Ep. #69: Harish Abbott, CEO & Cofounder, Augment

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Andrew Silver (00:00):
Hey FreightPod listeners.
Before we get started today,let's do a quick shout out to
our sponsor, rapido SolutionsGroup.
Rapido connects logistics andsupply chain organizations in
North America with the best nearshore talent to scale
efficiently and deliver superiorcustomer service.
Rapido works with businessesfrom all sides of the logistics
industry.
This includes brokers, carriersand logistics software

(00:23):
companies.
Rapido builds out teams withroles across customer and
carrier, sales and support, backoffice administration and
technology services.
The team at Rapido knowslogistics and people.
It's what sets them apart.
Rapido is driven by an insideknowledge of how to recruit,
hire and train within theindustry and a passion to build
better solutions for success.

(00:44):
The team is led by CEO DannyFrisco and COO Roberto Lacazza,
two guys I've worked with frommy earliest days in the industry
at Coyote.
I have a long history with themand I trust them.
I've even been a customer oftheirs at Molo and let me tell
you they made our businessbetter.
In the current market, whereeveryone's trying to do more
with less and save money,solutions like Rapido are a

(01:07):
great place to start To learnmore.
Check them out at gorapidocom.
That's gorapidocom.

(01:32):
Welcome back to another episodeof Freepod.
I'm your host, andrew Silver,joined today by a special guest.
Per usual, always a specialguest on the show.
Mr Harish Abbott, did I getyour name right?
It's perfect, perfect.
Welcome to the show, mr HarishAbbott.
Did I get your?

Harish Abbott (01:46):
name right, it's perfect.

Perfect.
Welcome to the show, sir.
How are you doing?
I'm doing great.
Thank you for having me.
So I think you've got one ofthe most successful backgrounds
of any of the guests.
I've had Multiple companiesfounded and sold, and the most
recent one for a big number.
But you're back at it, back inthe action, with your newest

(02:09):
business Augment.
As you say, the logisticsindustry.
Logistics is broken, I think,is the tagline on your LinkedIn
page.
Let's start right there.
How is logistics broken?
Why is it broken?
Let's start right there.
How is logistics broken?
Why is it broken?
You seem like a seeminglylifelong logistics guy with your

(02:33):
businesses.

So I'd love to hear your perspective there.
Yeah Well, first, I'm not sureabout being the most successful,
probably being the most lucky,I'll take that.
Most lucky, I'll take that.
Listen, logistics is, I mean,we all know it's the backbone of
sort of everything.
Right, like everything youtouch, the coffee mug on your

(02:55):
table 20 different companies,probably coordinated from Asia
to Chicago for that coffee mugto be on your table, and it's
complex, you know, it's ships,it's trucks, it's warehouses.
But if you think about you knowit's somewhat like these 20

(03:19):
companies, for them to get thatcoffee mug on your table have to
trade information and tradegoods.
Right, whether it's themanufacturer or the distributor,
or the freight forwarder, orthe shipper or the port on this
side, the crosstalk, or thetrucking company, they all have
to trade information.

(03:40):
It's like, hey, here's a billof lading, here's the container
number, here's the trailernumber on which this truck is,
this thing is at, here's the youknow bill of goods or things
like that.
And you know, the trading ofinformation today happens
through the lowest commondenominator Like you've lived it
right Like emails, phone calls,text, telegram, whatsapp, and

(04:05):
it happens very asynchronously.
So operators across these 20companies are trading
information asynchronously overemails and phone calls.
All of those methods are lowbandwidth methods.
They don't contain a lot ofthey cannot.
They're just designed to besmall bits of information that

(04:25):
can be traded.
So first, it overwhelms theoperators, right, like they have
to deal with hundreds of emails.
You know, the other day I wasat one of the brokerages and I'm
not kidding, the operator had800 unread emails, right.
So if you come in the morningand your inbox says 800 unread
emails like that's not a good inthe morning, and your inbox
says 800 unread emails, likethat's not a good sign, that's

(04:47):
not an industry that is, likeyou know, not broken.
But two, when you tradeinformation with these, you know
these methods of communicationwell.
Invariably at somewhere,someplace, it becomes five
o'clock for somebody, right.
But the trucks don't stop.
The warehouses are still takingappointments, but the

(05:10):
information trade stops.
Right, and that's the brokenpart about it, right.
Which is a truck unused for aday or a night is probably 800
bucks.
A warehouse missing anappointment and having labor
there could be hundreds ofdollars, right.
A container sitting on a portfor seven extra days because

(05:35):
they couldn't get the paperworkon customs done, you know, could
be thousands of dollars ofcapital cost, and so the
underlying thing is not stopping.
The perishability cost of theseassets is so high.
But the trading of informationis what's broken.
It's broken because we're usingfairly asynchronous, cumbersome

(06:01):
methods to trade informationacross these companies and so it
overruns the operators.
It creates waste in thisindustry, because when you stop
trading information, wastehappens.
The container stays on the portfor a few more days, you know
the intermodal doesn't getpicked up or it's dropped, but
then it's never picked up thenext time.

(06:22):
The appointment is missed.
That's what I mean by you know.
It's such a crucial industryand the operators are the heroes
here who are making this happen.
In spite of all this, like, inspite of getting 800 emails a
day, they're executing it.
You know, literally I was atanother brokerage and you know,
know, this person shows up at 6am and she drives for an hour to

(06:45):
get there.
So her day is starting at 5 am.
At 6 am she starts her day andshe's not leaving till like 5,
36 pm and then an hour back.
So that's a 14, 15 hour daysfull of emails, phone calls, and
it leaves her like very littleroom to do creative work, the
work that was advertised whenshe got the job.

(07:08):
She's not doing any of that.
She's not calling on theshipper and saying, hey, let's
look at this month's shipments.
Can we read out them?
Can we pre-schedule them?
Can we do a multi-stop for you?
Can we reduce your waste here?
Can we do a multi-stop for you?
Can we reduce your waste here?
Can we make it more efficientfor you?

(07:32):
But it's just so busy thatthere's no time for the creative
work.
But the only time is for hey,is this information there?
Oh, can I log it into my siloedsystem that nobody else sees
than me?
But I still have to log it.
And then let me downloadsomething and send it to
somebody else, and they're doingthe same thing.
They're logging it in theirsystem that no one else sees,

(07:54):
right.
So that's what I mean by it.
The brokenness is one of thelargest industries in the world,
like worldwide $10 trillion, inthe US 3.2 trillion, like a
tenth of our economy.
That powers every single thingthat we see and touch and feel
is still today run on tradinginformation in very, very

(08:19):
archaic methods, in my view.
You know so.
Oh, it makes it perfect.

So what I was going to say is anyone who has
ever worked in a brokerageperfectly understands the
problem you're describing.
So we're talking about the dayin the life of an operator, and
different brokerages anddifferent models have different
titles for this function, but itcould be logistics coordinator,

(08:48):
could be operations wrap,operations coordinator, account
manager.
You know the roles, thefunctions that I'm specifically
thinking about are people whoseresponsibility it is to take
care of existing customers, takecare of existing customers, and
in doing so you have to buildorders.

(09:08):
Depending on if your customeris sending things over via EDI
or if it's email, you're oftenhaving to manually build the
orders.
You have to schedule the orderspickups, delivery.
You have to communicatewhatever new POs are added to
the orders, whether that'sputting them in the system or
notifying the carrier that'sbooked on the order.
You have to communicate anyupdates, delays, and your day is

(09:35):
, you know, coming into 800unread emails is not a surprise
to me at all.
That is a common reality forthe operator in this business,
and the day is a constant.
It's a never-ending game ofcatch-up and you never are
really caught up.
So the way you describe thisproblem is pretty astute,

(09:58):
because, as someone who'soperated a brokerage.
You know that was the life of alot of our employees and
exactly what you said in that,the job someone thinks they're
signing up for is never that.
That's not what ends up comingto fruition, and it's not
intentional necessarily,especially for growing companies

(10:20):
.
It's like you're constantlygrowing and throwing more people
at the problem.
At least historically, that'sbeen the way we've always
addressed it.
That seems to be changing intoday's environment with
companies like yours.
I have no choice, because Istarted with this question.
I have no choice but to talkabout Augment for a bit here.

(10:40):
I was going to start in yourbackground and then get there,
but let's talk about Augment.
So this is the you know, thebusiness you're building today.
Augment is intended to solvethis exact problem, correct?

