about what we're doing is that we come in and work at

doing is that we come in and work at companies with the objective to really

companies with the objective to really map the way they work right now and

map the way they work right now and adapt the technology to their operations

adapt the technology to their operations rather than give them like something

rather than give them like something that's out of the box and that they just

that's out of the box and that they just have to figure out how to use. Uh, and I

have to figure out how to use. Uh, and I think that's kind of the beauty with

think that's kind of the beauty with what's happening with, you know,

what's happening with, you know, technology and AI right now is that

technology and AI right now is that there's there's a real uh ability to

there's there's a real uh ability to build and custom build which, you know,

build and custom build which, you know, historically was kind of reserved for

historically was kind of reserved for dev shops and and then like scale uh as

dev shops and and then like scale uh as a as a business model. Uh, but now we're

a as a business model. Uh, but now we're actually able, you know, to from a

actually able, you know, to from a baseline infrastructure really adapt use

baseline infrastructure really adapt use cases and workflows. So again, the way

cases and workflows. So again, the way we work with people is we kind of tell

we work with people is we kind of tell them tell us how you work. you know,

them tell us how you work. you know, we're from the same industry, so we

we're from the same industry, so we already somewhat understand, but there's

already somewhat understand, but there's there might be some specificities that

there might be some specificities that that you have as as a business. Uh, and

that you have as as a business. Uh, and so we can adapt to those on top of like

so we can adapt to those on top of like your sort of like standard forwarding

your sort of like standard forwarding workflows that you know, everyone has in

workflows that you know, everyone has in in this space.

in this space. So like in other words, if they have um

So like in other words, if they have um certain like nuance workflows for

certain like nuance workflows for clients, you can adapt uh agent Astro

clients, you can adapt uh agent Astro around that.

around that. Correct. And and that's actually a big

Correct. And and that's actually a big thing in the space like you know

thing in the space like you know customers have specificities like the

customers have specificities like the carriers you might be working with also

carriers you might be working with also have specificities of like you have to

have specificities of like you have to give me information in a portal or like

give me information in a portal or like I want you to send me an email or I want

I want you to send me an email or I want you to call me up you know like there's

you to call me up you know like there's different like modalities of

different like modalities of communications and we can kind of meet

communications and we can kind of meet people where they want to be met while

people where they want to be met while in or like inhouse uh making it more

in or like inhouse uh making it more efficient for the for the different

efficient for the for the different users. M and just uh coming from the

users. M and just uh coming from the supply chain space as well like how how

supply chain space as well like how how how long does it take you to convert

how long does it take you to convert skeptics which I imagine at first

skeptics which I imagine at first conversation there's some amount of like

conversation there's some amount of like uh this sounds too good to be true

uh this sounds too good to be true I think so it's a great question it's

I think so it's a great question it's interesting because I think like one of

interesting because I think like one of the big changes I've seen with AI is

the big changes I've seen with AI is that people seem to think that finally

that people seem to think that finally technology can actually help them right

technology can actually help them right I think there's kind of this weird thing

I think there's kind of this weird thing where you know system of records have

where you know system of records have been around for a while and you

been around for a while and you uh they weren't ney super good but they

uh they weren't ney super good but they were there but people were very

were there but people were very skeptical about them right it was really

skeptical about them right it was really hard to sell like even sort of like a

hard to sell like even sort of like a canban where you could just move things

canban where you could just move things from one stage to the next like people

from one stage to the next like people would much rather use Excel back then

would much rather use Excel back then but I think the the big change is um

but I think the the big change is um that like people are a lot less

that like people are a lot less skeptical right now and I think maybe

skeptical right now and I think maybe one of the one of the reasons I I would

one of the one of the reasons I I would assume is because like ultimately a lot

assume is because like ultimately a lot of people now use chat GPT even like

of people now use chat GPT even like they use Google right but it's just so

they use Google right but it's just so good compared to what you were just able

good compared to what you were just able to do in a browser before. I think it's

to do in a browser before. I think it's kind of done with a lot of the work of

kind of done with a lot of the work of us like convincing that things can be

us like convincing that things can be done, right? And I think the other thing

done, right? And I think the other thing that's happening is um you hear a lot of

that's happening is um you hear a lot of like press releases or or like sort of

like press releases or or like sort of uh announcements from even like larger

uh announcements from even like larger enterprises saying that you know they've

enterprises saying that you know they've used AI and it's working. Uh and I think

used AI and it's working. Uh and I think that also like does some work in terms

that also like does some work in terms of convincing smaller players that

of convincing smaller players that there's, you know, at least definitely

there's, you know, at least definitely something to look into, right? I think

something to look into, right? I think you obviously still have to pay your

you obviously still have to pay your dues and shows that it's show that it's

dues and shows that it's show that it's working. And I think that's true of any

working. And I think that's true of any tool. Uh but the the bar to people

tool. Uh but the the bar to people believing that it is possible that it

believing that it is possible that it would do that I think has lowered

would do that I think has lowered massively in in the last few months for

massively in in the last few months for sure. that I mean that's a I think

sure. that I mean that's a I think that's probably you can attribute the

that's probably you can attribute the explosion of uh small niche companies

explosion of uh small niche companies fixed like fixing really narrow problems

fixed like fixing really narrow problems in different industries. I think that

in different industries. I think that makes a lot of sense because skepticism

makes a lot of sense because skepticism drives down your entry point is so much

drives down your entry point is so much lower to distribute your product.

lower to distribute your product. Correct. Yeah.

Correct. Yeah. Yeah. So you you worked across um

Yeah. So you you worked across um consulting international trade and then

consulting international trade and then you scaled MetaMap. Um how did you

you scaled MetaMap. Um how did you decide from those experiences to make

decide from those experiences to make the move into freight forwarding? Well,

the move into freight forwarding? Well, so I think it's kind of going back to to

so I think it's kind of going back to to what I I knew as a consultant. So I I

what I I knew as a consultant. So I I worked in in a small boutique firm that

worked in in a small boutique firm that catered almost exclusively to supply

catered almost exclusively to supply chain companies, right? So

chain companies, right? So manufacturers, importers, exporters,

manufacturers, importers, exporters, logistics businesses. Uh and so I did a

logistics businesses. Uh and so I did a lot of work after that. You know, we had

lot of work after that. You know, we had a bunch of international clients. So I

a bunch of international clients. So I traveled, you know, all over the world

traveled, you know, all over the world for for that company, saw like a ton of

for for that company, saw like a ton of different ways of doing business and and

different ways of doing business and and use cases. Uh but back then you know um

use cases. Uh but back then you know um I shaved so I look younger than I am. Uh

I shaved so I look younger than I am. Uh but this is like a good 1015 years ago

but this is like a good 1015 years ago at this point. Um you know a lot of the

at this point. Um you know a lot of the work as a consultant was writing more

work as a consultant was writing more SOPs and checklists right

SOPs and checklists right uh very few projects around like tech

uh very few projects around like tech implementation. You know a lot of the

implementation. You know a lot of the work was again lived in like a kind of

work was again lived in like a kind of standard CRM but like a lot of it was

standard CRM but like a lot of it was also outside right. So I did that and

also outside right. So I did that and then as you said I joined a company

then as you said I joined a company called called MetaMap uh which was

called called MetaMap uh which was focused on KYCML. I think it's one of

focused on KYCML. I think it's one of the first atscale application of machine

the first atscale application of machine learning uh for like a business use case

learning uh for like a business use case which you know was online identity

which you know was online identity verification in this case of just

verification in this case of just background checks in general. Um and you

background checks in general. Um and you know I I think that's kind of where I

know I I think that's kind of where I caught the tech bug right because that

caught the tech bug right because that was for me my first like real tech

was for me my first like real tech experience. Um and you know a lot of the

experience. Um and you know a lot of the work of the work that we were doing back

work of the work that we were doing back then was ultimately automating like

then was ultimately automating like financial institutions back office teams

financial institutions back office teams right and freeing them up from you know

right and freeing them up from you know ultimately a super timeconuming task

ultimately a super timeconuming task that doesn't add a lot of value but as

that doesn't add a lot of value but as but that is a regulatory requirement so

but that is a regulatory requirement so you had to do it uh it just wasn't

you had to do it uh it just wasn't necessarily the best um the best way to

necessarily the best um the best way to spend your time right and I think that

spend your time right and I think that with you know logistics supply chain

with you know logistics supply chain freight foring there's obviously a lot

freight foring there's obviously a lot of that as well where you know you have

of that as well where you know you have to push a lot of paper because you have

to push a lot of paper because you have to uh but is that the best use of your

to uh but is that the best use of your time? Probably not, right? And so after

time? Probably not, right? And so after that metamap experience and when it you

that metamap experience and when it you know when I wanted to start you know

know when I wanted to start you know becoming an an entrepreneur myself, I

becoming an an entrepreneur myself, I figured why not bring back that

figured why not bring back that experience that I had at MetaMap and and

experience that I had at MetaMap and and try to bring it to to supply chain and

try to bring it to to supply chain and and logistics. Uh a lot of hubris as a

and logistics. Uh a lot of hubris as a first-time founder, right? Uh I think uh

first-time founder, right? Uh I think uh when when you work for a company that

when when you work for a company that has strong product market fit, you kind

has strong product market fit, you kind of assume it's easy. Uh, and so I've

of assume it's easy. Uh, and so I've been uh I've been navigating that ever

been uh I've been navigating that ever since.

since. Yeah. I I think that hubris is like a

Yeah. I I think that hubris is like a prerequisite to go off on your own and

prerequisite to go off on your own and build something great.

build something great. For sure. You but you you have to be

For sure. You but you you have to be very comfortable with being wrong.

very comfortable with being wrong. Yes.

