Retailgentic | Consumer Behavior & Retail Trend

In this special Retailgentic episode, guest host Kiri Masters turns the tables on our regular host, Scot Wingo (CEO, ReFiBuy). They unpack the next big retail revolution, Agentic Commerce, and what it means for shoppers, retailers, and brands as AI shopping agents like ChatGPT and Perplexity redefine discovery, loyalty, and checkout.

Scot shares the lessons behind ChannelAdvisor, Spiffy, and now ReFiBuy, plus why messy product catalogs, ad overload, and loyalty programs will determine who wins in the AI marketplace era.

⏱️ Key Timestamps
00:00 – Intro: Kiri turns the tables
02:15 – Scot’s early career & first startup
07:50 – Building ChannelAdvisor
14:30 – Lessons from going public
20:05 – Spiffy and the services economy
26:40 – Discovering AI → the spark for ReFiBuy
33:00 – What is Agentic Commerce?
38:20 – ChatGPT & Perplexity in the marketplace race
45:00 – The product catalog problem explained
51:00 – How ReFiBuy solves it with AI
55:20 – Retailers’ edge: loyalty, data, customer experience
59:00 – The next phase: marketplaces 2.0
1:02:00 – Close: The future of AI shopping

👉 Connect with Scot: https://www.linkedin.com/in/thescotwingo/
👉 Learn more about ReFiBuy: refibuy.ai

🧠 Want to stay ahead in AI commerce? Subscribe and follow along:
📰 Subscribe to the free Substack: retailgentic.ai 
📺 Watch episodes on YouTube & Subscribe for updates: https://youtu.be/4xI9qXNU_R0

Creators and Guests

Host
Scot Wingo
CEO and Co-founder, ReFiBuy.ai
KM
Guest
Kiri Masters

What is Retailgentic | Consumer Behavior & Retail Trend?

Retailgentic, presented by the ReFiBuy Media Network and hosted by Scot Wingo, is where Agentic Commerce meets retail innovation. We help retailers and brands prepare for the future of agent-driven shopping with actionable news, expert insights, dxis, forward-looking predictions, and practical guidance on Agentic Commerce Optimization (ACO).

So the other thing to keep in mind is
ChatGPT is the 800-pound gorilla, and

whatever they do, people are either gonna
follow, fa- fast follow, or they're gonna

try to, um, be very, very different.

So I think we're gonna have some people
will try to do what ChatGPT is doing and,

and try to be a little better at it or at
least keep up with them, but then others

are gonna go in def- different directions.

So it's gonna be interesting
to see what that means.

So we're gonna see way more of the
big guys even maybe get into certain

verticals in a bigger way or, um,
you know, focus on a different part.

Maybe some of them don't even care
about the bottom of the funnel, and

they just try to be better at research.

Or to your point, maybe they try
to get better at being the one

that's awesome at social proof.

Welcome to the Retailgentic Podcast,
where we explore the intersection

of retail, e-commerce, and AI
agents, which we call retailgentic.

Our mission is simple.

We are shining a bright light
on the new agentic AI shopping

megatrend, which will disrupt the
trillion-dollar digital retail market.

The stakes are high.

Google, OpenAI, Microsoft,
Perplexity, Amazon, Apple, Meta, and

Anthropic are all fighting to be the
consumer AI shopping agent winner.

There are gonna be some tough decisions
for online brands and retailers to make.

What happens to checkouts?

How should payments work?

Should you continue to invest
in retail media networks?

What does the world look like when
over half of your traffic is not

from humans, but from AI agents?

These existential questions
are the tip of the iceberg.

On this podcast, we're talking to industry
experts, analysts, and influencers

to get answers to these questions,
identify trends and best practices.

This is Retailgentic, your
podcast partner in navigating

our AI shopping agent's future.

Hello, and welcome to
the Retailgentic Podcast.

I am Kiri Masters, and don't
worry, Scott is here with me.

You will be hearing from Scott.

Um, but, uh, I, um, I'm the host
of another podcast called Retail

Media Breakfast Club, and an advisor
to Refybuy, and a huge fan of the

Retailgentic Podcast right from the jump.

But I realized after listening to
several dozen episodes by now that

we haven't actually heard from the
big man himself, Scott Wing, go about

his views on agentic commerce, on the
future of AI-enabled shopping, and

what exactly he's up to at Refybuy.

So Scott, I'm turning the tables on you
for this episode to hear from you on, on

your perspective on all of these things.

Thanks for having me today, Kiri.

It's, uh … This is the most relaxing
podcast ever, 'cause I didn't have to

come up with a clever intro, and, uh,
you have a much better accent than I do.

Much more interesting, so I, I
always enjoy hearing your accent.

Yeah, I hear, um, people ask where, where
I live, and I say, um, I live in Atlanta.

Um, so I call it my, the, the- Yeah.

… the deep, the true Deep South, uh, accent.

It really is.

Absolutely, yeah.

We had an office at Chann advisor in
Melbourne, so I've spent a fair, a

little bit of time in Australia, so.

Yeah.

But you didn't, you
didn't pick up the accent.

But I, I, I've learned
how to recognize it.

I know there's a whole Vegemite
and, uh, what's the other…

Uh, I forget the other one.

There's all these interesting
inter- differences when

you go to other countries.

Yeah.

Yeah.

Well, everyone l- walking around
standing on their heads down there.

The toi- toilets flushing
in the wrong direction.

So one, one of the questions that
you, uh, always ask your guests is

about their career to date and how
you ended up where you are today.