Yeah, it's intended to.
You know, I think we feel liketo fix this industry.
You have to make the lives ofoperators better.
You have to get them back towhat their advertised role was
this creative problem solving.
You're helping the shipper,you're building relationships,
you're negotiating.
You know things that were inthe job board and we believe the

(11:15):
way to do that is to take allthis tedious mundane you know
repetition out of their day.
You know, and so, yeah, sothat's what we're building.
We're building these.
You know AI teammates.
Think of them as like apersonal intern that everyone

(11:36):
will have.
You know they can interact withsystems.
They can call, they can email,they can text, so they have sort
of the skills that you wouldexpect from an entry-level
employee to have.
But they can follow theseworkflows and SOPs.
You know this business, as youknow, is a lot about SOPs.
Like, hey, if I have myenterprise customer Pepsi, they

(11:59):
want me to run my loads this way.
That's a 30-page document andyou don't want to miss that.
And in status quo, you havelike a team that like learns
those 30-page document and thenmake sure that every load is
sort of matching that SOP.
You know these.
We're calling it Augie, butpeople like Kristen in any way

(12:19):
they want.
There's lots of different namesnow, but Augie can like read
these SOPs, can like followthese SOPs to a T, and so it
takes away that tediousness.
It's like am I, you know, am Imaking sure that when I'm asking
for the truck it's belowrefrigerated minus 10, because
that's what my customer, donFoods, wants?

(12:42):
Okay, I can ensure that like,like Augies can do that.
And then they're taking over,like some of the you know sort
of basic workflows designed as,like you know, tracking loads,
collecting all the documents.
Like nobody likes collectingdocuments no, no one likes it
and like documents that haslumpers, then detentions, then

(13:06):
accessorials, some of those youfind out a few days after the
load is being delivered.
And now you have to go back andnow the you know the dispatcher
on the other side is like whatare you talking about?
I've sort of moved on, you know, and no one's getting paid,
commission's not getting paid,you know.
And so like, okay, let's nothave humans do that.

(13:27):
Like we've come pretty farahead, let's have you know
systems collect these documents.
They should be able tounderstand that, oh, this load
requires excess oils, it hasdetention.
Right Now, let me go collectthose.
And when I get the documents,can I like, match it up with,
like, is it for the right loadnumber and is it for, you know,

(13:48):
does it sign or is it not signed?
You know, like these were thethings before, like we would get
that we would forward it to anARAP or a finance person.
Then they get to it a weeklater and they're like, hey,
this is the wrong document, Ican't invoice.
Then they pointed back to therep.
Now the rep sort of moved on.
Oh my gosh, I need to deal withthis load that was delivered 10
days ago, but if I don't dealwith it, nobody gets paid, and

(14:13):
so these things add up.
All these are the word I wassaying.
In this world of AI, let's justmake all of these things so
that we can take on all thistedious work, so that the
operators can now focus onthings that we are.
Take on all this tedious workso that the operators can now
focus on things that we areuniquely positioned to do, like
judgment, right, like, oh,here's an enterprise customer.

(14:34):
You know we got these loads,but this load is going to run
negative margin.
But you know we've got tomaintain this relationship.
We have an rfp with them.
Okay, I'm going to run anegative margin, like that's a
decision.
Yeah, I can't take that humancan, because they have a puts
and takes on the relationship,the rfp.

(14:56):
How much negative margin?
What's the pulse on the market?
You know, or like relationships,you know some deep negotiations
, or when things break down,like you know, a truck's on the
way, you did a track and tracecall and the guy said I broke
down, I'm on the side of theroad, I need help.
Well, a bunch of people needsto now get coordinated, like the

(15:16):
customer needs to know thewarehouse needs to know an
extenuating circumstance hashappened.
You need to now figure out whatto do with this.
Do I send a new truck?
Do I have enough time?
Well, that's the time you wantthe operator to spend on.
That's the service level thatwe're delivering to the customer
and the shipper.
So Augment's vision is that,hey, it's all fixing logistics,

(15:39):
all starting by really bringingthe operator back to doing
creative work, decision-making,this stuff, and then let's take
all this tedious work away fromthem.

So it's interesting because I see a
clear.
So the best way to say this, Iguess, is relying on humans to
recall all of the rules thatexist for all of their customers
while playing the game of catchup constantly.

(16:14):
So you know, I think the bigchallenge for people is yeah,
there's an order, so for one, Iwant to focus on revenue
generating opportunities, and soI'm trying to spend my day on
the phone with customers orcarriers or whoever can help us
get the next load and make thenext dollar, and oftentimes
these issues that come up arenot issues that could certainly

(16:38):
make you any more dollars.
They're not getting you anotherload, they're almost.
They're generally a distractionfrom the opportunity to do
those things you want to do,which is a big reason why reps
don't want to deal with them.
You know, when you get that lumpor receipt from a load four
days ago and you know thecarrier sends it to you finally,

(16:59):
and now you have to deal withit.
It's like I can either deal withit now or I can focus on what's
making me money, and but by notdealing with it now, one I'm
delaying our chance to get paid,I'm delaying the carrier's
chance to get paid and mostimportantly, potentially is
there may be a rule that says iflumpers aren't turned in within
48 hours, they don't getreimbursed.

(17:20):
There are plenty of customerswho have rules like that,
whether it's for a lumper, fordetention, a layover, whatever
it may be, and over time thatmoney adds up.
Remember every rule for everycustomer and act accordingly to

(17:45):
ensure every load is taken careof within those rules while
dealing with the kind of nonstopfirefight of their day-to-day
job.
And so there is a lot that'smissed and it makes sense if you
could be certain that the AIwill.
Essentially, if Augie canalways like, it will never miss

(18:05):
a rule If you've told it A plusB equals C, or like, if this, if
AX, then do Y, and it just willalways do those things.
It does set up the businessvery nicely to have way less, I
guess, leakage or just kind offall out from people making
mistakes.

Yeah, that's right.
And the other piece here,andrew, is that you know it's
sort of counterintuitive ormaybe not, but you know when you
ask somebody for something andit's on top of their inbox, they
respond fast, right, likesomething happened, and like,

(18:46):
hey, andrew, you have this infolike right there.
Forward it right Now if I askyou that 10 days later it's
going to take you longer to findthat right.
So you slow down in responding.
You're like okay, I got thisemail.
Oh, I know, this happened 10days.
I have to go search, search,find it.
There'll be four emails.

(19:07):
I have to look at those fourwhich ones it is.
Maybe I'll get to it.
End of the day.
Sometimes you forget.
So now it's 11th day, maybe youforget.
I'll struggle.
Now, five days later, I sendyou an email I'm saying, hey
guys, it's 15 days, I need thatinformation.
So the longer you wait, thelonger it takes you know and

(19:28):
it's a human syndrome and sowhere these AIs are good at like
, they're setting alarms forthemselves at a load level.
Like Augie is setting an alarmsaying, okay, this load got
delivered, you know, maybe Itext the driver if that's what
the workflow says.
Within 10 minutes the driverdoesn't respond.
I need to email the dispatcherin like an hour.

(19:51):
And it's not missing that it'sdoing it at a load level and so
that makes it actually do thesethings faster too.
On the other side, theyactually like it because it's
timely, and so we as humans,let's say, if a rep is running,

(20:12):
let's call it actively 100 loadsright now that are active like
somewhere in motion.
That's like in your mind,you're setting 100 alarms and
there are 10 things that happenon the load and there's an alarm
that needs to extend for those10.
That's like a thousand thingsthat they have to keep track of.
I mean, that's impossible, likeexpecting a human to have a

(20:33):
thousand things time-wise, at anhourly basis, to keep track of,
right?
Well, this is the thing theseais are amazing at, so let's use
them for that, and then let'suse them for that.
And then let's use us for whatwe are great at, which is
judgment Right?
Nobody like AI doesn't havejudgment.
We have judgment.
We can do puts and takes on anuncertain outcome and incomplete

(20:57):
information.
We can come out of a decisionthat aligns with my company's
goals or values or objectivesRight, but if we don't give
humans the time to do that, thenthey're spending all their time
chasing these alarms and theyhave very little time to
exercise their judgment.
You know, and so that's what ourhope is that we can make this

(21:17):
happen.
And if we start with that, thenI think we can uncover, like
the second aspect of it, whichis the waste, that, okay, if you
get your operators doing greatthings, creative work now across
these 20 companies thatcollaborated to get your coffee
mug where it is at, we can nowmaybe reduce the waste between
them.
Now we can do more smarterthings across these companies.