Yes. Which is like, you know, it's it's like

Which is like, you know, it's it's like a specific kind of hubris, right? Like

a specific kind of hubris, right? Like you you have to not be too committed to

you you have to not be too committed to your idea, right? Like it's like a fine

your idea, right? Like it's like a fine line between that commitment and just

line between that commitment and just knowing to let go and focus on something

knowing to let go and focus on something else.

else. Yeah. One of the the mantras I always

Yeah. One of the the mantras I always tell myself is like strong beliefs

tell myself is like strong beliefs loosely held.

loosely held. 100%. Yeah.

100%. Yeah. Yeah.

Yeah. Strongly agree. Yeah.

Strongly agree. Yeah. So on on that um just building on that

So on on that um just building on that that question for you is like what is

that question for you is like what is one of the first things um that you were

one of the first things um that you were looking outside in even though you've

looking outside in even though you've been in the space for so long um you

been in the space for so long um you assumed you could automate it and then

assumed you could automate it and then when you actually deployed in real world

when you actually deployed in real world you said to yourself wait a second I

you said to yourself wait a second I have to make quite a bit of adjustments

have to make quite a bit of adjustments for my original hypothesis. Yeah, I

for my original hypothesis. Yeah, I think so. I'm I'm ultimately like a

think so. I'm I'm ultimately like a sales like business guy, right? I'm not

sales like business guy, right? I'm not I'm not a tech guy. Like obviously I

I'm not a tech guy. Like obviously I vibe code a lot more right now. Uh and I

vibe code a lot more right now. Uh and I send like little widgets to my team and

send like little widgets to my team and they tell me this is great, but like

they tell me this is great, but like behind it it's [ __ ] right? Like the the

behind it it's [ __ ] right? Like the the infrastructure is terrible. But but when

infrastructure is terrible. But but when I when I started this like it wasn't a

I when I started this like it wasn't a thing. It was really just me uh try to

thing. It was really just me uh try to try to figure out what to what to sell

try to figure out what to what to sell and what to build. I think there's a lot

and what to build. I think there's a lot of advice out there. or one of him that

of advice out there. or one of him that really resonates to me even though it's

really resonates to me even though it's easier said than done is like try to

easier said than done is like try to sell before building and so you know I

sell before building and so you know I spent a fair amount of time um and kind

spent a fair amount of time um and kind of using my network and people I knew

of using my network and people I knew from past experiences um trying to

from past experiences um trying to figure out like what would be useful for

figure out like what would be useful for these companies right um and you know I

these companies right um and you know I also wasn't

also wasn't married to a specific ICP at least

married to a specific ICP at least initially and so I I tried a bunch of

initially and so I I tried a bunch of different spaces different focus like we

different spaces different focus like we we we did importers exporters like

we we did importers exporters like specific industries We looked at

specific industries We looked at agriculture for a while which which I

agriculture for a while which which I thought was super interesting. Um you

thought was super interesting. Um you know eventually kind of moved on more to

know eventually kind of moved on more to like a logistical focus. Um but you know

like a logistical focus. Um but you know initially I think you kind of want to

initially I think you kind of want to try as much as you can to figure out a

try as much as you can to figure out a product before building it right because

product before building it right because you kind of have to understand the pain

you kind of have to understand the pain point and you want to make sure that

point and you want to make sure that this is a real painoint and you're not

this is a real painoint and you're not wasting your time especially in my case

wasting your time especially in my case where you know at least initially I

where you know at least initially I couldn't build anything myself right and

couldn't build anything myself right and so you know that meant like bringing on

so you know that meant like bringing on people that you potentially have to pay

people that you potentially have to pay if they're not co-founders so there was

if they're not co-founders so there was a cost to sort of launching a product uh

a cost to sort of launching a product uh and so you kind of want to do that with

and so you kind of want to do that with at least like some some confidence that

at least like some some confidence that like there's someone out there that

like there's someone out there that would use it. Um, so we did that and uh

would use it. Um, so we did that and uh and I think you know there was a big

and I think you know there was a big shift uh with this whole like chat GPT

shift uh with this whole like chat GPT moment where it kind of went from like

moment where it kind of went from like we're clearly building or trying to

we're clearly building or trying to displace like historical system of

displace like historical system of record maybe doing more like vertical

record maybe doing more like vertical SAS to like hey no like we can do you

SAS to like hey no like we can do you know something completely different

know something completely different which is more workflow based automation

which is more workflow based automation you know freeing up people's time etc.

you know freeing up people's time etc. Um, [snorts]

Um, [snorts] and what was interesting though is I I

and what was interesting though is I I think you you kind of realize that uh

think you you kind of realize that uh you are it's really hard to like even if

you are it's really hard to like even if it's clear that you're automating time

it's clear that you're automating time away from people to make have that an

away from people to make have that an impact on bottom line is a is a is a

impact on bottom line is a is a is a real hard one. Uh and so you know t

real hard one. Uh and so you know t still fighting that that upward battle.

still fighting that that upward battle. Uh but I think it was like kind of um

Uh but I think it was like kind of um you know a whole phase of like trying to

you know a whole phase of like trying to look for for the product market fit and

look for for the product market fit and what we actually wanted to do as a

what we actually wanted to do as a company and I think ultimately you want

company and I think ultimately you want to spend as much time as possible

to spend as much time as possible talking to potential customers trying to

talking to potential customers trying to figure out what the real pain points are

figure out what the real pain points are and then from that like start building

and then from that like start building products that hopefully they like and

products that hopefully they like and and end up using right but there's a lot

and end up using right but there's a lot of like top of funnel work before

of like top of funnel work before anything. Uh and then you know the hard

anything. Uh and then you know the hard part is also pushing for adoption right

part is also pushing for adoption right like the tool is there uh people now

like the tool is there uh people now have to use it how do you push for that

have to use it how do you push for that yeah I think uh just pulling from some

yeah I think uh just pulling from some of the things you said there's a lot of

of the things you said there's a lot of great insights there one thing that came

great insights there one thing that came to mind is that when you automate these

to mind is that when you automate these workflows and you mentioned this earlier

workflows and you mentioned this earlier now there's an unlock of time to apply

now there's an unlock of time to apply yourself somewhere else but I think the

yourself somewhere else but I think the first time that unlock comes um some

first time that unlock comes um some companies and some staff don't actually

companies and some staff don't actually know where to apply themselves and

know where to apply themselves and there's some period perod of time in

there's some period perod of time in which they're in this ambiguous zone

which they're in this ambiguous zone trying to figure that out.

trying to figure that out. Yeah. I think it's not sometimes, it's

Yeah. I think it's not sometimes, it's every single time, right? I think

every single time, right? I think but and I think actually if I was to

but and I think actually if I was to give any advice to anyone who wants to

give any advice to anyone who wants to to start a business um you know that is

to start a business um you know that is focused on automating something. I think

focused on automating something. I think you want to focus then the mistake I

you want to focus then the mistake I personally made for instance is assuming

personally made for instance is assuming that just taking the process where it's

that just taking the process where it's very consuming and you know you can

very consuming and you know you can automate a lot of it assuming that's

automate a lot of it assuming that's like the biggest pain point or that is

like the biggest pain point or that is what's going to drive adoption is often

what's going to drive adoption is often time not right just because of what you

time not right just because of what you just said right like you free up all

just said right like you free up all that time that doesn't mean they're

that time that doesn't mean they're going to do anything with it

going to do anything with it right uh I think often time you know

right uh I think often time you know what ends up happening is it's also like

what ends up happening is it's also like people are not rolling their thumbs in

people are not rolling their thumbs in an office, right? Like they're swamped

an office, right? Like they're swamped to start with. And so what ends up

to start with. And so what ends up happening is the quality of the work is

happening is the quality of the work is actually better, but all that time that

actually better, but all that time that you've saved doesn't really convert to

you've saved doesn't really convert to anything on the P&L, right? You're still

anything on the P&L, right? You're still very much like a cost to to these

very much like a cost to to these people.

people. Uh and that's something I I I heard I

Uh and that's something I I I heard I learned the hard way where, you know, it

learned the hard way where, you know, it was very clear that we were automating

was very clear that we were automating and saving a lot of time, right? and and

and saving a lot of time, right? and and we could measure it and it was measured

we could measure it and it was measured in like FTDs, right? It wasn't like I'm

in like FTDs, right? It wasn't like I'm saving you like two three hours here and

saving you like two three hours here and there. It's like, you know, this is like

there. It's like, you know, this is like a full-time employee and and you're

a full-time employee and and you're still a relatively small business. So,

still a relatively small business. So, it is uh substantial, but you know,

it is uh substantial, but you know, owners would just tell you and operators

owners would just tell you and operators would just say that person is still

would just say that person is still there. He's still doing the same thing.

there. He's still doing the same thing. Right. Right.