So let's start there.

Yeah, I'll give the short version, and you
can answer and ask any questions you want.

So I have a engineering background.

So I decided to go, um, uh, to
University of South Carolina, which

is in Columbia, South Carolina, and
did a degree in computer engineering.

At, at that time, computer science
was in the math department,

and I knew I didn't wanna…

You had to take like four years of
very serious math, and I knew, I

knew I had a couple years in me, but
I didn't wanna take that much math.

So the computer, the engineering side.

And I thought I wanted to
work on hardware, but ended

up being better at software.

So I did that, and then I went to
Raleigh, North Carolina, where I live

now, and got a master's at NC State.

Um, and then, uh, got a job at a
startup in Connecticut and realized I

liked startups but not cold weather,
so moved back here to this area,

and now I'm on company number five.

So the, the brief tour of those is
first company was called Stingray

Software, and it was super nerdy.

We did developer tools
for other C++ developers.

So, um, that was a lot of fun.

Today, we didn't know how to call
it this back then, but it's called

DevTech, so developer technology.

So all the companies like Cursor
you hear about and that are just,

like, moonshotting are now DevTech.

So I like to think I was
30 years early on that.

Um, but that was fun.

That was a bootstrapped company, and we
ended up selling it to a public company,

so that was a really good experience.

Um, and then the, the proceeds from that
we took, and I became enamored with this,

uh, thing called the internet and the
web browser, and it was, uh, you know-

Marc Andreessen popularized with, uh,
what he did with Mosaic and whatnot.

And, um, I'm a big Star
Wars fan, if you haven't…

I think you probably know that already.

But I started, and I-- So I had this
deadly combination of, uh, I had made

some money from selling a company,
uh, the internet was just being born,

and I got really into auction sites.

And, uh, I'm a big
collector of various things.

Uh, and I started really spending a
lot of time on these new, new way of

doing things, buying online called
e-commerce and, uh, online shopping

malls is what we said back then,
which we don't really say anymore.

It's kinda funny.

And, um, I realized that there
was, uh, an opportunity to start

another company in this and ended
up starting a company called Auction

Rover, and it had three pieces to it.

It had an auction search engine, so it
could search hundreds of auction sites.

It, uh, had, uh, let's see, a buyer
agent, so you could say, "I wanna

buy this Star Wars collectible,
and it's on these six sites."

And the way auctions work is there's a
sequence of closings, and so it, it would

automatically put that together for you
and say, "All right, here's your budget.

If I don't win the first one, I'll keep
bidding," and until you win this item

for your budget, uh, if, if possible.

Uh, and then the last thing we
built was a tool that would let you

sell across multiple auction sites.

So that went, uh, that was crazy.

That was during the dot-com bubble, and
within six months, we had raised some

money, which is the first time I did that.

Uh, and then we started getting
acquisition offers and ultimately

sold that to the guys that invented
paid search called go2.com.

So that was really exciting 'cause I
really liked this idea of paid search.

Um, they ultimately, uh,
got in a, a branding dispute

with Disney, who had go.com,

had to rebrand to Overture, and
then ultimately got bought by Yahoo.

So it was, like, a lot of chaos
in that part of the world.

But they, they ended up buying us, um, so
I got to see the very early days of paid

search, was really fascinating to see.

Um, and then, um, Google started its
ascendancy, and the Overture guys

didn't take it very seriously, and
ultimately, they realized that was a

mistake, and they called one day and…

Well, a-after the dot-com
bubble burst, the only auction

site left was really eBay.

So you don't need a search engine
or a bidding engine, but our selling

tool became oddly popular with
what we now call power sellers.

So just, even just selling on eBay,
our, our tool was designed for kind

of a next level thing, versus on
eBay, you have this thing called Sell

Your Item, which is, like, very…

It's, like, an 8,000 form, you
know, field form you have to

fill out to sell one widget.

So we had streamlined that.

Uh, and then suddenly, big companies
started using that during the dot-com

bubble burst to liquidate inventory.

And so we had Motorola, IBM selling
millions of dollars worth of stuff

using this pretty simple little selling
tool we had built, and we started

beefing that up all on Overture's dime.

Then they called one day and said,
"We're gonna shut you guys down

because we're gonna go buy all these
other search engines, AltaVista,

Dogpile, and all these things, and
combine them all into ours to take on

Google and put them out of business."

That, spoiler alert, that didn't work.

And it went the other way.

And we said, "Okay, but we wanna…

We have this kinda wacky thing
that we think could be an idea.

We wanna, uh, basically buy
that back and spin it out."

And, um, we were fortunate in that
we got eBay to invest in that.

That was, like, their one
requirement, so we actually went to,

uh, kinda now famous Jeff Jordan.

He was just a regular dude back then.

Uh, and he helped fund that for us.

Um, so that, we spun that out
and called it ChannelAdvisor.

So that was, uh, the origin
story of ChannelAdvisor.

What year was that?

It's 2001.

Okay.

Yep, and the idea was someday Sites
like eBay are for people to sell to

people, but someday businesses will sell
to people on there because they, the

people are there, the buyers are there.

And these marketplaces, we no
longer called them auction sites,

now they became marketplaces.

We thought there'd be-- It's such
an awesome way to shop, we thought

there'd be hundreds of them.

Well, we were right, but it
took six years to get our next

one, which was very painful.

So the next one-- Fortunately,
the next one, these folks from,

um, Seattle called and they said,
"Hey, this is super secret."