(21:39):
You know, and so that's sort ofour, I would say, objective
number two in logistics is that,one, make the lives of
operators great, but two, let'sreduce the waste here, and if
you reduce the waste, we makelogistics sort of better.
There will be more of it.
It's just the nature, you knowwhen you make something better,
there's more of that.

Are you looking to grow your brokerage?
Are you struggling to land newcustomers in these challenging
market conditions?
Look within so many companiesthat tender you freight
throughout the domestic UnitedStates also have business coming
out of Mexico.
A year ago I understand why youmight not have seen that
freight as an opportunity, buttoday Cargado exists and that

(22:20):
means any load coming into orout of Mexico is now an
opportunity for you to support,and that means any load coming
into or out of Mexico is now anopportunity for you to support.
In just over a year I've beenable to see Cargado go from
ideation to launch to rapidgrowth.
It's amazing to see how manylogistics companies have been
able to use Cargado to expandinto Mexico to grow their
business.
Cargado is the first platformthat connects logistics

(22:46):
companies and trucking companieswho are moving freight into and
out of Mexico.
If you move Mexico freight orare planning to reach out to
Cargado today at cargadocom,that's C-A-R-G-A-D-Ocom.
So a couple of simple questions.
I guess One is it possible forAugie or the AI to miss an alarm
?
Just to stay with that analogy,if you give it a rule or give

(23:11):
it an SOP, such as if a load isbooked with a carrier Augie must
call the driver one hour beforethe pickup appointment if there
is no update in the load.
If you give it that clear rule,can it ever miss it?

I mean, you never say never, but very rare, very
rare.
Like you're looking at you know, maybe less than than you know,
half a percent or less thaneven that, like 0.1.
We haven't seen that.
But systems can go wrong.
Uh, you have to, obviously.
You know um and and so but.

(23:56):
But even if it misses, it'sgoing to be better than humans
for sure.

I just am curious to understand it.
So that's why I'm asking thatquestion.
It's not meant to be criticalof it, it's more just to like
understand.
I'm pretty sure I understandthe answer is if it has a rule
and you know the situation iswithin that rule, it will
execute the rule Like it's astrustworthy as you are going to
get with that I think where it'snot great at Andrew today is

(24:25):
that you know, let's say you do.

It's an hour before you make the check-in
call and the person said hey, Idon't know like I'm ready, but
I'm not in the truck, and youknow I'm waiting for something,
but I think I have it.
I have to check in my email.
So this is sort of complex call.
It's not very clear.
Like check, check, check, check, check, equipment check.

(24:53):
It's like there's some graysthere, right.
And now we can set a rule inthe workflow to say like hey, if
there are some grades, like ifthe equipment is not matching or
if, I don't know, the reeferisn't cooled enough, whatever
the rule is, then eitherescalate to your rep or at least

(25:17):
inform the rep.
Or you could say like maybecall back and say like hey, we
need these things to be done.
We need to call back Now.
Interpretation of that, becausea call can be quite complex or
an email chain can be complex.
You have to interpret like hey,is it really matching every
single thing that the SOP saidor not?

(25:40):
There I would say it's liketoday, at least for us it's like
97, 96% us, it's like 97, 96percent.
So there are three to fourpercent of the cases where the
adherence on those is nothundred but something as binary
as I can't call within an hourthat it does very well and
there's lots of software workthat's going on now that like

(26:03):
improves that adherence from 96,97 to 99, a little bit higher,
yeah.

So when did this problem and the solution you're
developing, when did that firstcome to mind for you?
How long ago did you startthinking about this?

Well, I've been thinking about the problem.
At Deliver, we were a consumerof freight.
We also had built our littleown freight brokerage.
Because we were consuming somuch freight, it made sense for
us to build our own a little bit.
Deliver was a very tech-centriccompany.

(26:46):
We were mostly engineers and wehad consuming freight and this,
and every time we sort of grewfreight.
You know, it just occurred to melike we're just not creating,
you know, leverage here, likethere should be leverage, you
know, in a good business, whichis you're doing more with the

(27:07):
same number of people or you'redoing better service with same
number of people.
And that's where I started tolearn about sort of the freight
word.
And you know I chalked it up tosaying like hey, this is a
fragmented space, the softwaresare also quite.
There's so many different typeof tms's, for instance, and at

(27:29):
that point there wasn't a way.
But I think as ai came in, youknow, literally two and a half
years ago, um, about summer oflast year, I started to like
think about prototyping andsaying, huh, for the first time
we can meet people where they'reat If they want to do Telegram.
Ai can do that, it canunderstand the message.

(27:52):
It doesn't need an API.
It can turn an email almostinto structured sources really,
really well.
Or it can turn a phone callinto a structured, almost
API-like response quite well,way better.
So that was like the sort ofmoment for me to say like, ah,
maybe this is the time where wecan meet the industry, where

(28:13):
they're at right.
We're not changing, we're notasking the industry to say, like
stop using email or go intothis one centralized system and
everything will be magical.
We're saying you can do thebusiness that way.
We're just giving you an AIemployee, an AI teammate that
can do exactly the things thatyou're doing exactly on the same

(28:33):
system.
There is no system to learn andhow.
Maybe that is a breakthrough.
And so I think it was summer oflast year and then we started
building a bunch of prototypes.
We shadowed a lot of operators,like I think, like 60 operators
.
We shadowed, like literallyevery day We'll just sit behind

(28:54):
them and we see what they do,because there's no better
learning.
You know and and this I thinklearned very, learned, very
quickly that, yeah, there issomething here, like when we
shadow these off, like we canbuild something that can help
these operators become like youknow, give them more superpowers

(29:19):
, right so?

yeah, were there any kind of unique takeaways or
surprise?
I know you said you had alittle bit of a freight
brokerage at Deliver but I knowthat wasn't the focal point of
your time.
I'm just curious as youprepared to kind of dive
headfirst into freight brokerageand you spent all that time
shadowing operators, were yousurprised by anything or any
kind of interesting takeawaysfrom that time?

(29:42):
That kind of were lightbulbmoments for you as you got going
here A lot of redundantcommunication.

Like you know, in brokerages you do these
customer email groups andeverybody who is either a rep or
a demand rep or an ops rep ispart of that customer email
group.
And so there were 50 emails.
Those 50 are going toeverybody's inbox and
everybody's looking at those 50,even if they don't have to take

(30:11):
action or it's not evensupposed to inform them.
And just the amount ofinformation overload that comes
with that and I know why.
Sort of it landed where itlanded, which was like you know,
people leave their vacation butloads are running 24 seven.
How do you keep everybodyinformed?
I sort of it landed where itlanded, which was like you know,
people leave their vacation butloads are running 24 seven.

(30:34):
How do you keep everybodyinformed?
But man, the amount of overloadbased on that was just mind
boggling.
You know like we literallyanalyze so many inboxes of
people to say, like out of that,like which emails could I
delete?
And nothing in your day to daywill change.
And it was a big percentage.

It's like, depending on the person, it
could be over 90%.
I mean it's a crazy number.
It's such a good.
I'm glad I asked that questionbecause your answer is so
perfect.
I mean, the amount of emailcommunication that is wasteful
in brokerage is it's disastrous?
I mean it is.
I think about my own inbox andI may be a part of the problem

(31:10):
as an ADHD guy who just wasgenerally disorganized but,
having spent a lot of timecustomer facing and selling, I
brought in customers.
I would be on the initial emailchain and then three years
later I'm getting emails aboutlumpers on an account as the CEO
of the business and it justgets put into a folder that
never really gets opened.

(31:31):
But each of those folders has4,000 unread emails in it.
It can easily get out of handit can get out of hand.

And I think one of the problems there, in my
view, is that you know it's sortof, you know it's hard to
distinguish signal from thenoise and because among those
emails there are 10 that youreally need to action, but you

(32:03):
have to go through the 50 tofind the 10.
Now you've made very hard forthe person to find those 10.
And it's very easy to missthose 10 because you're so busy.
So I think this whole signal tonoise ratio is not great and
we're working on that problemand we hope we will have a good

(32:23):
solution there.

Yeah, so I'm curious as you got started here
in the last year, what kind ofthings did you pull from your
experience?
I mean, you had this incrediblesuccess with Deliver and a
couple other businesses you hadstarted.
I'm just curious if there areany really important things that

(32:46):
you as a CEO, as a leader,thought to yourself like I need
to make this a cornerstone, as aleader, thought to yourself
like I need to make this acornerstone, like I learned from
my experience and this needs tobe a crucial part of how we
build augment.