Right. Right. Um and so I think you kind of have to

Um and so I think you kind of have to really find something where it's not

really find something where it's not about time saving, but it has like an

about time saving, but it has like an impact on the P&L, whatever that is.

impact on the P&L, whatever that is. Right. So it could be, you know, working

Right. So it could be, you know, working on on on cash flow. It could be working

on on on cash flow. It could be working on top of funnel, bringing you business,

on top of funnel, bringing you business, recuperating business. I think that

recuperating business. I think that typically are um places where you create

typically are um places where you create value from day one for these businesses.

value from day one for these businesses. And so then they're more willing to use

And so then they're more willing to use and it just makes more sense uh even

and it just makes more sense uh even from just a a pricing model perspective

from just a a pricing model perspective often time rather than say like hey you

often time rather than say like hey you know I know you spent 10 hours on this I

know I know you spent 10 hours on this I can take that time away but

can take that time away but realistically no guarantee they do

realistically no guarantee they do anything with that freed up time. Yeah,

anything with that freed up time. Yeah, I think I think like you touched on um

I think I think like you touched on um perhaps like the outcomedriven pricing

perhaps like the outcomedriven pricing model which I think quite a few like

model which I think quite a few like companies are experimenting

companies are experimenting experimenting with um because like

experimenting with um because like you're right it's it's if you're not

you're right it's it's if you're not impacting P&L then it's almost abstract

impacting P&L then it's almost abstract until someone someone fully relies on it

until someone someone fully relies on it then when you remove it I think the

then when you remove it I think the abstract part is gone. Um and and

abstract part is gone. Um and and building on that like question for you

building on that like question for you in that space is where do you think like

in that space is where do you think like the what's the threshold where it's not

the what's the threshold where it's not an assist and it's genuinely doing a

an assist and it's genuinely doing a piece of the job that was done before by

piece of the job that was done before by a person? I I mean I think it's it's

a person? I I mean I think it's it's it's almost like a product development

it's almost like a product development question like I think you oftent times

question like I think you oftent times start as like a co-pilot somehow on like

start as like a co-pilot somehow on like a specific task and then you have like a

a specific task and then you have like a human in the loop component right

human in the loop component right because when you start you want to make

because when you start you want to make sure that the work is still done

sure that the work is still done correctly

correctly uh and then over time you probably start

uh and then over time you probably start having you I I think the the phasing is

having you I I think the the phasing is you have a human in the loop you know

you have a human in the loop you know initially the human often has to correct

initially the human often has to correct something that correction at some point

something that correction at some point gets to basically it's always approved

gets to basically it's always approved by the the operator and so at that point

by the the operator and so at that point you can just say well I can put this

you can just say well I can put this part in photo but I think it's like you

part in photo but I think it's like you you really divide like a workflow like

you really divide like a workflow like in in very specific steps to get there

in in very specific steps to get there and so I think eventually you

and so I think eventually you potentially have a an agent that does

potentially have a an agent that does like a whole set of tasks that could

like a whole set of tasks that could ultimately be attributed it to what you

ultimately be attributed it to what you know uh an operator would be doing. Uh

know uh an operator would be doing. Uh but I think it's it's a phase thing and

but I think it's it's a phase thing and there's a lot of chat around like you

there's a lot of chat around like you know that that's that already exists. I

know that that's that already exists. I don't think that's really true. Uh

don't think that's really true. Uh because I think there's a clear adoption

because I think there's a clear adoption problem. There's still like a lot of

problem. There's still like a lot of risk in terms of accuracy hallucination.

risk in terms of accuracy hallucination. You know harness is obviously getting

You know harness is obviously getting way better. uh but you know I think some

way better. uh but you know I think some in some cases specifically if you have

in some cases specifically if you have you're implementing AI in um in

you're implementing AI in um in workflows where there's a direct impact

workflows where there's a direct impact on P&L that's often like fairly risk

on P&L that's often like fairly risk risky as well right it's kind of

risky as well right it's kind of correlated and so the risk of getting it

correlated and so the risk of getting it wrong is actually fairly high uh and so

wrong is actually fairly high uh and so you want to avoid that being true um but

you want to avoid that being true um but yeah honestly I don't know where where

yeah honestly I don't know where where the switch happens it's still so early

the switch happens it's still so early uh in that like I think right now what

uh in that like I think right now what I'm seeing is just people being more

I'm seeing is just people being more efficient, being able to do more, but

efficient, being able to do more, but you know, it's not like the uh

you know, it's not like the uh ultimately even like the tools that we

ultimately even like the tools that we build, they still look like a dashboard,

build, they still look like a dashboard, right? Like it's not like a first,

right? Like it's not like a first, right? Um and I I don't know that people

right? Um and I I don't know that people really have figured out this this UX UI

really have figured out this this UX UI yet either, right? I know that some

yet either, right? I know that some companies have some interesting ideas,

companies have some interesting ideas, but it's early, right? TBD if that's

but it's early, right? TBD if that's like the way forward.

like the way forward. Yeah. And I think for for in your like

Yeah. And I think for for in your like in in supply chain and logistics, the

in in supply chain and logistics, the leap to agent first uh UIUX seems too

leap to agent first uh UIUX seems too big of a leap to make to actually like

big of a leap to make to actually like transition um operators and like I think

transition um operators and like I think to your point a dashboard is familiar

to your point a dashboard is familiar even though if it's agent powered. Um I

even though if it's agent powered. Um I think that is the right transitionary at

think that is the right transitionary at least UIUX to to deploy.

least UIUX to to deploy. Yeah. And I honestly do not know what

Yeah. And I honestly do not know what the next stage is right like right now

the next stage is right like right now it seems to be working that way. uh you

it seems to be working that way. uh you kind of have to meet people where they

kind of have to meet people where they are, what they're used to. It's already

are, what they're used to. It's already a fairly big leap from, you know, a lot

a fairly big leap from, you know, a lot of people use historically Excel and

of people use historically Excel and Google Sheets. Uh very little actually

Google Sheets. Uh very little actually use their system of record consistently,

use their system of record consistently, right? They use it for data entry, but

right? They use it for data entry, but often times what you still see is I've

often times what you still see is I've done all the work outside of my system

done all the work outside of my system of record and when I'm closing that, you

of record and when I'm closing that, you know, shipment in this case, I'm

know, shipment in this case, I'm entering the data at the end in in the

entering the data at the end in in the system record. And so that's already a

system record. And so that's already a big leap to say like well you know

big leap to say like well you know forget these sort of like siloed uh data

forget these sort of like siloed uh data centers and you know move onto a

centers and you know move onto a platform that's already a big one. I

platform that's already a big one. I think also in supply chain it's it's

think also in supply chain it's it's probably one of the few ecosystems where

probably one of the few ecosystems where there's so many people and so many cooks

there's so many people and so many cooks at the table right cooks in the kitchen.

at the table right cooks in the kitchen. And so

And so it it makes it and both internally and

it it makes it and both internally and externally by the way, right? Like it's

externally by the way, right? Like it's not only just like the different people

not only just like the different people you're interacting at the business you

you're interacting at the business you work for, it's also all the different

work for, it's also all the different interested parties you're communicating

interested parties you're communicating with. Um I think that creates a lot of

with. Um I think that creates a lot of complexity in terms of how do you net

complexity in terms of how do you net build systems that talk to each other.

build systems that talk to each other. Uh because at least until everyone is

Uh because at least until everyone is willing to communicate that way, right?

willing to communicate that way, right? like you're still sort of handcuffed to

like you're still sort of handcuffed to what the other person is doing, right?

what the other person is doing, right? And if the other person is, you know,

And if the other person is, you know, tech forward, that's great. If they're

tech forward, that's great. If they're not, you know, you kind of have to work

not, you know, you kind of have to work with it.