Uh, it was Amazon.

"We're working on a marketplace."

Uh, a lot of people didn't take that
seriously at the time because you probably

remember this, but, but Amazon had tried
four or five times to have where you

could sell on Amazon, and ultimately the
marketplace is what, what-- that we know

today, the three P, one P, and all that
jazz, uh, was what really kickstarted it.

What did they try?

Uh, I'm actually not familiar with
that, like pre-marketplace days.

What did they try that didn't work?

The, the problem was it
was always a separate site.

It was like a microsite.

So they had Amazon
Auctions, Amazon Stores.

They even had…

They called it, uh, merchant.com,

um, where they actually
were the infrastructure.

They were basically the Salesforce
Commerce Cloud or Shopify for

Toys R Us, Borders, Circuit City?

I can't remember the other one.

Several, several of the really big
retailers that, that ended up not,

not going well for, for any, for
any of those peop- You know, most of

those people went out of business.

Amazon obviously didn't.

Uh, and, uh, yeah.

And then ultimately, the reason they
thought Marketplace would work is

they tried it to sell used books.

That's how they discovered it.

And the, the secret was, instead of it
being over here, all the people were going

to the PDPs, and they just hung it onto
the PDP where there was already traffic.

In hindsight, it's kinda obvious
that you should do that, but no

one had ever done that before.

They, they kind of-- And they
discovered it through the used

category starting on books.

Um, and then they just figured, well,
you know, we can also put new on there.

This is what we call a buy box today.

Yep.

And then it, it ultimately
became the buy box.

Yep.

Yeah.

So we leaned in on that very hard.

That ended up being very good.

Uh, long story short, we
raised ninety million.

We survived the great recession
of '08 and '09 just barely.

Um, and we got to the other side of
that, and every competitor that we had

had been cleared out by the recession.

And suddenly Wall Street decided this
thing called software as a service was

really amazing and the cat's pajamas, and
they wanted more companies to go public.

And we, we survived.

And it-- So it was like
Salesforce, Omniture, and, and

then we were kind of behind.

Salesforce was really big
at that point already.

Mil- you know, uh, probably like
three hundred million or something.

Omniture was knocking on a
hundred million, and we were

at like sixty going to eighty.

So we were able to go public in that,
that interesting circumstances that,

that kind of cleared out the, the IPO
window in two thousand and thirteen.

So as a, as an entrepreneur, that
was kind of always my, um, you

know, one of my bucket list items.

So that was a very fun experience to, to
take a company public and, and do that.

Are you glad, like overall w- uh,
because I'm, I'm sure that that was

like a really interesting experience
and, you know, en- entrepreneur bingo

card, taking a company public is one,
um, probably on a lot of people's bingo

cards, but, uh, uh, very few achieve

But I'm sure it brought about
a lot of challenges as well.

Like in sum, are you
glad that you did that?

Or if you had your time again,
would you have stayed private?

I think in the balance of it,
I'm, I'm happy we did it Yeah, I

think, I think in, you know, it,
it wasn't all positives for sure.

You're, you're right.

There's a lot of negatives.

The thing that's most surprising,
uh, as having now done it, what I got

wrong was I thought being a public
company would be great and there's all

this transparency, but there's oddly
less transparency when you're public.

Um, 'cause you have these hedge…

So, so the public company investors,
you have retail, which is individuals

that go on the stock market, but most of
it is, um, you know, uh, institutional

investors, and they have several buckets.

Mutual funds are good and they, they tend
to be very smart and think of long term.

Um, and then you have these hedge funds
and they're, uh, a l- many of them are

good, um, but a lot of them are, are
kind of, they do very dubious things.

And, you know, so what they do
is you, you have this period of

time where you know your results,
so you can't really say anything.

Uh, it's called a, a window,
um, and a quiet period.

And so they know you're going into your
quiet period, so they'll put out some

fake news about you and they'll say like,
"Channelyzer are shedding customers,"

and, you know, no one, they're, you
know, "We think it's gonna go down 80%."

And, and they do that where you can't
defend yourself, so it's kind of a,

a really nasty trick that they do.

So I, I thought we wouldn't
really, you know…

You see that happening with like
Carl Icahn and all these like big

guys on CNBC with like trillion
dollar companies and stuff.

And I thought we'd be immune to it being
like a smaller company, but no, there's,

you know, for everything bad that you
hear, there's like another hedge fund that

got, does it all the way down the stack.

So stuff like that was like super annoying
because, you know, I like, I kinda

naively believe that if we put a great
product out there and make customers

happy, that things will work out.

And they, they do in the long
run, but the short run is very

frustrating doing all that.

See, I'm in the, I'm in the wrong
business, Scott, because I rake some

companies over the coals occasionally
and I don't make money from it.

Yep.

So what you need to do is go put a short
on them and then rake them over the coals.

Yeah.

I'm just-

Yeah.

That was, uh, that was fun.

And then, uh, just to finish the
Channelyzer story, so I moved to

chairman in 2015 and started Spiffie.

Um, and then, uh, stayed on the board
at Channelyzer and then in 2022, we, we

got to the point where we realized being
comp- being public was hard and we needed

to like do some things to fix some stuff.

And there was a private equity
firm that wanted to combine us with

CommerceHub, who had like an EDI
solution and a lot of, lot of brands.

And we realized we were probably
over-concentrated in retailers for

what we did because a lot of retailers
ended up, um, at this particular moment

in time not, not surviving, right?