Yeah, man, there's so many.
But let's see, I think cultureis, you know, it's one of those
things like trust, like it takesa long time to build but you
can lose it instantly, andculture is similar.
Like you need to build veryresilient cultures means you

(33:20):
have to work every day at it.
And it's not just about likehey, you write five pretty
values on your powerpoint decksor you know, on your, as a mural
on your on your office, and andlike have employees on the
first day like read it or orread a document about it.
But it's like how you makedecisions every day, right, how

(33:43):
you show up, how do you maketough decisions?
Are those all aligned with thevalues?
And being extremely intentional, that when they're aligned, to
point it, why we made a decisionthat was aligned with that
value.
You know, like if you sayputting customers first, which
is one of our values, that's animportant value, but you're

(34:04):
short-circuiting some customersor telling them something that's
not true, well, employees aregoing to see that very quickly.
It's like, no, that's not yourvalue.
Like you're saying it it's yourvalue, but you're not.
But you have to like, live it.
You know, you have to live itevery single day.
And I think if you do that,then your next 50 employees,

(34:24):
your next 100 employees, theybecome the, I guess, the bearers
of your culture for the next500.
And so it's extremely importantthat you know you get that
foundation right, you know, andso, like I think, when I started
Augment, like the first two,three days, we just wrote values

(34:48):
.
There was no code written, youknow, we were just shadowing
people and we just wrote a setof values.
We socialized among ourselves,we debated every single word of
it and we said, okay, what arethe mechanisms we're going to
use to uphold these?
How are we going to incorporatethat in our hiring processes?
How are we going to incorporatethat in letting people go,

(35:08):
people who are not a fit, andlet's just talk about that.
Let's just put all of that in apaper so that when we have a
disagreement, we can now go backto this and saying, hey, this
is what we agreed on.
Right, we agreed that we'regoing to be honest and direct
and that's one of our values.

(35:29):
And like that means disagreeand commit or that means we're
separating ideas from people,and like that, like, if there's,
let's just go back to that and,like you know, become sort of
like our constitution, if youwill, that when things can't
agree, like hey, this is whatour in in america, our founding
fathers sort of agreed on, how,when we were building the, the

(35:52):
country, I think almost treatingit at the same level of
reverence, I would say, is onething that I'm sort of carrying
forward Because it pays individends.
You know, once you do that, yourhundred, your next few hundred.
Now a decision could be made inan office, whether you are
there or not, you're maybeseveral layers removed.

(36:12):
It's going to be made inaccordance to the values and if
you can do that then you'rereally building a scalable
business.
But if you have to be in thatroom to make sure the decision
is made well or not made well,then it becomes very hard to
build a scalable business.
You know, becomes very hard tobuild a scalable business.
You know.
I think that sort of I would sayyou know, listen, like focus

(36:36):
and customer delight, like allsoftwares when you start is you
know it has edges, it's notperfect, it takes a while to
build great, great software.
You know, and I think one thingwe probably didn't do as well
early on in deliverable, but wefixed it I think we're doing a

(36:57):
better job here is reallysupplementing that with customer
success teams.
So, very early on, you knowwhen people are implementing
Augie.
You know we have ex brokers inour team who sort of felt the
pain, who had those 800 emails,who are then going and
implementing Augie.
We have ex-brokers in our teamwho sort of felt the pain, who
had those 800 emails, who arethen going and implementing
Augie and they're like OK, well,and even if the software is not

(37:19):
doing like 10 out of 10 things,the 2 out of 10 things it's not
doing, or 1 out of 10 thingsit's not doing, they're able to
either tease out the solutionfor them or bring it back in the
real terms, back to the productand engineering team, like, hey
, we guys really need to buildit.
We face it.
Our customers are facing it.
We're just having that eyes andears of the customer deeply

(37:41):
penetrated in the company.
Is is probably you know, but butthen?
But then there's somethingdifferent, andrew.
Like AI is moving faster thananything I've seen so far.
You know it's making mucheasier to write software.
You have to write software in adifferent way though, you don't
, you know.
So, like all that stuff I hadto let go.

(38:03):
You know, I had to rethink.
Like you know, like in atraditional software, you think
about unit test cases and yousay like, oh, I'm gonna
enumerate, I'm gonna write theseunit test cases, and then I can
just deploy it.
You think about unit test casesand you say like, oh, I'm going
to enumerate and I'm going towrite these unit test cases, and
then I can just deploy it andyou put those test cases in your
CI CD like deployment pipelines.
But in the AI world theenumeration is many.

(38:23):
Like when the AI talks tosomebody or sends an email, the
number of cases are infinite onhow you will get back the
responses, how does it need torespond back.
And then now you have guardrails.
It's like, okay, I want you toachieve this objective, but
here's the guardrails for itDon't ever do this, don't ever
do this.
So the software you're writingis very different because the

(38:45):
number of cases it is addressingis certainly a whole lot more.
Addressing is certainly a wholelot more, you know.
So also had to change and alsohad to move fast.
Because you know, like we it'sit's moving so fast that we've
gone from zero to like 75 peoplein a short amount of time and
because it's just so much tobuild so quickly, you know so.

Attention, freight brokers and 3PLs.
Greenscreensai is transforminghow freight professionals price
and quote freight.
With ai powered pricing andreal-time market intelligence,
greenscreens delivers accurate,reliable predictions to help
stay competitive in a constantlyshifting market.
Whether managing spot rates orlong-term contracts,
greenscreens empowers brokers toquote with confidence and boost

(39:31):
profitability, removing theguesswork from freight pricing
strategies.
Trusted by over 220 brokerages,greenscreens is leading the way
in the future of freightpricing.
And now there's Illuminate,greenscreens' latest product,
designed to shine a light ondeeper freight market insights.
Illuminate providesunparalleled visibility into
spot and contract freight trends, giving users a clearer view of

(39:59):
pricing fluctuations and marketconditions to inform smarter,
more profitable decisions.
Visit greenscreensai, slashthefreightpod and discover how
Greenscreens and Illuminate canhelp win more business more
profitably.
I'm curious about the elementof change with respect to how
technology is built and how AIis built, and or how the
technology using AI is built,and I guess, is that concerning

(40:23):
at all to you?
Or like how do you think aboutthe way to?
How do you think aboutpositioning Augment to be like
the leader sustainable product,knowing that every three, six
months, there's something newand maybe better than what I'm

(40:57):
currently scaled and buildingwith?

Yeah, that's a really good point.
So you know, like, if you thinkabout, in that scenario, you
know you have you can.
Either you have two choices youcan lean in or you can wait

(41:20):
because the problem is going toget solved, right.
So you can say like, oh, I'mgoing to wait for that next
breakthrough, or I will lean inhere, but I will adapt.
You know, obviously we believein the latter, which is like, if
you lean in here, but I willadapt, Obviously we believe in
the latter, which is like, ifyou lean in here, you learn
enough about the space, youlearn enough about the quirks,

(41:41):
you learn enough about the usersthat your adaptation would be
better than somebody who'scoming in six months later with
that new tech, for instance, ora new version of OpenAI or
ChatGPT or Cloud or underlyingLLMs, because, fundamentally,
all of us are using one of those, and that's what's changing so

(42:03):
fast.
I think the second piece is.
What is not changing, though,andrew, is that, like, like AI
is still a tool, right.
Like ultimately, you have toeither deliver something better,
cheaper or faster, like it doesnot matter, like whether it's
AI or not, and so you have todeliver productivity or revenue

(42:27):
gain to the business, right, andthat's not changing, right.
And so staying focused on that,deeply focused on hey, what is
the business value?
With a viewpoint that thisunderlying LLM is going to
improve radically.
So you're also starting to takebets and saying I think the
LLMs will be here six months ayear from now.

(42:49):
I have to start thinking aboutthose use cases today will be
here six months a year from now.
I have to start thinking aboutthose use cases today, all in
line with, like this businessuse case of like I'm building
something better, cheaper orfaster.
It has to be that you know.
And so I think, like yourquestion was like how do you do
that?
I think you lean in, we do test.