with it. Um because on the service level, they

Um because on the service level, they might just be amazing, right? And I

might just be amazing, right? And I think ultimately,

think ultimately, you know, we're still very much in a

you know, we're still very much in a space where service is more important

space where service is more important than the tech itself, right? Like, you

than the tech itself, right? Like, you know, people will tell you they could

know, people will tell you they could just be operating on pen and paper. as

just be operating on pen and paper. as long as the service is excellent, I

long as the service is excellent, I don't really care, right? I would that

don't really care, right? I would that vastly outweighs um you know my decision

vastly outweighs um you know my decision in terms of who I choose to work with vs

in terms of who I choose to work with vs like you know whatever tool I use. Yeah,

like you know whatever tool I use. Yeah, I think like that touches on I guess

I think like that touches on I guess exception handling which is like a a

exception handling which is like a a normal instance in logistics and um on

normal instance in logistics and um on that topic like today if um an agent is

that topic like today if um an agent is wrong like what have you thought about

wrong like what have you thought about to uh design some of the level of of

to uh design some of the level of of trust because you're handling real

trust because you're handling real shipments, customers etc. Um what

shipments, customers etc. Um what happens in those instances? Yeah, I

happens in those instances? Yeah, I think right now the the way we handle it

think right now the the way we handle it is that as little as possible agents

is that as little as possible agents don't do any work with customers

don't do any work with customers directly, right? And so there's always

directly, right? And so there's always like a verification there.

like a verification there. Um they're they're more like words

Um they're they're more like words facing at this point. Um you know there

facing at this point. Um you know there are adwords facing on some specific

are adwords facing on some specific things but it's it's really just like

things but it's it's really just like automated updates which by the way

automated updates which by the way there's no AI behind it, right? It's

there's no AI behind it, right? It's it's just a timer of like sending an

it's just a timer of like sending an update or whatever it is. So there's no

update or whatever it is. So there's no risk there whatsoever.

risk there whatsoever. Yeah.

Yeah. Um but but a lot of it is still very

Um but but a lot of it is still very much inwards facing um and and you know

much inwards facing um and and you know on the inwards facing there is a lot of

on the inwards facing there is a lot of exception handling. There are some

exception handling. There are some mistakes that are being made. I think at

mistakes that are being made. I think at that point it's about building systems

that point it's about building systems that register those those issues and uh

that register those those issues and uh learn from them. Right?

learn from them. Right? And by the way the learning you know

And by the way the learning you know some of it is automated. The vast

some of it is automated. The vast majority right now is still manual. And

majority right now is still manual. And what I mean by manual it's like we see

what I mean by manual it's like we see what the the exception was uh and we

what the the exception was uh and we tackle it from a product perspective

tackle it from a product perspective right uh there's obviously you know we

right uh there's obviously you know we have uh agents like surfacing issues

have uh agents like surfacing issues like uh compiling data for us you know

like uh compiling data for us you know looking for patterns etc but you know

looking for patterns etc but you know they're not really fixing systems on

they're not really fixing systems on their own at least at least right now

their own at least at least right now right like we're pushing in that

right like we're pushing in that direction for for specific topics

direction for for specific topics um but but I think a lot of it is still

um but but I think a lot of it is still you very human driven uh because the

you very human driven uh because the decision making is is is well I call it

decision making is is is well I call it like tribal knowledge like a lot of that

like tribal knowledge like a lot of that information is in the operator's head

information is in the operator's head and so it's also really hard to just say

and so it's also really hard to just say well I'm just going to trust an AI agent

well I'm just going to trust an AI agent how regardless of how like imagine even

how regardless of how like imagine even like a really good one like top tier you

like a really good one like top tier you know bestin-class I don't think they

know bestin-class I don't think they have all the information because I know

have all the information because I know where that information lives and it's

where that information lives and it's not in data it's in like that guy's head

not in data it's in like that guy's head so unless that guy has a neural Lincoln

so unless that guy has a neural Lincoln plant. Uh kind of hard to take that

plant. Uh kind of hard to take that decision without him, right? Um and so

decision without him, right? Um and so it's a learning process. I think you

it's a learning process. I think you accumulate data, you accumulate like

accumulate data, you accumulate like patterns as a result. And then from

patterns as a result. And then from there, you know what's true is that

there, you know what's true is that there's more and more agentic workflows

there's more and more agentic workflows in what we do, but it's still very

in what we do, but it's still very limited. It's still very controlled. Um

limited. It's still very controlled. Um and I I know maybe I'm I'm like a

and I I know maybe I'm I'm like a control freak. I know that some people

control freak. I know that some people are way more like open about it, but I

are way more like open about it, but I think specifically for supply chain,

think specifically for supply chain, there's so much gray. Uh, which is by

there's so much gray. Uh, which is by the way why RPA never really worked,

the way why RPA never really worked, right? Because that's also not new

right? Because that's also not new technology.

technology. Um, I think we're still like in the

Um, I think we're still like in the initial phases of of figuring out where

initial phases of of figuring out where where and and how to build correctly on

where and and how to build correctly on a lot of these different topics. You you

a lot of these different topics. You you mentioned earlier just like um the work

mentioned earlier just like um the work is siloed and fragmented and then

is siloed and fragmented and then eventually it hits the system of record.

eventually it hits the system of record. Is what was one of the driving reasons

Is what was one of the driving reasons or is that the driving reason why you

or is that the driving reason why you built a new system of record instead of

built a new system of record instead of just building a layer on top of a TMS or

just building a layer on top of a TMS or or other

or other tunes? Yeah. So, so just to be clear

tunes? Yeah. So, so just to be clear right now we do still rely on on TMS's

right now we do still rely on on TMS's right. So we've you know I think

right. So we've you know I think eventually what we've built looks like

eventually what we've built looks like an AI first system of record. um which

an AI first system of record. um which you know it's it's kind of like a

you know it's it's kind of like a context graph with a UX UI layer on top

context graph with a UX UI layer on top of it right and then connected to

of it right and then connected to different data sources that enriches it

different data sources that enriches it and makes it more efficient and surfaces

and makes it more efficient and surfaces improvements uh potential improvement on

improvements uh potential improvement on its own you know you start getting into

its own you know you start getting into predictive modeling uh as well in the

predictive modeling uh as well in the future uh but because we're starting we

future uh but because we're starting we still very much rely on on system of

still very much rely on on system of records uh there's a couple reasons for

records uh there's a couple reasons for that like um I would say the main one is

that like um I would say the main one is often times they're connected to

often times they're connected to regulatory databases.

regulatory databases. Uh and so you kind of still have to go

Uh and so you kind of still have to go through through that. Uh and the other

through through that. Uh and the other reality is that again it's it's a very

reality is that again it's it's a very social ecosystem, right? Like I'm still

social ecosystem, right? Like I'm still very much um you know it's not like um

very much um you know it's not like um you know when you work with your lawyer

you know when you work with your lawyer for instance, you have a one-to-one

for instance, you have a one-to-one relationship with your lawyer, right? Uh

relationship with your lawyer, right? Uh so your lawyer sends you something, you

so your lawyer sends you something, you send something back, you know that's

send something back, you know that's that's kind of it. in in our space, you

that's kind of it. in in our space, you know, there's five to 10 different inter

know, there's five to 10 different inter interested parties when I'm moving a

interested parties when I'm moving a shipment,

shipment, right?

right? And they all have different tools,

And they all have different tools, different systems, etc. So, you have to

different systems, etc. So, you have to reconciate all that. And the reality is,

reconciate all that. And the reality is, you know, there's like a system called

you know, there's like a system called EDI. You probably heard of it. Like a

EDI. You probably heard of it. Like a lot of people are connected to each

lot of people are connected to each other with EDI, right? And I just can't

other with EDI, right? And I just can't rip that off uh on on day one uh because

rip that off uh on on day one uh because uh you need the business to keep

uh you need the business to keep operating, right? So you have to

operating, right? So you have to compromise uh at least until you get to

compromise uh at least until you get to a point where you're in that position to

a point where you're in that position to to rip things off um without impacting

to rip things off um without impacting the business. I think that's ultimately

the business. I think that's ultimately what the most important thing is. And

what the most important thing is. And obviously with you know infinite

obviously with you know infinite resources in terms of you know manpower,

resources in terms of you know manpower, money uh and time you know you can build

money uh and time you know you can build something and rip off everything on day

something and rip off everything on day one. Uh but you know resources are

one. Uh but you know resources are limited by nature. Uh and so these

limited by nature. Uh and so these things take time.