So Amazon's put a lot of big retailers
out of business and smaller ones.

Um, and then, uh, you know,
suddenly brands were coming on,

um, and going direct, uh, at a…

And we had some good solutions for
those, but we wanted to accelerate it.

So it made sense to kind
of do that transaction.

So we sold the company to, um, it went in
a take private to a private equity firm.

So I haven't really been involved with
that since 2022 when that happened.

Um, and then they rebranded it to Rhythm,
which I take no responsibility for.

It's R-I-T-H-U-M.

So I, I try not to have things
misspelled when, when I try to create

my own brands, but there's a lot of…

It's, it's very trendy to do that.

So like Miracle in our space and a bunch
of other companies have wacky spellings

So you mentioned Spiffy, but you
didn't explain what it is yet.

Yep, yep.

So the idea for Spiffy came during
the ChannelAdvisor Roadshow.

I was in New York, and, um, I
had my first Uber experience.

And as an e-commerce person, I thought,
all right, if we think about e-commerce

at a really big level, we've seen,
you know, physical retail go digital.

So analog to digital is like…

And Andreessen Horowitz, uh, has this
whole theme, software eating the world.

And I thought, if you look at GDP, uh,
uh, you probably haven't looked at US

GDP, but, you know, um, I don't know if
in Australia you study GDP in school, but-

Not on a daily, not as a daily habit, no.

Yeah, but in school you kinda
learn about it when you- Yeah

take an econ class.

I always just thought it was interesting.

Yeah.

And, and, you know, I'm a CNBC
junkie, so they talk about GDP

every quarter when it comes out.

And if you look at it, you know, one
of the, in, in the consumer side,

there's a big B2B part, but over on
the consumer side, there's consumer

goods, which is about a third, and
then two thirds is consumer services.

And, you know, you're, you have a
house, you're in Atlanta, I have a

house, and you've got kids and a car.

And if you looked at your dollars
you spend per month, it's way

more on services versus goods.

Um, you know, so, uh, that includes
things like restaurants, eating

out, um, the groceries, goods.

So, you know, if, if you look at that, a
balance, you look at the whole economy,

we're very much a service economy.

And I thought, "What's that
look like when it goes digital?"

And I, I thought, "Well,
it looks like Uber."

It doesn't have to be the…

It's an app, and it's like, it's kinda
like what e-commerce did for products

where, you know, it used to be someone
in Macy's in New York decided mint

green is gonna be the color, and that's
kinda like what you got in the store.

And then, then that all got democratized,
um, and by digital, where now, you

know, influencers influence what
the people at Macy's buy, and fast

fashion lets you almost, like,
pick what you want to almost…

You're almost at this customization phase,
but you also have total transparency in

the transaction and way more optionality.

So I thought, "That's
gonna happen to services."

So that was the idea of Spiffy is
figuring out what's a service that

we could really innovate on and make
digital and see, see what happens there.

So we, we chose car care.

Um, so we started with car washing, added
tires, oil, and started with consumer, um,

and then discovered, you know, the real
secret to unlocking that was more B2B.

So there's all these fleets, so rental
car fleets and that kind of thing.

So I did that from 2015 to,
uh, you know, late 2024.

Um, also around that time,
uh, I'd been on the shop.org

board with Jason Goldberg, so
started The Jason and Scot Show.

So, so he's, he's a ad agency person
at Publicis, and I'm a marketplace

person, and we always enjoyed…

He thinks a lot about stuff in
the store and how that's going

digital, um, payments, grocery, a
lot of stuff I don't know about,

so I'd always learn a ton from him.

And then I talk more about marketplaces
and Pure e-commercey stuff.

So we thought we enjoyed
our conversations.

We thought certainly
other people would, so

Why, of course,

obviously.

Yeah.

Why not, why not have a podcast?

So, so we did…

You know, we've been- we've
done that, and we're up into the

hundreds of episodes on that.

It's, uh, it's slowed down a little bit.

He's, he's kind of progressed.

I always give him a hard time, but
he's, he's actually, you know, I don't

know if there's, like, some number,
but he's gotta be, like, top five

people at that company or something.

So he's quite busy, and if his client
list, you know, e- every big retail

client is one of his clients, and
he probably talks about Walmart

and stuff, so, uh, yeah, I…

You know, when, when you have a client
like that, it can be kind of demanding.

So, so he's, he's less
available than I am.

Yeah.

And, and just a big shout-out
to the Jason and Scot Show.

I think it certainly in, in my
world, it was a huge influence.

Still is a huge influence.

I listen to every episode, probably
have listened to every episode

since 2015 when you started out.

And, um, I started a podcast not, a
couple years after that, um, being

inspired by what you, you all were
doing, and since then, the, you know,

the, the podisphere in e-commerce
and retail has just exploded.

So you gu- you guys were certainly
one of, one of the first, and

certainly still one of the most
popular podcasts in that space.

Yeah, when we were the first
and earlier, it was easier to

win awards, we, we, we found.

Yeah.

It's a lot, lot…

The competition's- Early run of
the game … a lot harder now.

When you're one of one, you're,
it's easy to really stand out.

What developed over time is, uh, Jason's
gotten more curmudgeony with, with age.

I'm a, I'm an eternal optimist, and he's
kinda, like, drifted towards more … And

I think people really enjoy the, the…

It's got, like, a point,
counterpoint kind of a thing.

So we, we end up disagreeing a lot,
so it's not a, it's not a podcast

where we end up agreeing many times.