(43:09):
Almost We've built a frameworkwhere, when new LLMs come, we
rapidly test against those andsee, in terms of those three
dimensions, better, faster,cheaper.
Are we improving?
If it's not improving on that,we don't actually care.
We don't care about thebenchmark.
It hit on a math test Like, oh,it improved the benchmark on

(43:31):
this new math test from 93 to 97.
We'll try it and we'll run, youknow, thousands of simulations
against this new llm and saying,did it make things better,
faster or cheaper?
If the answer is no, we'llstick with what we have.
If the answer is yes, we'llswitch.
But then you have to build.
You know infrastructure andtechnology that can easily
switch those things.
We can run real life scenariosvery fast so we can know like

(43:55):
nothing Augie's in the worldtoday that is handling tens of
thousands of loads.
They're not going to getdisrupted when I change this
underlying engine that powerstheir intelligence or their
actions.
So we had to build a lot ofsoftware that can do that and
that's a new requirement in theLord.
That, like we didn't, I didn'thave to worry about it in
Deliver or in my previousbusinesses.

(44:16):
But now you have to.
You have to, you know, designyour things from day one,
assuming that one.
You're going to try multipleLLMs and you have to try them
differently every three monthsor six months.
You know.

Yeah, that makes sense and it's interesting you
mentioned because everyone'susing similar or the same
underlying LLMs or technology tobuild on and I think to kind of
broaden that.
This space has gotten verycompetitive very quickly and
there are every week there'sanother announcement of a

(44:54):
company, either started by someyoung kid out of silicon valley
who is just eager to make hisname, or it's someone like
yourself, an established,successful previous founder, who
sees an opportunity and wantsto participate.
I don't know what the number is,whether there are 30 or 40 or

(45:18):
50 different AI companies comingin to automate work in freight
and make brokers more efficient.
It seems like everyone hassomewhat of a different approach
.
Some companies are just focusedon appointments to start, some
are just focused on appointmentsto start, some are just focused
on APAR, some are just focusedon phone calls to carriers for
tracking, and then somecompanies are kind of

(45:40):
bootstrapping early with smallteams, small raises.
Others are going big andraising a bunch of money,
getting a large team set up andtrying to attack several things
at once.
How do you think about what hasbeen your approach specifically
?
Obviously, you guys just raiseda bunch of money, so definitely

(46:02):
not bootstrapping over here,but I'm curious how you have
thought about entering into thespace if you're approaching one
problem at a time or all of themtogether, and kind of why
you're choosing the path you'rechoosing.

Yeah.
So we're taking sort of a morehorizontal, broader approach,
right?
So we're saying, hey, there aremaybe 30 to 40 different
workflows, traditionallymultiplied that by customer SOPs
or configurations.
So now I'm dealing withhundreds of different workflows

(46:37):
in a business and ultimatelyAugie needs to handle a good
chunk of them.
And our reason is that everysubsequent email, phone call
text or a system login getssmarter when Augie has context
over the load, right, like inyou ran brokerages, why did you

(47:03):
like?
There's a reason probably whyyou said like, hey, you own, you
know, this customer for me,right end to end.
And I'm assuming the reason wasyou have the full context, so
when the next situation happensyou can handle it.

Yeah, certainly that, and accountability.

And accountability.
Yeah, exactly, so our approachis that it's no different, for
we're approaching almost Augieto be a teammate and and be
accountable, but we think it canonly be when it sees hey, it
was built, the load was built.
This was the customer SOP.

(47:42):
It required the followingthings for the equipment.
It required the followingthings for Lumpur we understand
that while building the loadOkay.
Rescheduling the load based onthe requirements oh, okay, we
can now help them track andtrace load or assign it to a
carrier.
Or when I'm collectingdocuments, all of those things
are helpful, like if I'm justchasing documents without

(48:06):
knowing like, hey, these guysrequire this customer requires
everything to be submittedwithin 24 hours.
Well, I'm going to run at adifferent pace for that customer
because otherwise I'm notgetting it.
But it helps to have theentirety of the context of the
business, the object, whetherit's the load or the customer in
your case and so our approachis that that makes for a better

(48:30):
experience of Augie completingthe task.
So that's been our thing, andand and, like the other piece,
which I'm quite different thanthe rest of the world and the
market here is that you know, alot of people think about like
hey, email automation or phonecall automation is like this

(48:54):
superpower, right?
And I think they're missing thepoint, which is it's not about
phone automation, it's aboutgetting work done.
Right, like the best phone callis the call never made.
It's a defect that we have tocall somebody.
It's not a feature, it's adefect that we have to call
somebody.
It's not a feature, right, like, if you could get the work done

(49:16):
without calling anybody, wouldyou prefer that than to actually
have more calls, like everybodywould prefer.
Like, oh, we just want to getthe work done without call.
You can't do that unless youhave an understanding of the
load, unless you have theunderstanding of the customer,
right, you can't do that.
So we think if you start to likedo point solutions, saying, oh,

(49:37):
just automate my calls, you maybe solving the wrong problem.
You're reducing defects butyou're not reducing.
You're like you're handlingdefects but you're not reducing.
You're like you're handlingdefects but you're not reducing
them.
You know like the industrywould be better off if this were

(49:58):
to happen without so manyemails and so many calls, right?
And so that's maybe quite a bitof a different approach that we
take.
We say like, hey, our job is toget this, you know, load tracked
to the specifications of yourcustomer as specified in their
30 page SOP.
Ok, what's the best way to dothat?

(50:20):
You know, maybe it's not topound the driver every four
hours, like where you're at.
Maybe it's to better at, likeyou know, seeing where the ELDs
are and why are we not gettingthat?
You know, um and so.
So that's one area where Ithink I I differ quite a bit
from some of the noise in thisindustry, which is like people

(50:41):
are just so excited about callsor emails, or, you know, I'm
like, no, that's not the problemso it's interesting because, as
I've kind of learned about thispart of the space and I'm
definitely I appreciate thatexplanation.

I think that really is helpful for
understanding how you thinkabout the problem and,
specifically, how you thinkabout the solution.
I've heard challenges withlet's focus on appointments for
a second um, one of the bigappointment scheduling companies
that has gotten a lot ofattention.
Um, I've heard challenges fromfrom their customers who've said

(51:20):
you know, it's designed toautomate the appointment process
but when something goes wrongand it breaks for whatever
reason, there's an exceptionwhich happens with.
Whether it's appointments oraparAR or tracking, there are
always exceptions.
That's the name of the game inthis business.
But what I've heard is this isa company that has a relatively

(51:40):
small team and what it does isessentially just breaks.
It kicks it back to theemployee at the company and they
have to then jump in and dealwith it, which and companies
like this are touting we'vereduced 70% of the need for your
employees to be involved inappointments.

(52:01):
Okay, that's good, you'resolving part of the defect
problem there.
However, it's not necessarily afully baked solution or an
end-to-end solution if everytime there's an exception, it
kicks it back to my team to haveto deal with it.
Now, on the other end of thespectrum, some solutions I've

(52:24):
heard are kind of likeAI-enabled BPO, where you're
going to outsource the entirejob to us and we will leverage
AI to do as much of it as we can, and that which the AI can't do
, we have a person who's goingto do the job itself.
Yeah, it sounds like yourbusiness is kind of in the

(52:45):
middle there, in that you wantAugie.
You know, if my operations teamhas five people with Augie
essentially now has six andAugie can do a lot of work and
you know, when something breaks,the team sees it because
they're working with Augie andthey jump in.
Am I describing right how youare, or do you feel like you're
more on that AI-enabled BPO?

(53:08):
End of the spectrum.

No, we are not sort of a BPO.
I think our approach is thatwhen there is an exception,
augie has the tools to do a setof emails and phone calls and
others that are needed toresolve it.
It's going to try that.
So let's say the appointmentwasn't booked because the

(53:31):
windows were too short and thereis nothing available on the on
the pickup facility to meetthose windows.
Now we've got a new set ofwindows, but this is sort of a
hard requirement.
Now it requires a quick messageto either the customer or the
customer rep and saying, hey,I'm not getting an appointment.

(53:52):
This is the next best we can do.
What do you want to do?
Now?
That's a message, right?
Augie can take that and say,okay, let me run that message
internally or externally, getthe response, then maybe pass it
to either the scheduling systemor its own scheduling system
and saying, oh, let me see if Ican get that so it can at least

(54:13):
handle a higher percentage ofthese exceptions.
And only after it has failed onsome of those it will then say,
okay, I need to know and I needto know Escalate.
Think of it as a junior levelemployee who've tried my best
based on what I was given as anSOP.
Now I need to report to mymanager that, hey, I've tried

(54:35):
these things and I'm at astandstill.
You need to get involved.
And so we hope to reduce thoseexception handlings to a much
smaller percentage, because ithas these communication tools,
both internal and external,right.
It sits on Slack or Teams inyour employee stack.
It's not an app, it's on youremployee stack.
It sits there and it has accessto your customer groups and it

(55:01):
will message them and they'llrespond back and then it can
take that message and say, ok, Iknow what to do next.
You know, almost very similarto what a remote employee would
do, you know.