things take time. Yeah. And I think like um maybe this is

Yeah. And I think like um maybe this is well understood or maybe it's not like

well understood or maybe it's not like these systems are entrenched and

these systems are entrenched and standards for a reason like they've

standards for a reason like they've worked for decades even though they uh

worked for decades even though they uh might be perceived as like slow or

might be perceived as like slow or outdated or very manual. Um so yeah I

outdated or very manual. Um so yeah I think your your insight there is the

think your your insight there is the right one where you have to work in

right one where you have to work in parallel until there's some moment where

parallel until there's some moment where maybe um complete replacement is

maybe um complete replacement is actually viable

actually viable and it's a maybe and and I think you

and it's a maybe and and I think you have again it's it's kind of that that

have again it's it's kind of that that founder hubris right like I am going to

founder hubris right like I am going to replace and displace the the historical

replace and displace the the historical system and record that might be your

system and record that might be your vision and mission

vision and mission uh you also have to be okay with the

uh you also have to be okay with the fact that you most probably actually

fact that you most probably actually have to envision a world where you want

have to envision a world where you want to be very successful even if you depend

to be very successful even if you depend on them and you actually can't displace

on them and you actually can't displace them, right?

them, right? Um so you know it's it's always that

Um so you know it's it's always that fine line between hubris and and

fine line between hubris and and reality, right? Like what's uh where

reality, right? Like what's uh where where you move and that line moves as as

where you move and that line moves as as you grow as a as a company,

you grow as a as a company, right?

right? Um I think for me right now there is a

Um I think for me right now there is a reason why these companies are so

reason why these companies are so entrenched. They're doing some things

entrenched. They're doing some things incredibly well. Mhm.

incredibly well. Mhm. Um and and I don't think we need to

Um and and I don't think we need to necessarily completely displace them to

necessarily completely displace them to be successful ourselves, right? Um and

be successful ourselves, right? Um and by the way, we're in no position to do

by the way, we're in no position to do it today anyways, right? So [laughter]

it today anyways, right? So [laughter] we have to work towards that.

we have to work towards that. Mhm. Yeah. I think like uh like to your

Mhm. Yeah. I think like uh like to your point like resources and and time and

point like resources and and time and people are like the biggest factor and I

people are like the biggest factor and I think in the logistics space I'm sure

think in the logistics space I'm sure you've seen uh Happy Robot has had like

you've seen uh Happy Robot has had like a a monster um just level of success and

a a monster um just level of success and and implementation. I mean they're

and implementation. I mean they're making like movie grade trailers now

making like movie grade trailers now about their product which is like quite

about their product which is like quite an interesting thing to see. Um but

an interesting thing to see. Um but still I think uh adoption and having you

still I think uh adoption and having you know voice agents handle all these tasks

know voice agents handle all these tasks in real world operations I think uh

in real world operations I think uh trucking companies still might often ask

trucking companies still might often ask to be looped into a human operator on on

to be looped into a human operator on on that side. So um to your point I think

that side. So um to your point I think the vast majority of existing systems

the vast majority of existing systems are still being touched more than 50% of

are still being touched more than 50% of the time. So um yeah sorry go ahead.

the time. So um yeah sorry go ahead. Yeah. And I think the unique insight

Yeah. And I think the unique insight from Happy Robot and it's actually

from Happy Robot and it's actually incredibly smart is that the agent is

incredibly smart is that the agent is ultimately a purchasing agent, right? In

ultimately a purchasing agent, right? In the sense that like you're asking for a

the sense that like you're asking for a price, you are asking for a rate, you

price, you are asking for a rate, you are uh asking for a availability and so

are uh asking for a availability and so the other person on the line on the

the other person on the line on the other end of the line, which you know

other end of the line, which you know often times is still very much a human,

often times is still very much a human, they still have an incentive to answer

they still have an incentive to answer you.

you. Right. Right.

Right. Right. Yeah. Yeah.

Yeah. Yeah. Because they're selling. I think the

Because they're selling. I think the reason why this model works so well is

reason why this model works so well is because it's it's a purchasing engine

because it's it's a purchasing engine ultimately,

ultimately, right? And it's like an update engine as

right? And it's like an update engine as well where where you know, but it kind

well where where you know, but it kind of goes hand in hand in in sort of like

of goes hand in hand in in sort of like if you were using that same technology

if you were using that same technology regardless of how good it is to sell

regardless of how good it is to sell something. Um I think that would be a

something. Um I think that would be a lot harder, right? Like if and you know

lot harder, right? Like if and you know there's like I'm sure you get called by

there's like I'm sure you get called by like random AI voice agents on a fairly

like random AI voice agents on a fairly right and and you hang up, right? Uh,

right and and you hang up, right? Uh, and so I think like they they've really

and so I think like they they've really understood where this technology fits

understood where this technology fits well in their workflow and what works.

well in their workflow and what works. But by the way, like I don't know if

But by the way, like I don't know if you've seen, but they also expanded to

you've seen, but they also expanded to other industries, right? They don't do

other industries, right? They don't do just freight anymore.

just freight anymore. Um, I think also what they're realizing

Um, I think also what they're realizing is that it's not necessarily a big

is that it's not necessarily a big enough space uh for them to reach the

enough space uh for them to reach the numbers they want to reach on those sort

numbers they want to reach on those sort of processes alone. So, they're getting

of processes alone. So, they're getting more horizontal, but that's also because

more horizontal, but that's also because it's it's a I mean, if I was I think

it's it's a I mean, if I was I think their pitch initially was,

their pitch initially was, and don't hold me up to it, I'm

and don't hold me up to it, I'm paraphrasing, but you know, we're going

paraphrasing, but you know, we're going to build the the agentic voice for for

to build the the agentic voice for for like freight, right?

like freight, right? And I think right now their pitch is

And I think right now their pitch is probably much closer to uh we're going

probably much closer to uh we're going to build the the agentic voice for

to build the the agentic voice for procurement, right? which which is

procurement, right? which which is probably which is what actually enables

probably which is what actually enables you to um expand to other industries

you to um expand to other industries because procurement is not just in

because procurement is not just in logistics it's kind of everywhere which

logistics it's kind of everywhere which I think is just super smart um and

I think is just super smart um and expands their time and that's also where

expands their time and that's also where they raised a lot of money because I

they raised a lot of money because I think investors are seeing that like the

think investors are seeing that like the potential for uh the size they're going

potential for uh the size they're going to reaches is a lot bigger now

to reaches is a lot bigger now right yeah I mean the founders must have

right yeah I mean the founders must have had some early insight naming their

had some early insight naming their company as generic as possible happy

company as generic as possible happy robot so it's quite quite clever yeah

robot so it's quite quite clever yeah you can do everything yeah for Yeah,

you can do everything yeah for Yeah, it's true. Um, so you know, some of this

it's true. Um, so you know, some of this what we've been talking about so far is

what we've been talking about so far is like the importance in supply chain

like the importance in supply chain where relationships will become or or

where relationships will become or or continue to be a driver of business

continue to be a driver of business success even when um agents are

success even when um agents are increasingly deployed. A question I had

increasingly deployed. A question I had for you is that what does that mean in

for you is that what does that mean in practice to develop or strengthen your

practice to develop or strengthen your relationship building skills in this

relationship building skills in this space if you're you're an operator?

space if you're you're an operator? Well, I think like the way I think about

Well, I think like the way I think about it is in a world where everything is

it is in a world where everything is automated, like what still matters,

automated, like what still matters, right? What do we do as people? Um, and

right? What do we do as people? Um, and I think that is relationships, right?

I think that is relationships, right? Uh, the the example I I like uh and it's

Uh, the the example I I like uh and it's very personal because it's what I

very personal because it's what I believe for my kids. For instance, I I

believe for my kids. For instance, I I do not know what school and university

do not know what school and university looks like in 15 years, which is when

looks like in 15 years, which is when she's going to be going through that.

she's going to be going through that. And that's scary, right? Um, and and for

And that's scary, right? Um, and and for me, what I believe is true is I I want

me, what I believe is true is I I want her to be able to speak as many

her to be able to speak as many languages as possible because for me

languages as possible because for me personally, that's something that's a

personally, that's something that's a skill that I've always valued because I

skill that I've always valued because I think that's like a great unlock to

think that's like a great unlock to understanding someone else's background,

understanding someone else's background, culture, and creating a relationship.

culture, and creating a relationship. Uh, and I believe that's ultimately

Uh, and I believe that's ultimately going to be true in the future. And and

going to be true in the future. And and the way I describe it is if you assume

the way I describe it is if you assume that tomorrow you get live translation

that tomorrow you get live translation through your AirPods, right? So like

through your AirPods, right? So like you're talking to me in a different

you're talking to me in a different language, I understand you instantly in

language, I understand you instantly in I don't exactly know how that would work

I don't exactly know how that would work because there's always some of some sort

because there's always some of some sort of a delay. But anyways, assuming that

of a delay. But anyways, assuming that works, you know, we will obviously be

works, you know, we will obviously be able to understand each other, right?

able to understand each other, right? Like that's the massive unlock in

Like that's the massive unlock in business. That's huge, right? M

business. That's huge, right? M uh but I think you know let's say you

uh but I think you know let's say you know your your language is is is uh is

know your your language is is is uh is English and mine is French you will

English and mine is French you will value the effort I make to speak English

value the effort I make to speak English to you um by learning English much more

to you um by learning English much more than just me relying on the translation.

than just me relying on the translation. Sure. Yeah.