Yeah.

Well, uh, yeah, uh, and that is
what I think people appreciate.

I certainly do because it is helpful,
um, to hear the other side, and we're

gonna get to agentic commerce in, in a
little bit and talk about some of the

skeptic, uh, skeptic's arguments, but I,
I think it is a very useful, um, practice

to listen to the, the other side of an
argument because these are, these are

arguments that your company might be
having internally about, you know, three

P versus one P on Amazon or, um, pretty
big consequential decisions, and it is

helpful to, to know, um, what, what, what
might be Argued on, on the other side

Yeah.

You can't-- The best way to win a debate
and be persuasive is to kinda know

what's coming at you and have a counter,
countermeasure to that and, and on down.

Yeah.

Right.

So I remember when you'd announced Spiffy,
and it was a huge, seemingly a huge

departure from your e-commerce roots.

But you ended up coming back
after, after that diversion.

So what brought you back
into e-commerce after Spiffy?

I, I've been watching AI very
closely, even from, like, GPT-1,

and started to get excited about it.

And once ChatGPT came out, I couldn't,
I couldn't help but think that this

was gonna … Th- as an entrepreneur,
this is a once in a lifetime

opportunity to start something in AI.

And I actually didn't really
have e-commerce as a thing, um,

that, that I … You know, I was
just kinda like, "I wanna learn.

I just have this voracious
appetite to learn more about

it, and I wanna learn more.

And then once I do, maybe there'll be
something that, that comes out of that."

So, so it was really, really just that.

I was kinda like … And Spiffy had kind
of gotten to a point where it was at

scale, and I had done, you know … It,
it really became a scaling thing.

And, and I liked … The part of startups
I like is that zero to one, they call it.

You know, where you kinda go
from an idea to some scale.

Um, and yeah.

So, so that was the general idea.

So in the fourth quarter of '24,
um, I left Spiffy to just kind of,

you know … I, I also have this
side thing where I do a little

investment fund here in the Triangle.

I try to support … I've had a lot
of mentors that gave very freely of

their time, so I try to pay that back,
um, to our little ecosystem here.

So I've done some things, um, that are
all under this umbrella term called

tweeners, which are startups that are
in that, that zero to one phase and

trying to f- They're a little bit past
what we call product market fit, but

they're trying to learn how to scale, and
I've done that three or four times now.

So, so I try to help companies there.

So trying to, um … It's not scalable
to do it one-on-one, so trying to do

it a little more programmatic through
some content and some other things.

So I have this little fund.

So I started spending
more time on that too

Right.

Well, th- that leads us sort of to
our present moment with RefiBuy.

I wanna get, I wanna get back to RefiBuy
in a little bit and talk about the

problems that you're solving there.

But, um, I wanna switch gears to just
start talking about agentic commerce now,

and then we can come back to RefiBuy.

So this is something that you've been
asking your guests about is with agentic

commerce, and maybe you wanna expl-
you know, define that first, what, what

is, what are your predictions around
what percent of, um, online shopping

will happen agentically in the future?

Yeah.

So, uh, the, my definition of agentic
commerce is commerce that, uh,

happens inside of a gen AI engine.

Um, and I, I like to call them answer
engines just to be very specific.

Uh, and, um, you know, it, it impacts
the shopper journey in some way.

I don't think it has to
always go to the transaction.

I have a pretty broad tent for
this, so it, it can, you know-- And,

and I like this framework research
find and buy the shopper journey.

And, you know, everyone has a different
way of saying it, discover and whatever.

There's, you know, consideration.

There, like, there's all these different
ways, but I like research, find, and buy.

Uh, and, uh, you know, so, so I actually
think-- So what percentage of e-commerce,

you know, the PayPal CEO has said 25%.

I think what's more interesting is
since I've kind of, kind of been

out but had a toe in and come back
in in a bigger way into e-commerce,

we've, we've slowed down dramatically.

And I think the root cause, if I try to
look objectively at that, and you're not

gonna like this answer, but I think one
of the root causes is we, we've overdone

it on advertising, um, especially Amazon.

Amazon and Google.

So, so they-- You get to this kind
of tipping point in ad load where in

the early days of Amazon ads, I think
it, it actually surfaced interesting

things you wouldn't have found before.

But now it's gotten way overdone
on the way they, they do ads.

And I think they've gotten
addicted to, you know, you could

see-- I don't fault them for it.

I, I would've fallen into the same trap.

But you could see maybe
they're having a hard quarter.

You know, they had this period
where Jeff Bezos left and a new

CEO, he's gotta hit the numbers.

And it just becomes so easy for Amazon,
I think, to hit their numbers by saying,

"Well, let's throw five more ads on, on
every, on every search results page."

And, you know, the, the way Amazon ads
works, they actually call it out now, it's

about a $60 billion line item for them.

And I-- You know, what is
the cost to serve an ad?

If the customer-- They've
already acquired the customer.

They're already paying the
cost to serve the page.

It-- They, they put some load
on that to make it seem like

it's not super high margin.

I think it's, I think it's, it's like 59.9

billion just drops to the bottom line.

So, so I, I, yeah, yeah, I think it's
just this pure margin juice that, that

has become pretty addictive to them to
kind of hit their numbers and grow a lot

of the other things they need to grow.

And which is fine until
there's an alternative.

Um, but I think that is what has slowed us
down, is a lot of this ad load throughout

the system has, has really made it
less exciting and fun to do e-commerce.