And what's been the initial feedback in terms of
well, I'm curious both aboutsoliciting new business and
existing.
Let's start with soliciting newbusiness.
What I'm curious about is do wehave AI fatigue yet?
Like, are there as you reachout to brokers saying, hey, you
know I'm Harish.
Or your team says, hey, we wantto show you Augment?

(55:35):
Are there people who are likeI've had seven demos for AI
companies in the last few weeks.
I just can't do any more.
I've already got three demosongoing.
Are we there yet?

I think we're very close.
I think we're very close.
There is just a lot.
There's a lot, yeah, and likewe were having a prospect call
this week and we finished it andthat's what the person said
like guys, just fii, this is myfifth demo of this week, you

(56:08):
know.
So, uh, it's very easy to buildsolutions in the ai, like not
only like how ai is helping youknow logistics, like it's
helping engineers to buildthings very fast, right, so, um,
so there's lots of companies inthis space and um, and you know
, I can, like I would sympathizewith the decision makers on the
other side, like like, how doyou separate?
You know, what do you reallyneed?
Um, so, yeah, I think they'rethere.

(56:29):
They're almost there in my viewand for you.

How are you navigating that kind of fatigue
to stand out like?
How do you take that situationwhere someone tells you this is
the fifth one I've had this weekand turn that into at least a
trial?

Yeah, well, I think a couple of things.
I think one is, you know,fortunately so far I think we
are one of the few, or maybe theonly one, like who's sort of
looking at a very holisticorder-to-cash system versus a
point solution, and so I thinkthat message and value prop is

(57:11):
connecting that if they do thatover time they can think about
an order-to-cash transformationhere for the business.
They can think about, you know,in order to cash transformation
here for the business.
But two is we're sort of saying, hey, like to see how, like
Augie can internally communicate, externally communicate, let's
get started with a couple ofsimple workflows without you

(57:31):
having to spend a lot of timeinto integrations, you know.
So it could be like, hey, giveus some flat files and we'll
spin it up and you get to see itin action, hopefully in days,
because, as you know, the TMSsare not fully well designed for,
they're not the most friendlyin terms of integration.
The APIs are not the cleanest,I would say For sure, and some

(57:56):
are good, some are not For sure,you know, and some are good,
some are not.
So and so and and and, likefreight's been in a tough, in a
tough space in the last threeyears, like it's been really
tough market for freight and youknow every, every business is,
you know, feeling it in some wayor the other Right and so for

(58:16):
them, any initiative, aspromising as it is, it's a new
investment.
They have to put people in itto make it successful, they have
to put engineering behind it,and so that's like the biggest
hurdle, where they see thepromise and like, okay, I see
the promise, but how do Ijustify putting a few
engineering resources or athird-party engineering firm or

(58:37):
even our own business time tosee this through, like if it's
all just you know, shiny techwhich has been pitched to
freight along a lot of times,you know it's not like, you know
they'd like and like, why isthis not one of those you know?
And so we get to say like, okay, maybe we can do like a simple

(58:57):
flat file.
You get to see it, you get tofeel it, you know, and I think
those between those two things.
But you know, so far, like youknow, we're fortunate, we have
some very large businesses thathave trusted us early on and
opened their doors and almostco-building, and so taking some

(59:21):
of those case studies has helpeda lot where you're showing real
, meaningful productivityimprovements.
Operator happiness and operatorhappiness as you know, like just
the recruiting and trainingcosts in this industry is very
high.
If your operators are burningout Like, that itself is a very
massive boost.
You know if your operators arehappy they're producing more

(59:44):
revenue, but they're alsoquitting less.
You have to train less people.
Your classes are shorter,smaller, and so you know taking
those metrics have been helpful.

But yeah, at Molo we built a great company and
I'm proud of the work we did.
We knew when to ask for helpand sometimes that meant going
outside of our own company.
I'm proud we built an ecosystemof trusted partners like
Metaphora.
When we needed differentiatedindustry expertise in business
consulting or technologyservices, we looked at Peter
Ryan and the team at Metaphora.

(01:00:15):
They've consistently deliveredvalue in the transportation and
logistics space for over adecade for mid-market and
enterprise brokers, for shippers, carriers, private equity and
freight tech companies.
At Molo we use Metafora tosolve problems we simply
couldn't on our own.
Metafora is the only partneryou should trust to help you win
, whether that's doing ops andtech diligence, growing revenue,

(01:00:37):
optimizing spend or selectingand building software.
Go check them out atmetaforanet that's
M-E-T-A-F-O-R-Anet.
Yeah, so turnover and tribalknowledge are two significant
issues in brokerage and they'reway bigger when you pair them
together, because havingaccounts that are run on tribal

(01:01:01):
knowledge by people who've doneit for a long time, it goes well
until you have the turnover andsometimes it's employees
quitting and so that accountloses its rep.
Sometimes it's something biggerthan that and it's a company org
change and all of a suddenyou've got 50, 100 accounts that

(01:01:22):
are changing hands and all thetribal knowledge can be lost, or
the game of telephone by whicheveryone plays to disseminate
information, doesn't go well.
I think one of the biggestconcerns I've heard shippers
have one of the things that theystart to look at over time
within a business or a providerof theirs is how often the team
is changing, because it's almosta guarantee when you make a

(01:01:45):
change on the account and atleast one person changes,
there's almost guaranteed to bea problem, and it could be a
small one, but at least one loadwill be messed up because one
bit of information was noteffectively passed from person A
to person B.
So I do see how the right AItechnology it's Augie, or

(01:02:07):
hopefully it's Augie can helpboth of those problems.

Harish Abbott (01:02:13):
Yeah, yeah, yeah, I think that you've sort of
summarized it better than I have.
Like you know, we're reallylike when we like all given it
sees travel knowledge, it triesto retain it.
Now, you know, when it seeslike a dispatcher phone number
that's not in the tms, it willgrab it, it will persist it, um,
or the dispatcher says I'm nolonger the guy, like you've got

(01:02:35):
to contact this person now.
Like it just won't be, like oh,I'm going to call this person,
move on, it would actually oh,I've got a new number, new
contact, I'm going to save it.
You know, just, the directoryinformation is so dynamic and
it's moving, but if you look atthe tms itself, it doesn't
represent the dynamism of what'shappening in the real world.

(01:02:56):
Like people move their, theirnumbers change, their emails
change, their roles change, butTMS is relatively static because
it's like burdensome to go andfind something in an email and
then replace it.
You know, these are the areaswhere Augie is good at like it's
gonna, it's gonna get that infoand it's going to update your
TMS.
So you always have like themost recent.
You know update.
But yeah, turnover, and I mean Iwent deep into understanding

(01:03:21):
like the P&Ls of different sizebrokerages and just the cost of
training and turnover reallysurprised me.
If you ask me for the surprise,that was the second part, I did
not imagine to be.
Just the cohorts when you know,really surprised me.
Like, if you ask me for thesurprise, like that was the
second part, I did not imagineto be, you know, just the
cohorts when you hire,especially on the sales side,

(01:03:44):
and then those cohorts, how manyyou know, stay with you within
60 days, 90 days, 180 days.
Just the amount of money youspend for 180 days to get them
trained before they actuallystart producing for you and then
you know, it's like it's just,it was a much higher number than
I had envisioned before.

Andrew Silver (01:04:03):
So you make a great point and what that has me
wondering is how do you takethat insight and apply it to a
revenue model?
Because what essentially you'regetting at is the solution
you're creating.
It has this kind of immediateimpact in solving the physical

(01:04:25):
problem that you assign it,which is to okay.
You've now made the operator'sjob less monotonous.
You've made the operator's jobless time consuming, demanding,
so it doesn't have to be 15 hourdays, which means you've

(01:04:48):
reduced turnover.
You've reduced the amount oftraining that has to be done.
So you see, like once you zoomout a little bit, you're saving
a company a lot more money thanjust the dollars it takes to
schedule an appointment.
And I'm just curious how tothink about that from the
revenue side when you have thesekind of second, third tier

(01:05:13):
efficiencies you're creating forthe business.