Sure. Yeah. Right.

Right. Yeah. And and I think that's kind of the

Yeah. And and I think that's kind of the key thing with relationships is again in

key thing with relationships is again in a world where everything is AI um I

a world where everything is AI um I think one clear component where people

think one clear component where people will be important is building

will be important is building relationships and that will ultimately

relationships and that will ultimately drive you know who you decide to work

drive you know who you decide to work with who you decide to interact with um

with who you decide to interact with um etc etc because there's a world where

etc etc because there's a world where because everything is so automated the

because everything is so automated the service level is very similar from one

service level is very similar from one to the next right and so how do you make

to the next right and so how do you make that decision um it's in the

that decision um it's in the relationship that you build with someone

relationship that you build with someone and the trust that you build with that

and the trust that you build with that person. Um, so yeah, that's that's kind

person. Um, so yeah, that's that's kind of way I think about it. And so, you

of way I think about it. And so, you know, relation like logistics was always

know, relation like logistics was always an incredibly relationship driven space,

an incredibly relationship driven space, right? Um, you know, you send business

right? Um, you know, you send business to the guy that you've known for like 20

to the guy that you've known for like 20 years and when he moves to the next

years and when he moves to the next company, uh, you work with him and you

company, uh, you work with him and you continue and you bring your business

continue and you bring your business over.

over. Um, I I think that's still going to be

Um, I I think that's still going to be true in the future regardless of how

true in the future regardless of how good technology gets.

good technology gets. Yeah. No, I I think you also um might

Yeah. No, I I think you also um might have given uh at least uh one answer to

have given uh at least uh one answer to what someone can do with um an increased

what someone can do with um an increased amount of time is invest in language

amount of time is invest in language skills.

skills. Yeah, 100%.

Yeah, 100%. Yeah, I think that's uh I think that

Yeah, I think that's uh I think that like especially um in there's so many

like especially um in there's so many Spanish speakers in the United States um

Spanish speakers in the United States um and then depending where you're

and then depending where you're operating in the world like

operating in the world like understanding Cantonese, Mandarin, like

understanding Cantonese, Mandarin, like there's so many language skills that

there's so many language skills that keep you get an instant bond when

keep you get an instant bond when someone who's Yeah. So I think that's

someone who's Yeah. So I think that's and at least from my experience speaking

and at least from my experience speaking the language it's not just about

the language it's not just about learning the language like you learn

learning the language like you learn just a lot more through that process

just a lot more through that process about the person's culture the way they

about the person's culture the way they think cuz a lot of like the mentality

think cuz a lot of like the mentality behind how people think based on like is

behind how people think based on like is present in the way they speak and their

present in the way they speak and their language right like the concepts and the

language right like the concepts and the language

language um you know for instance I think English

um you know for instance I think English has way more words than French um and

has way more words than French um and and so that obviously has an impact in

and so that obviously has an impact in terms of how you think about things and

terms of how you think about things and how you think about the world. So your

how you think about the world. So your ability to speak both

ability to speak both um enables you to understand better how

um enables you to understand better how the other person thinks rather than just

the other person thinks rather than just like a translator. Right.

like a translator. Right. Right. Yeah. Yeah. Yeah. Yeah. It's it's

Right. Yeah. Yeah. Yeah. Yeah. It's it's true. I mean like English language has

true. I mean like English language has phrases like um don't beat around the

phrases like um don't beat around the bush. But but I mean like that's most of

bush. But but I mean like that's most of of sales is speaking in long poetic

of sales is speaking in long poetic terms instead of a very concise direct

terms instead of a very concise direct message. And um I learned that lesson

message. And um I learned that lesson working with folks out of Germany like

working with folks out of Germany like use use much less words to convey your

use use much less words to convey your message. So uh that's a very strong

message. So uh that's a very strong point as well. Um

point as well. Um well and also don't beat around the bush

well and also don't beat around the bush if you don't speak English and it's an

if you don't speak English and it's an auto translation. You will not

auto translation. You will not understand what that means, right? It's

understand what that means, right? It's just like okay [laughter]

just like okay [laughter] what are you talking about?

what are you talking about? That's true. That's true. Like you might

That's true. That's true. Like you might think this person's coming to attack my

think this person's coming to attack my garden. Um so another question I had for

garden. Um so another question I had for you was around um what like just

you was around um what like just building your company. I I guess like

building your company. I I guess like the question I had is so you started

the question I had is so you started like you said find out uh what the

like you said find out uh what the market thinks about your idea before you

market thinks about your idea before you build around the product. Um at that

build around the product. Um at that time like were you considering uh like

time like were you considering uh like using an external team temporarily to

using an external team temporarily to get like when you started to get good

get like when you started to get good traction from the market or how did you

traction from the market or how did you view that as a founder? You mean in

view that as a founder? You mean in terms of hiring people to help you build

terms of hiring people to help you build on the tech side?

on the tech side? Yeah. Yeah.

Yeah. Yeah. Well, for for me, I had to like I had

Well, for for me, I had to like I had like close to zero technical skills.

like close to zero technical skills. Um but I think I was lucky enough that

Um but I think I was lucky enough that because I had been in the ecosystem

because I had been in the ecosystem before, I knew some people, I knew some

before, I knew some people, I knew some like, you know, really good engineers.

like, you know, really good engineers. Uh and so that kind of leaped that gap

Uh and so that kind of leaped that gap for me, right? But but this is again

for me, right? But but this is again this is before the chat GPT moment right

this is before the chat GPT moment right like if I had started after that it my

like if I had started after that it my founder journey would have looked very

founder journey would have looked very different right uh I think if I was if I

different right uh I think if I was if I was starting from something from scratch

was starting from something from scratch I would still go through the the process

I would still go through the the process of like talking to customers and trying

of like talking to customers and trying to understand like what their pain

to understand like what their pain points are and spend a lot of time on

points are and spend a lot of time on that. Uh but but I think and I know I

that. Uh but but I think and I know I would I would go probably in in that at

would I would go probably in in that at least that second meeting with you know

least that second meeting with you know uh a vibe coded prototype that I could

uh a vibe coded prototype that I could show uh and and you know talk about it

show uh and and you know talk about it and and actually interestingly enough I

and and actually interestingly enough I think a lot of time when you do these

think a lot of time when you do these discovery meetings you show prototypes I

discovery meetings you show prototypes I mean back then you showed like flat

mean back then you showed like flat prototypes of like hey this is

prototypes of like hey this is potentially what it would look like

potentially what it would look like right and this is where the button would

right and this is where the button would be and you would be able to move it

be and you would be able to move it around like I used to do that right and

around like I used to do that right and I The big difference now is you probably

I The big difference now is you probably would get into these conversations at

would get into these conversations at least in the second one with working

least in the second one with working products almost or or something that

products almost or or something that looks like a working product, right?

looks like a working product, right? Like I don't know if you remember but

Like I don't know if you remember but Figma uh before like you would use but

Figma uh before like you would use but it was like pure design, right? But you

it was like pure design, right? But you could click on buttons and it I I used

could click on buttons and it I I used to do that, right? And so it looked like

to do that, right? And so it looked like a product.

a product. Totally.

Totally. But now like but now it's it's it's a

But now like but now it's it's it's a work in product to some extent, right?

work in product to some extent, right? You could give it to them. it probably

You could give it to them. it probably breaks like fairly quickly, but you

breaks like fairly quickly, but you know, you in theory they could start

know, you in theory they could start using it, putting data in there, uh, and

using it, putting data in there, uh, and it could work, right? Um, and so I think

it could work, right? Um, and so I think that that would make a big difference,

that that would make a big difference, but to to your question, back then I

but to to your question, back then I didn't have a choice, right? So I kind

didn't have a choice, right? So I kind of had to partner with, uh, with people.

of had to partner with, uh, with people. Uh, but but I think now that the the

Uh, but but I think now that the the level of excuse is lower, right? like

level of excuse is lower, right? like the the barrier to entry uh you know I

the the barrier to entry uh you know I think if you want to be a founder and

think if you want to be a founder and you want to start specifically a

you want to start specifically a technology founder right because there's

technology founder right because there's millions of other types of founders but

millions of other types of founders but you know you you can just invest in

you know you you can just invest in yourself spend a lot of time on figuring

yourself spend a lot of time on figuring out how to use these tools uh and start

out how to use these tools uh and start building on your own um to get like that

building on your own um to get like that initial traction

initial traction uh and then eventually you know you you

uh and then eventually you know you you want a team because I think it's

want a team because I think it's honestly your role as as sort of the

honestly your role as as sort of the founder

founder um to to to build the product, right? At

um to to to build the product, right? At least if you're more like a business

least if you're more like a business founder um and your your role is more

founder um and your your role is more sort of like the vision and and making

sort of like the vision and and making sure that you have the adequate

sure that you have the adequate resources to to sell that vision, etc.

resources to to sell that vision, etc. Yeah. Yeah. No, that makes sense. And

Yeah. Yeah. No, that makes sense. And and given that like you started just

and given that like you started just before um which which we correctly

before um which which we correctly labeled like the chat GPT LLM explosion

labeled like the chat GPT LLM explosion moment. Um, now that you're past that

moment. Um, now that you're past that moment, are you designing your team to

moment, are you designing your team to be small like not needing to expand into

be small like not needing to expand into Okay. Yeah, go ahead.