There's pockets of that.

TikTok Shops is one, um, you know,
where that kind of has, you know…

And, uh, we can talk about this, but
everyone for a very long time, Suthrith

is like the most famous skeptic on
this, that, you know, we'll never

have this intermingling of social,
social networking and in commerce.

And, and a lot have tried and, and only
TikTok seems to have succeeded with that.

There's an interesting path there
because I think that's, um, one,

one of these cases where it, it, it
didn't work a couple of times, and

so then there was the assumption,
"This will, this will never work.

This is not how people wanna shop."

But I think that there was sort of…

With, with a lot of con- uh,
drastic consumer behavioral

shifts and technological shifts,
like there's a warm-up period.

So, you know- Webvan launches, they
screwed up their economic model, it didn't

work, and then 20 years later, Instacart
comes, comes in with a different capital

structure and just win- wins the day.

And, uh, I, I think that that there
might've been a warm-up period with

some of these, like, social commerce
failures in, in the past, and when TikTok

Shop kind of arrives on the scene with
something that wasn't, like, really

dramatically that different to what
previous social commerce attempts were.

But at that stage, the, the habit
had been built in so deeply with

consumers that it just finally clicked.

And, and part of it was it just made it
such a natural part of the experience.

So you're watching the influencer
already talk about products,

and instead of going through…

They just basically tied it together
very nicely, something that was already

hap- It's kinda like the Amazon, how
it took a couple iterations to get

the, the product page listing, you
know, multiple listings on one PDP.

It, it took a-- I think TikTok had
a very similar insight, that they

kind of tied it together in a very
unique way that we, you know…

E- even, even on Instagram,
they haven't really succeeded

on doing that the same way.

Um, okay.

So we got a, we, we, we got a
little off, off track with the

agentic, uh, shopping predictions.

So, so I think, so, so I wish…

I'm more excited by the forecasts
that show agentic commerce

accelerating e-commerce, and I
think that's what's gonna happen.

And, and that, that gets-- It always blows
my mind where it, like, depending-- Jason

and I always argue about this, 'cause
you can take some categories in and out.

But somewhere between 16 and
18% of, of sales are online.

But it, but you and I are
probably, like, 80 or 90%, right?

Because we're, we're busy
people, we don't have time.

I still go to the store, and there are
certain things I like going to the store.

But, you know, I don't go to
the store for deodorant and,

and all these, like, things.

It just shows up, right?

There's huge category differences
in, in DIY and hardware, for example.

That is a category that's, that's been,
it remains very dominant in store today.

Um, different reasons, but yeah.

So, so you're not an either/or,
um, you're not arguing either/or

that agentic commerce is going to
take, take share from e-commerce.

You sort of see it as an
accelerant for e-commerce

taking more share from in store.

Is that-

I think, I think there'll be phases.

Phase one, it will raise,
rise the tide very quickly.

And you may even see e-commerce
itself expand because of it.

Um, and then phase two, if it can be
A significantly better experience,

then it will start to eat e-commerce.

And I, I think it's gonna get there.

Is-- I'm pretty good at predictions, and
I used to be pretty good at timing the

predictions within a year, and this is
very, very hard to predict in this space

because of this exponential growth curve.

It's like it messes with
your head because of…

You also have the factor of compounding.

So yeah.

So it, uh…

Like when we started, I thought ChatGPT
would come out with a checkout or a,

a marketplace in two to three years,
and it basically took seven months, so.

Yeah, I was actually just writing
something up today 'cause, um, Prime Big

Deal Day has just finished, and, um, I
haven't seen what Amazon self-reported

numbers are always super squarely, right?

They never actually tell you what-

It's always like, "We were excited
by tens of millions of things

that happened on some days."

And you're like, "What?

Yeah."

That's right, and you can't, you can't
square them back to previous years

'cause it'd be like, "We sold this
many TVs," and then the next time it's

like this, "We sold this many Xboxes."

So you can never really tell
exactly what's going on.

I wonder how big, what the team looks
like that obfuscates all this data.

They're like, "Nope, we've
talked about TVs before.

We can't talk about that."

Yeah.

Well, I'm, I'm sure they're
using AI for that now.

But I think that, uh, from what I've
seen, and there was a lawsuit actually,

um, filed a, a few months ago after
the July Prime Day, um, event And it

was ar- around, um, price manipulation
and that the deals that you would

see on, on Amazon during those events
were not actually proper deals at all.

And, and this isn't like
Amazon actually setting prices.

In many cases it's a, it's
a merchant, as you know.

But this is one area that I think that
agentic commerce is really gonna solve

for, which is I bought a few things on
Prime Big Deal Days a coup- few days ago.

I'm not sure if I r- I got a genuine
deal, and I'm, I'm in a situation where

it's n- not a material thing for me.

But I think about the fact that many of
the, the US states are in a recession now.

People are very cautious.

Tariffs are increasing prices, and there
is a lot of distrust amo- uh, between

consumers and retailers about just
whether they're getting a good deal.

And this is a terrific use case for
agentic commerce because th- what,

what better way to price compare in
seconds across different retailers

than through an AI, um, sidekick?

Yeah.

I think the thing that takes us from
phase one to, to phase two in my

methodology there is agentic browsers.

Because you can kinda get there with
browser extensions, but they're always so

hacky and, and annoying, and they don't,
they don't work across devices and stuff.

You're like, gotta install the
Honey app everywhere and all

this jazz to get a good deal.