Yeah, short answer is today we are not.
Today we are really thinking oflike amount of work Augie does
and ensuring that the businessesget, like you know, 7x ROI on
that.
Basically, you know, I treatthat as more like you know, if
you're producing the secondthird order effects, which I
believe would be prettymeaningful, they might take 12

(01:05:40):
months to fully pan out.
My thinking is that that shouldhopefully make Augie very
sticky and the most tenuredemployee in the company.
That's enough payback.
If Augie remains the mosttenured employee in the company,
that's enough payback.
If Augie remains the mosttenured employee in the company,

(01:06:00):
that's a good payback.

Yeah, it maybe is just most helpful on the front
end in how you solicit andcommunicate the opportunity,
because two years from now, forexample, I think you could go
through.
I know Arrive is a publicdesign partner of yours.
Yeah, I bet, if all goes toplan and you build this thing
the way you want to, in twoyears I bet you could look at

(01:06:23):
Arrive's business and, outsideof the very specific things that
Augie worked on, you could zoomout and see some of the effects
to turnover and cost oftraining going down.
And then maybe you're justselling that on the front end to
other providers, like hey forother providers, like hey for
other companies.
We've reduced turnover by.
We've seen turnover reduced by20% in the first 24 months that

(01:06:47):
Augie was implemented and ithelps you because I'm just I
don't like you, I don't knowthat.

I see a way to actually charge for that, but
yeah, I hate charging for thingsthat are just so first, far out
in future, but to you yourselfdon't know, you know, you don't
know the total impact of it.
Right, like you know, greatbusinesses get built when they
leave a lot of money on thetable.

It's a good point .
It's an interesting comment.

Yeah, if you look at some of the biggest
businesses, I don't know likeNetflix, it's a great business.
The value that the Netflixoffers to a person who is a big
consumer of movies is certainlyway more than I don't know $12

(01:07:39):
or $20 a month that they'recharging you.
You know way more right thevalue, but they're just they're
not going to.
I think.
Or, like you know, google has agreat business where you know
you're getting so muchinformation from them, or chat
GPT today, like at $20 a monthand it's making your life like
20 more productive and likethat's a lot of value it creates

(01:08:02):
.
You know they don't want tocapture most of it, um, but it
it sort of, I think, pans out inretention like yeah the
stickiness is what matters,right?

yeah?
Yeah so you did raise, I think,25 million bucks or more, 25.
Yeah, yeah, 25.
I'm just curious again.
This is more me thinking aboutthe general approach and how
everyone's kind of had thisdifferent entry point to this
industry.
I get why, with a more holisticapproach, a horizontal that you

(01:08:35):
know, you probably do need togo bigger.
But how are you thinking aboutwhy did you pick 25 million, why
did you pick 8VC, and how didyou get to that number and how's
it gonna be deployed?
Yeah, Okay.

So I think 8VC I work with 8VC folks for Deliver
and so it was sort of an easier.
We both know each other'sstyles on how we build
businesses.
They've been a big supporter atthe Deliver journey from very
early on.
Their philosophy is veryfounder-centric, which appeals

(01:09:19):
to me.
So that was sort of an easierdecision in at least early
phases of building Augment andthey were interested in this
space, the amount I think youknow in early days.
It's like we know there's alike there's a potential to

(01:09:43):
build a large business.
You know we're starting off withfreight, which is a big
industry, but the long-term goalfor Augment is beyond freight
too.
We're, you know, we're buildingfor the logistics world, like
warehousing and so on and soforth, like warehousing and so
on and so forth, and so we knewthat we needed a broader
footprint for some of the verycomplex communication problems.

(01:10:04):
So I'll give you an instancewhere, let's say, when you run a
business, or even in your mediabusiness, you probably get 20
pings a day.
Now, a week later, latersomebody else will ping you on
same thing, and now you're ableto connect those contexts in

(01:10:26):
your mind.
Right, okay, this happened,this happened, this happened,
and then you're able to respondwith those contexts that are
being connected.
You know, um, it's uniquelyhuman.
We never stop to think about it.
But that's so cool that we can,like, have hundreds of contexts
going and a month later,sometimes a year later, a friend

(01:10:47):
of yours might ask you aboutsomething and you're like, oh
exactly, yeah, this is what Idid, you know, and just respond
and just instantly connect.
And Augie needs to do the same.
There is no difference, right?
If Augie needs to be a usefulteammate to somebody in a
brokerage that has 2,000 people,there is at least 2,000
contacts going maybe more, right?

(01:11:10):
And it has to keep track ofevery single context over time,
so that two months from now, ifsomebody asks Augie about
something which started twomonths ago, it could respond
just like a human would.
So we knew like these problemsare very complex problems.
They take a lot of engineeringwork to like actually architect,

(01:11:32):
think and write and then solvefor it, right, and so for that
we required, you know, areasonable amount of money to
get started to build sort of abroader team, you know.
So now we have, like, and theuse of the money is engineering,
it is like we have 75 people.
Of that, I think close to 60are engineers, and so it is

(01:11:55):
purely, you know, building verytalented mission-driven
engineering teams.
But yeah, I mean exactly, itwas 30, 45.
I mean, I think it's sort ofthis dance between you want to
raise enough but like, give aslittle dilution as possible, and
you know, you trade it withyour VC or your partner to say

(01:12:21):
like okay, we're building for abigger mission, we're solving
some very complex engineeringproblems.
It's going to take somesignificant investment to solve
these problems.
Well, okay, but we also don'twant to be diluting ourselves
before too much right now,because we know we can start to
show revenue traction, customertraction fairly shortly.

(01:12:42):
And so it's a judgment callAndrew, at some point You've got
to make one.
But I knew that a $7, $10million round at this stage for
what we want to do just wouldn'teven get us far enough to show
the proof points.

And so what's your timeline in your head that
you think?
What's the next milestone foryou in terms of how you think
about the business growth andthe deployment of these funds
and how long they should lastyou versus another raise?
How are you thinking about thenext few years?

I always think about.
I like run for my business,like to at least have 18 to 24
months of runway.
You know that's sort of amental model in my that I carry
forward, um, and so at everypoint I'm sort of evaluating
that and saying, okay, do I havethat?
If not, maybe we need to raiseor we need to cut the burn or we

(01:13:41):
do increase revenue, whateverwe need to do.
So we have enough time becausemarkets can change.
But the big milestones for usis it's basically very simple,
early, early stages.
It's about delivering customerdelight, which in our case is

(01:14:03):
actually very straightforward,which is tremendous ROI and our
ROI, fortunately in thisbusiness, is highly measurable
your operator happiness.
Do you need too many offshoreor not that many offshore people
?
Are you running more loads withthe same number of people?
They're very measurable things.

(01:14:25):
That's one good thing aboutfreight which I also learned
after spending time is that youknow how measurable everything
is.
Not many industries are like sounitized at the fundamental,
like business object level.
In this case, everything isunitized at the fundamental,
like business object level.
In this case, everything isunitized at the load level.
And like, you know, loads perrep, or spread per load, or you

(01:14:48):
know anything, you, everybrokerage, we may not have the
highest fidelity, but they dohave a measure of that, you know
.
And so so those are, I wouldsay, the metrics, and.
But but really I'm like and I'mnot even saying it to be like
this you know, we, we, we, wejust hired our head of ops and

(01:15:08):
either he we I just this morningat the chat, like the only goal
is customer delight.
There is no other goal.
You know, if you do that,everything follows.

There is no other goal.
If you do that, everythingfollows.
Yeah, I prescribe to thatmedicine, so I can pick up what
you're putting down there.
How about, like five years fromnow?
What does Augment look like?

What does a successful five-year trajectory
look like for you?
Well, hopefully we're servingmany more industries in the
broader logistics supply chainarea.
You know we're impacting a lotmore business, but really I
think like, if you look at thesize of the logistics industry
that's 3 point, some trillion inthe US, 10 trillion broadly I
think there's at least 10% waste.

(01:15:59):
Would you agree?
Yes, 10%, at least At least.

There's so much.

There's trucks running empty, there's trucks
laying over, there's a lot,there's scheduling labor, but
the truck doesn't show up.
There's just so much waste,right?
So if you look at a $10trillion global world with at
least 10%, that's like atrillion dollar of waste.
Could we cut that In five years?

(01:16:28):
Hopefully we've made.
Cut 10% of that.
That would be a cool goal.
It's a hundred billion.
That's what we're going forhere.