Okay. Yeah, go ahead. Yeah, I think if if you're able to and

Yeah, I think if if you're able to and and it's it's hard, right? Because it

and it's it's hard, right? Because it also depends on your network, uh, who

also depends on your network, uh, who you know, etc. But again, back to the

you know, etc. But again, back to the relationship thing, like you kind of

relationship thing, like you kind of want to invest in your network as early

want to invest in your network as early as possible anyways. Uh, and I think in

as possible anyways. Uh, and I think in as much as possible, not in transa

as much as possible, not in transa transactional way. um you want to build

transactional way. um you want to build like really good relationships. But to

like really good relationships. But to your question in terms of how I think

your question in terms of how I think about the team, uh yeah, I want it to be

about the team, uh yeah, I want it to be much smaller because you just know that

much smaller because you just know that you can do a lot more. Um I think, you

you can do a lot more. Um I think, you know, you want to work with excellent

know, you want to work with excellent people and that could take that could

people and that could take that could mean different things, right? Depending

mean different things, right? Depending on on what they're doing. uh and then

on on what they're doing. uh and then you want them to be as enabled as

you want them to be as enabled as possible with the right tools and and

possible with the right tools and and you you probably end up spending a lot

you you probably end up spending a lot on tools as a result but their output is

on tools as a result but their output is 10x right so so it's still like uh

10x right so so it's still like uh incredibly uh valuable and I think for

incredibly uh valuable and I think for for instance you know when I was working

for instance you know when I was working at MetaMap we had a a pretty significant

at MetaMap we had a a pretty significant uh sales team like it it was a it was a

uh sales team like it it was a it was a big team right um and I think you know

big team right um and I think you know it was kind of the the time that we were

it was kind of the the time that we were living in uh it was what you needed uh

living in uh it was what you needed uh from a resource perspective to reach the

from a resource perspective to reach the objectives we wanted to reach. Um it

objectives we wanted to reach. Um it would look very different if we were

would look very different if we were doing it today, right? Just because we

doing it today, right? Just because we live in a different time where I believe

live in a different time where I believe you can do a lot more with a lot less.

you can do a lot more with a lot less. And that's true from tech, but that's

And that's true from tech, but that's also true for anything, right? I think

also true for anything, right? I think it's not just limited. I think for tech

it's not just limited. I think for tech it's it's very clear because engineers,

it's it's very clear because engineers, you know, have that mentality of like

you know, have that mentality of like using the new tool out there and and

using the new tool out there and and really like making the most out of it.

really like making the most out of it. It's kind of in their DNA, right? That's

It's kind of in their DNA, right? That's what they are engineers. Uh and so that

what they are engineers. Uh and so that adoption is is very quick and you see

adoption is is very quick and you see how you know they're producing 10x more

how you know they're producing 10x more than they were producing before. Um but

than they were producing before. Um but I think it's it's true for everyone,

I think it's it's true for everyone, right?

right? Yeah. I I think like um you to to your

Yeah. I I think like um you to to your point like so much of uh the convergence

point like so much of uh the convergence that's happening between product

that's happening between product engineering and sales is in the

engineering and sales is in the imagination of a person and having like

imagination of a person and having like this outlet to place your imagination

this outlet to place your imagination that can quickly spin up a real world um

that can quickly spin up a real world um prototype or or working product. Maybe

prototype or or working product. Maybe the backend's not there yet. But I

the backend's not there yet. But I remember that time too where you're

remember that time too where you're sitting and clicking through Figma and

sitting and clicking through Figma and making all the touch points that way you

making all the touch points that way you want your prototype to perform in a

want your prototype to perform in a meeting. So um I think like the

meeting. So um I think like the imagination is the unlock where before

imagination is the unlock where before maybe it was restricted because you had

maybe it was restricted because you had to go and build these teams but that's

to go and build these teams but that's removed. Um and and on that like just

removed. Um and and on that like just building on that a question for you. So

building on that a question for you. So if I'm a freight forwarder and I

if I'm a freight forwarder and I register my business today and I'm not

register my business today and I'm not going to try to use any legacy

going to try to use any legacy technology or workflows or processes

technology or workflows or processes aside from what's necessary. Um what

aside from what's necessary. Um what would you build differently from day

would you build differently from day one?

one? That's a good question. uh like if I was

That's a good question. uh like if I was starting a freight forer from scratch

starting a freight forer from scratch like what would I build first? I don't

like what would I build first? I don't know that I would build anything. I I

know that I would build anything. I I think it really depends on what you're I

think it really depends on what you're I think I would look for what I believe

think I would look for what I believe are are good enough tools to uh operate

are are good enough tools to uh operate my business

my business that ultimately are good enough and and

that ultimately are good enough and and also like cheap enough. Cheap is not the

also like cheap enough. Cheap is not the word but more so like they're not going

word but more so like they're not going to eat away at my margins. And then I

to eat away at my margins. And then I would really invest in figuring out well

would really invest in figuring out well I would invest in two things. I would

I would invest in two things. I would invest in market data right so and that

invest in market data right so and that potentially you want to build some tools

potentially you want to build some tools around that like top of funnel uh tools

around that like top of funnel uh tools uh and then I would really invest in uh

uh and then I would really invest in uh figuring out what I can really solve for

figuring out what I can really solve for my customers that you know their current

my customers that you know their current uh provider isn't solving for. But I

uh provider isn't solving for. But I think it kind of goes back to you know

think it kind of goes back to you know spending time with customers to

spending time with customers to understand what you should build from a

understand what you should build from a tech perspective and then like building

tech perspective and then like building that very similar I would on the service

that very similar I would on the service layer very similar approach uh you know

layer very similar approach uh you know I would just spend time with shippers

I would just spend time with shippers trying to understand like what's painful

trying to understand like what's painful for them right now um and then and and

for them right now um and then and and actually that would probably drive

actually that would probably drive things that I would build that I don't

things that I would build that I don't think is currently covered by you know

think is currently covered by you know the tools I bought in the market but I

the tools I bought in the market but I would first like use of the shelf tools

would first like use of the shelf tools to get the business running, right? Like

to get the business running, right? Like you need that. Uh and then I would go

you need that. Uh and then I would go out, spend a lot of time with with

out, spend a lot of time with with shippers, try to understand what's

shippers, try to understand what's painful for them, and then try to solve

painful for them, and then try to solve for that.

for that. Uh and it could just be that it's just

Uh and it could just be that it's just like a a service offering, right? It's

like a a service offering, right? It's not even like a tech offering

not even like a tech offering potentially.

potentially. Right. Right.

Right. Right. Yeah. I mean, it also Sorry, go ahead.

Yeah. I mean, it also Sorry, go ahead. Yeah. Because at the you know, the

Yeah. Because at the you know, the bottom line here is like what matters is

bottom line here is like what matters is your service much more than your tech,

your service much more than your tech, right? So the tech is at the service of

right? So the tech is at the service of at Yeah. that the tech is at the service

at Yeah. that the tech is at the service of your service. It's it's repetitive,

of your service. It's it's repetitive, but it's anecdotal, right? Like what

but it's anecdotal, right? Like what really matters is um the the level of

really matters is um the the level of service you bring to people.

service you bring to people. Yeah, I think you I think you just uh

Yeah, I think you I think you just uh made it I just you accidentally just

made it I just you accidentally just gave a great quote that the tech is at

gave a great quote that the tech is at the service of your service. That is

the service of your service. That is like fantastically true. Um, and it's

like fantastically true. Um, and it's probably missed still by a large vast

probably missed still by a large vast majority of people building tech today.

majority of people building tech today. Yeah. And I think it's because of the

Yeah. And I think it's because of the nature of what it means to be uh, you

nature of what it means to be uh, you know, a tech software vendor.

know, a tech software vendor. Right.