So, so if you're gonna be a
value hunter, it's a lot of work.

And, but, you know, so the
convenience-oriented consumer

loves to save money too, so they
like convenience and saving money.

If we can do both, that's even better.

Like on your Prime deals, if you could
find that they were, you know, 20%

less, you would've done it, even if it
may have taken longer for them to ship.

I think if we start living in an
OpenAI or a, uh, you know, a ChatGPT

browser and it just does all that
for us, it, it will make, you know…

And, and just like having this persistent
agent across devices and, and things,

I think it, that's how you get to this
level, this, that whole next level.

Yeah.

I think th- th- there's a few, the
one, one argument that I've heard

make is like, okay, e- e-commerce
today, it's, it's pretty simple.

Like what, this is a, um, a
solution in search of a problem.

That's what I've heard a number of times.

And I look at online shopping
now, I'm like, well, it actually

is kind of- Annoying, as you
said, there's ads everywhere.

That's one issue.

There is the opaque pricing issue.

There is still inconsistent product
information where, like I was looking

for a pair of boots the other day and
product listing, first it says it's

leather, and then it says it's synthetic.

And I'm like, "Well, luckily I
already researched these boots,

and I know that they're leather,"
otherwise I wouldn't be able to tell.

Um, and then manipulated
social proof is another one.

We have ratings and reviews that
are gamed so frequently, um, that

it's hard to tell what's like a gen-
what is a genuine review or not.

And then, you know, other things
like returns, policies, shipping, you

know, prices and things like that.

So I th- I think that
tho- those are among…

I, I tend to disagree with
you about the advertising.

Well, the, the search experience
is broken too, and, and, you know.

So all this adds up to the discovery
experience is really broken.

So- Right … you know, and what's
happened is we've optimized for keywords

so much, the search, the search bar works
best if you enter like three or four

words, and if you get really descriptive,
it just like falls off and it gives you

like just stuff that doesn't match at all.

E- and Amazon used to be immune from
that, but they've, they've fallen into

it, and I think it's a combination of,
um, you know, over-optimizing on keywords,

people, people gaming the SEO on Amazon,
you know, the intra-Amazon SEO, and then

also, um, the lack of product data, and
then just the massive number of products.

They have something like 600 million
SKUs, and when you have that million,

that many SKUs, it becomes impossible
to find what you're looking for.

All right.

So did we…

So d- did you have anything else that
you wanna add on, on projections?

You're, you're thinking about this a lot.

So the other thing to keep in mind is
ChatGPT is the 800-pound gorilla, and

whatever they do, people are either gonna
follow, fa- fast follow or they're gonna

try to, um, be very, very different.

So I think we're gonna have some people
will try to do what ChatGPT is doing and,

and try to be a little better at it or at
least keep up with them, but then others

are gonna go in def- different directions.

So it's gonna be interesting
to see what that means.

So we're gonna see way more of the
big guys even maybe get into certain

verticals in a bigger way or, um,
you know, focus on a different part.

Maybe some of them don't even care
about the bottom of the funnel, and

they just try to be better at research.

Or to your point, maybe they try
to get better at being the one

that's awesome at social proof.

So, so I think we're gonna
go through a phase of kind of

sameness and then like this big
dispersion of what everyone's doing.

Yeah.

And ca- you've spoken a little bit about
categories as well, and there's some

that have distinct sort of needs and, and
levers to pull, like fashion, for example.

That seems to be an
interesting category play.

Yeah, there's like 10
vertical startups there.

Um, Daydream's the big one that,
that you hear about, but there's

another, like, uh, like 10 to 20.

There's new ones every day, and, um,
I've gotten to know a lot of the agentic

payment people, and they work with a
lot of these people, so they give me a

little heads-up that all these are coming.

Um, so there's a lot of those.

Uh, most of them are in apparel.

There's a couple in grocery, uh, and yeah,
if I look at the history of e-commerce,

the ones that made the most sense that
survived were in auto parts and B2B.

So, so that's kinda my prediction, that
there's things about that category that do

require a very different buyer experience.

Um, but you know, maybe, maybe
I'm wrong and, you know, some

of these fashion ones are…

I, I'm a little, uh, g- some blindness
to fashion 'cause I basically

wear the same T-shirt every day.

So I am not a fashionista.

Uh, uh, I don't care if my
boots are leather or synthetic.

I think I've got the same
T-shirt as you, actually.

Well, you do, but- Yeah … yeah.

And I also spread the T-shirts.

Uh, uh, and the…

You know, but I do, you know, whenever
I s- uh, I was at retail, um, club, and

several women in the thing I was doing
were lighting me up on that, that no,

fashion was very different and, you know,
they need a stylist and this kinda thing.

And, you know, I, I, I'll, I'll
keep an eye on that and see

if I'm wrong, but I just think
ChatGPT is probably good enough.

And, and by having a, just the
amount of resources they have,

it's hard, it's hard to acquire
customers in the consumer world.

It's just way, way harder
than you can imagine.

Do you think that this could be different
to, to e-commerce in, in the fact that,

uh, lots of these consumer surveys
that I've seen where people are asked,

"Okay, hypothetically, if we think
about a fu- a future with agents…"

First of all, most of these studies
have to first explain what a shopping

agent is, like what the concept is to
consumers, and then they either hear

about it and say, "Yeah, that sounds
great," or, "No, that's terrifying.

I don't wanna have
anything to do with that."