No one will say you're not ambitious.
If you can cut $100 billion ofwaste in five years, you're
going to be hired as the nextDoge leader to go to the
government and cut waste there.
Yeah.

Well, I think, I mean, I think there's some
that's like it's exciting.
You know, like that's what gotme started with Augment, which
was like hey, like there is, inour own little way, a shot at at
at making this, this word thatpowers everything we do better.
Reduce the waste here.

(01:17:19):
And if you've seen everywhere,when you cut the waste, more of
it happens, it actually doesn'tshrink the industry ever, it
just makes more usage of it.
And it's a very simple thingwhich is like, if you think

(01:17:41):
about, there was, you know,there's like printing presses
and there's a bunch of booksbeing written, and then internet
came and then, like the numberof content just exploded.
You know, because we sort ofreduced any of the waste that
was part of.
Like you know, because we sortof used any of the waste that
was that was part of.
Like you know, when you printeda bunch of books, some of them

(01:18:01):
were never read and so there wasa cost borne by everybody.
It was very hard to access anddistribute.
Now, you know so, anytime, likeyou know, if you think about
like these big platforms, likebefore, like Shopify type
platforms, they were just likemaybe 30,000 online businesses
selling online, you know, andnow there's like 3 million and
because Shopify like reduced thecost of, you know, setting up a

(01:18:24):
store and running a world-classstore and setting up the
payments.
So every time you do that, justthe market like grabs it and
like expands in ways that nobodyever imagines.
You know, and I think the sameis true for logistics I don't
think people would think thatit's a fixed power industry.

(01:18:45):
I actually don't think it is.
I think the more efficient wemake it, the more of it will be.

Yeah, I'm trying to.
It's hard to like visualize it,but I think I'm with you.

I'm just trying to see what comes as waste
reduces, but I definitely I seehow it leads to much as we have
access to.
There is probably a coffee mugmaker somewhere in the world
who's unable to reach youbecause logistics is so
cumbersome and expensive.
Yeah, but if you cut the waste,that person will be able to

(01:19:33):
reach you.

Yeah, I'm with you, there will be more.
Yeah, yeah, okay, I see.

And so I think that's where the exciting part
is.
And it enables commerce and youknow, trade and commerce is
what civilizations are based off.
Right, like ultimately, youwill be kind of like need trade
and commerce to, to try so andneed trade and commerce to

(01:20:01):
thrive.
Okay, well, you went a lot inphilosophy here.

Yeah, we're getting deep there.
I was trying to think of how towrap coming off of that.
As we wind our time, I think Iusually go to some advice and I
definitely want some from you.
So give our audience a piece ofadvice on building businesses
in the logistics space.
What's your cardinal rule, thatkind of your most important

(01:20:32):
guiding principle, that everyday when you wake up you're
thinking I got to do this or thebusiness has to do this, that
aspiring entrepreneurs can maybegrab from you and carry into
their own world.

Yeah, listen, my learning in logistics is that
it's messy, it's complex, ithumbles you like no other
industry.
You know it is a very, verylike.
You know it is.
It is a very, very like.
You know my.

(01:21:07):
I tell you I'm perpetuallyhumbled by the complexity of
logistics.
Like there has not been asingle day where you, like you
wake up like I think I had theanswer.
Then, oh, there's another edgecase.
Oops, never thought about it,you know.
And but I think in that is thisbeauty, right, it's just this
messiness and complexities, butthis opportunity and the beauty
is that if you want to go deepand truly understand it, like
why?
And I think, like wrestle withit, like there is so much to be

(01:21:30):
done here, it's such a bigindustry, um, but but like don't
approach it with like hey, myviews.
Like don't approach it withlike hey, my views.
Don't approach it like I've gotit figured out from like 50,000
feet and like I know the answer.
No, like, go in, you know, gosuper deep.
Like spend time with theoperators, spend time with the
frontline guys, reallyunderstand what they do, why

(01:21:52):
they do it.
You know there is a reason forit and work with them to make it
better, reason for it and workwith them to make it better.
You know, like, don't likedesign a solution, like in a
room, and then like, hey, I'vegot the shiniest.
And then, like you know, comeand use it.
It's just very hard because thecomplexity is so high and it's

(01:22:14):
for a reason.
You know, I think my biggestrule in building businesses is
that get close to the operator,just get close, just spend time
with them.
You learn so much People whoare actually doing the work.
Maybe the CEO buys your product,Maybe they are the ones who pay
for it, but the real test is isyour operator loving it?

(01:22:37):
Are they using it?
But the real test is is youroperator loving it?
Are they using it?
You know, and guess what, like,if you can make that, set the
ladder happen where they love it, they use it and they get value
from it.
Yes, the CEOs will pay for it,you know.
But, like, don't get the CEOsto pay for it.
Just forget the operator Atthat point.

(01:22:57):
At some point in a few years,the shine's going to come off
and then it's like it will beunused, paid license, which
there are plenty in thisindustry.
That's another thing.
Oh my gosh, the amount ofunused paid licenses is a lot in
this industry.
And so that's my big thing,which is like get close to the

(01:23:21):
operator, like get as close,feel the problem, be in their
shoes, build for them, and thenI think you work, work, work
upwards.
If you do that, like roi comes,customer delight comes.
You know all the businessmetrics sort of start to
hopefully take shape.

Great advice.
Know that customer.
Yeah, know that customer Livein your customer's office.

Yeah, exactly Exactly.
Yeah, I mean, we have peoplewho are like batched in, you
know they're indistinguishabletoday from they are an augment
employee, or you know one of theone of our customers' employees
.
But also, don't just stop atthe prioritization of the

(01:24:06):
leadership, because thesebusinesses are so complex and
they're so large.
The leadership could think ofhey, I think these are my top
four priorities, that's great,but, like, go to the level where
they're actually using it andthen connect the two.
You know, sometimes there is aseparation, right, like or like

(01:24:30):
dissonance between those two.
Yeah, and it's important tobridge that dissonance.
But you can't do that unless youlike you know spend time.
You know that, unless you likeyou know spend time, you know
with, uh, with with theoperators, you know, like there
was one, one business we were at, like they're like, oh, we
don't, like we ban telegram, wedon't use it.

(01:24:50):
Okay, we, it's like that's notof use to us.
Like we know how you cantelegram, but we don't use it.
And I'm like, okay, we spentlike six hours with your
operators.
Every single one uses it, everysingle one uses it.

We didn't even talk about that.
The difference between what theleaders think the team is doing
and what the team is actuallydoing.

Yeah, it is because the dispatchers are in
Eastern Europe, because that'sthe only tool they use, you know
.
And so, and they don't theyhate calling the driver when on
the road.
So they say we want to justtelegram the dispatcher and
always get the response.
And well, you have to recognizethat as a true need, you know
right.
And so now, if you start tocall the driver through an AI

(01:25:39):
bot, who's not happy?
It's the operator.
So you're going above me, doingsomething that I hate doing.
It's annoying my carriers.
You're calling their driverswhen I have a method that's
working.
So lean in that method that'sworking, and so I think you have
to break that dissonance quitea bit, in my view.

Yeah, I'm with you 100%.
That's a great point.
It's a good one to end on.
We're coming up on time here,so listen, this was great.
I appreciate you playing thegame with me.
It's funny.
I had an idea I wanted to talkabout your past and these
companies, and then we startedwith.
Logistics is Broken.
That took us to Augment, and wejust spent an hour and a half

(01:26:26):
talking about that, which was agreat conversation.
At least I enjoyed it, so Ihope you did too.

Yeah, I loved it too.
No, you learn from all andsometimes, when you're even
speaking, you're like rethinkingand learning, but it's always
good to talk to.
You know an operator like you.
You've built businesses, you'veseen the pain, you've seen.
You know the highs and lows ofthis business right.
So it's always good to talkwith someone like you.
So thanks, andrew.

Well, I appreciate you coming on
cheering for you and your team.
You've hired one of the best inHaley Weiner.
Shout out, haley, she won'tlisten to this, but she's one of
my favorites that I've everworked with.
She'll do great for you.
No-transcript.

All Episodes

Episode Transcript

Popular Podcasts

24/7 News: The Latest

Stuff You Should Know

Crime Junkie

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Ep. #69: Harish Abbott, CEO & Cofounder, Augment

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}24/7 News: The Latest

Stuff You Should Know

Crime Junkie

All Episodes

Ep. #69: Harish Abbott, CEO & Cofounder, Augment

24/7 News: The Latest