Right. Right. Because you have to you

Right. Because you have to you ultimately to to make money, you have to

ultimately to to make money, you have to sell a fairly similar system to a large

sell a fairly similar system to a large group of people

group of people that on paper look similar, but in their

that on paper look similar, but in their day-to-day are actually very different.

day-to-day are actually very different. Right. And so you often

Right. And so you often miss on, you know, whatever the 20% that

miss on, you know, whatever the 20% that makes the business special,

makes the business special, but it's hard as a as a software vendor,

but it's hard as a as a software vendor, you know, tech company to to cater to

you know, tech company to to cater to that 20%. Because,

that 20%. Because, you know, can you cater to the 20%

you know, can you cater to the 20% variance to everyone uh in in like your

variance to everyone uh in in like your your market? Like, you know, probably

your market? Like, you know, probably with AI like you get closer to bridging

with AI like you get closer to bridging that gap, but I think it's still there,

that gap, but I think it's still there, right? And so that's ultimately why you

right? And so that's ultimately why you you fall short on, you know, building

you fall short on, you know, building techn.

Yeah. No, it's a good line. I think you

should. Yeah. Yeah. It's uh it should be

should. Yeah. Yeah. It's uh it should be added to the um startup poster wall, you

added to the um startup poster wall, you know, like there's always Yeah. Yeah. Um

know, like there's always Yeah. Yeah. Um just moving into uh like to I have a few

just moving into uh like to I have a few like rapid questions, not necessarily to

like rapid questions, not necessarily to do um with ancient Astro, just general.

do um with ancient Astro, just general. Uh first one is uh which business leader

Uh first one is uh which business leader has influenced you and and how you think

has influenced you and and how you think about building companies and why.

about building companies and why. Oh wow. Uh so so many I have it's funny

Oh wow. Uh so so many I have it's funny like I'm I'm rereading right now um the

like I'm I'm rereading right now um the hard thing about hard you can't see it.

hard thing about hard you can't see it. The hard thing about hard things.

The hard thing about hard things. Y

Y um so I guess Ben Horowitz uh it's

um so I guess Ben Horowitz uh it's interesting because the my my kid has

interesting because the my my kid has put like a lot of stickers on on the

put like a lot of stickers on on the book right now. Uh, and so I think it's

book right now. Uh, and so I think it's like the new version of also why am I

like the new version of also why am I doing the hard thing about hard things,

doing the hard thing about hard things, right? It's because now I have a family

right? It's because now I have a family and and we're working towards that,

and and we're working towards that, right?

right? But, uh, yeah, I guess like Ben Horitz

But, uh, yeah, I guess like Ben Horitz is is a big one. But there there's like

is is a big one. But there there's like so many out there, right? Um, you know,

so many out there, right? Um, you know, that are very inspirational. For sure.

that are very inspirational. For sure. Yeah. I I mean that's a if you if you're

Yeah. I I mean that's a if you if you're out there and you haven't read that yet,

out there and you haven't read that yet, you should read it. It's a great book.

you should read it. It's a great book. Yeah. Yeah.

Yeah. Yeah. 100%. It's a It's no [ __ ] This is

100%. It's a It's no [ __ ] This is hard. It might not work.

hard. It might not work. just know that it might not work. I

just know that it might not work. I think uh that's that's a that's a pretty

think uh that's that's a that's a pretty big one actually when when you start as

big one actually when when you start as an entrepreneur,

an entrepreneur, right? Yeah, agreed. Um and so outside

right? Yeah, agreed. Um and so outside of uh building your company, what helps

of uh building your company, what helps you quiet your mind?

you quiet your mind? It's so easy. Um my wife and kid uh you

It's so easy. Um my wife and kid uh you know I I'm lucky that I started on this

know I I'm lucky that I started on this journey uh with uh we were already

journey uh with uh we were already married uh with my wife. uh that has

married uh with my wife. uh that has like always centered me. I think also

like always centered me. I think also the advantage is she works for a large

the advantage is she works for a large tech company.

tech company. Uh she's been working in tech for you

Uh she's been working in tech for you know 12 plus years at this point. So she

know 12 plus years at this point. So she also really understands it and so I

also really understands it and so I think you know she also works very long

think you know she also works very long hours so do I and so like it's a good

hours so do I and so like it's a good match and so I think uh she's really my

match and so I think uh she's really my sounding board. I hope I am also

sounding board. I hope I am also sometime hers. Uh so that helps and and

sometime hers. Uh so that helps and and I think you know we've had a chat now

I think you know we've had a chat now for for a few years. Um it it's what's

for for a few years. Um it it's what's great about that is that like children

great about that is that like children have a rhythm, right? So if you want to

have a rhythm, right? So if you want to spend time with your child, which I do,

spend time with your child, which I do, um you have to follow that rhythm. And

um you have to follow that rhythm. And so that has been like a really big one

so that has been like a really big one for me where uh you know, wake up in the

for me where uh you know, wake up in the morning. Uh well, I have to go wake her

morning. Uh well, I have to go wake her up, get her ready, go to school. Uh and

up, get her ready, go to school. Uh and you know, obviously I work throughout

you know, obviously I work throughout the day. Uh and then you know, she she

the day. Uh and then you know, she she uh she thankfully has a nanny after

uh she thankfully has a nanny after school. uh because I can't stop working

school. uh because I can't stop working at at 4 p.m. Uh but you know when it

at at 4 p.m. Uh but you know when it gets closer again to um to like bedtime,

gets closer again to um to like bedtime, this is where again I take over and and

this is where again I take over and and I do that. So I think that also has

I do that. So I think that also has added like that that nice rhythm to to

added like that that nice rhythm to to uh to my entrepreneurial journey. So

uh to my entrepreneurial journey. So there's nothing as like you know I wake

there's nothing as like you know I wake up and I go to the steam room or like I

up and I go to the steam room or like I I work out or whatever it is. It's it's

I work out or whatever it is. It's it's like a family driven uh rhythm but uh it

like a family driven uh rhythm but uh it works super well for me. I think it's

works super well for me. I think it's it's it's refreshing. It also at least

it's it's refreshing. It also at least personally um kind of tells you why you

personally um kind of tells you why you do these things, right? And and why

do these things, right? And and why you're you're you're working towards

you're you're you're working towards what I'm working towards.

what I'm working towards. Yeah. No, I think that's amazing. I

Yeah. No, I think that's amazing. I mean, I have a very similar family

mean, I have a very similar family configuration to yours and um you get

configuration to yours and um you get into this very serious mindset and then

into this very serious mindset and then your three-year-old says, "Let's have a

your three-year-old says, "Let's have a dance party." And you just instantly

dance party." And you just instantly break out of that and then switch back

break out of that and then switch back in. So,

in. So, 100%.

100%. Yeah. That's awesome. Um, so what does

Yeah. That's awesome. Um, so what does success look like for you ultimately

success look like for you ultimately with Agent Astro?

with Agent Astro? Yeah, it's kind of a joke. Um,

Yeah, it's kind of a joke. Um, I I've I've been saying it a lot in the

I I've I've been saying it a lot in the last few last few weeks because I've

last few last few weeks because I've been asked this question by a bunch of

been asked this question by a bunch of people. Um, I think for me what success

people. Um, I think for me what success looks like is when antitrust agencies uh

looks like is when antitrust agencies uh come up to us and try to break us apart

come up to us and try to break us apart because we've created monopolies. I

because we've created monopolies. I think that's a funny one. Um, and what I

think that's a funny one. Um, and what I like about that one specifically is that

like about that one specifically is that it doesn't set a number. I'm I'm not fan

it doesn't set a number. I'm I'm not fan of setting numbers because I think it's

of setting numbers because I think it's it's kind of like it kind of caps you in

it's kind of like it kind of caps you in in your ambition.

in your ambition. But the entry trust one I think is uh

But the entry trust one I think is uh does that and and is also a funny one,

does that and and is also a funny one, right? Like so you know that's that's my

right? Like so you know that's that's my objective. I think I'll tell the team we

objective. I think I'll tell the team we are starting to be successful because

are starting to be successful because like antitrust people are looking at

like antitrust people are looking at what we're doing.

what we're doing. That's a very uh that's a very that's a

That's a very uh that's a very that's a very interesting way to answer the

very interesting way to answer the question. I really appreciate that

question. I really appreciate that because I think that is like the you're

because I think that is like the you're so big that regulators are finally

so big that regulators are finally watching you and decided that you cannot

watching you and decided that you cannot be any bigger. Um so that's a cool way

be any bigger. Um so that's a cool way to approach it. I like that. Um okay. I

to approach it. I like that. Um okay. I mean uh that's a wrap. Pier, like this

mean uh that's a wrap. Pier, like this was awesome. I really appreciated

was awesome. I really appreciated chatting with you. Thank you for joining

chatting with you. Thank you for joining me today.

me today. No, thank you for having me.

No, thank you for having me. Behind the [music]

Behind the [music] behind the

behind the one [music] thing.

one [music] thing. Yeah, we build we build we build it real

Yeah, we build we build we build it real behind [music] the