And then within that, um, i- if
people are interested- The, the

commonly stated category that they're
interested in is grocery and household

purchases, because that is tedious
for most people, it's monotonous.

I'm buying the same things.

I have stated preferences for different
categories, like I, I, I will buy private

label in these categories, but I must have
my, like, preferred coffee brand or else.

Um, so that's like a stated interest
from consumers that they would like

to have some automation with that.

But then where there's also, I think, an
amazing use case for fashion or something

like that, where, where there could be
a more immersive shopping experiences.

And this is where I think like the
current sort of, um, use case of AI

is still limited to what we could
see in the near future with video and

virtual try-ons and things like that.

So those two categories are at completely
different ends of the spectrum and,

and actually tend to be the ones that
consumers at least say that they are

interested in experimenting with.

Yeah, I think this is
what gets me excited.

We're gonna have a lot of magical new
experiences we never experienced before.

So, you know, the one I, the one I tend
to gravitate to is when it, it surfaces

things that I hadn't thought about that
are, like, really interesting to me.

And, you know, so the, the
pulse thing that ChatGPT is

doing is really, really good.

And I could see that being, you
know, it may surface to you and

say, "Hey, Kerry Falls here.

I know you've been looking for boots.

Here's a couple cool ones,
and I know you like leather."

And, you know, so that, that's
like this really cool, magical…

It, it's kinda like that stylist
thing I think people are looking for.

But it's way-- It's more
proactive than reactive.

Almost all of our
shopping's very reactive.

Um, yeah.

And, you know, so some of the proactive
stuff it's gonna be able to do so it

knows you so well is gonna feel like
this, this really, is really helping

you in a, in a kind of magical way.

So, um, let's talk now a bit more about
what RefiBuy solves within that ecosystem

where, where-- 'cause you've just sort of
let the cat out of the bag at, at Retail

Club that you mentioned and, and starting
to work with retailers and brands on this.

Yeah.

So, so to understand where the
idea for RefiBuy, you have to go

all the way back to Channeltizer.

And the, the surprising thing about
Channeltizer is we learned very quickly

once we started helping big brands and
retailers sell on marketplaces, the--

there's this weird thing in e-commerce
where everyone's product catalog, which

is, you know, it's pretty important.

It's what, basically what
you're selling online, and

it's a, it's a hot, janky mess.

And it's because there's no standard.

So, so I'm an engineer and, you know,
when you buy, like, anything in…

If you make a car or anything like that,
every part has a specification, and you,

like, have to know, you know, what's the
voltage of this thing, and all this stuff.

Product catalogs have no specification.

So your product catalog
could be totally different.

Every one is actually very
bespoke and different.

And that's because every company from,
you know, Amazon and Walmart all the way

down to Kerry selling some stuff on eBay
has a different way of managing this.

You know, down here you're
using spreadsheets, and you're

using maybe Shopify, and you're
doing some copying, pasting.

You got some little tools.

And then up here you've got SAP
and some ER- you know, all these

different enterprise systems.

So, so ultimately, the product catalog
is the stuff, the data you paste

together in some way to get a PDP up
and sell stuff and track inventory.

So when these things would come to us,
it was very interesting 'cause, 'cause

our job at ChannelAdvisor was to take
literally thousands of companies' catalogs

coming at us every day in real time,
and then get them out to, you know,

hundreds of marketplace on the other side.

Well, the marketplaces on the
other side, they're the opposite.

They have very strict requirements.

Um, so strict that, uh, you know this,
but I'll say it for the audience.

So strict that it's-- we have
to take that very seriously.

So, so Amazon and eBay, I'll just pick
on them, but every marketplace has this.

If you, if you mischaracterize a
product and the buyer gives you a ding

on that, there's several weight forms
that come in, like a return and stuff.

One of them is a significantly
not as described.

We call it a SNAD.

And, uh, if you get SNADs on Amazon,
you get three of them, your business is

basically kicked off Amazon for life.

No judge, no jury, no Done.

You know, that's it.

One shot, you're done.

No speed bumps, you're dead.

And, you know, so we're sitting here r-
helping people run, you know, Dyson's two

hundred million dollar Amazon business,
and if we mess up, you know, the image

on a vacuum cleaner and, you know…

Also, it's, it's very insidious
because it's, like, transactional.

It's not just the SKU.

Like we, we can't say,
"Oh, that was just one SKU.

That's one SNAD."

It's like three people bought
the wrong thing, we're toast.

So it's very, very important that we take
this completely unorganized data and get

it to these guys in an extremely organized
and va- and validated way, or else.

So then the other thing you could do--
So one side is like, "Oh my gosh, if we,

if we give a lot of data, that's risky."

So then you're, you're like,
"Well, I'm incented to give

as little data as possible."

But then if you didn't send
images, it wouldn't sell, right?

Or if you didn't give as many
attributes like is this leather

or not, it's not gonna sell.

So then you, you have, all right,
be cautious and don't sell anything,

be aggressive and get nuked.

So we had to, like, figure
out how to navigate that.

So the long story short, um, it
ends up being a very hard problem

that I've spent thirty years in.

We've developed our whole own vocabulary.

But, but I'll give you a couple examples
so you kind of, like, understand.

The first thing that you have
to do is canonicalize all this.

So, so it's like I gave you sixty
decks of cards all scrambled, and I

said, "All right, Kerry, you need to,
like, unscramble these because we're

gonna send them to eBay and Amazon."

And, you know, and all-- By the way,
they all, they all have, like, very

similar, you know, um, not face cards.

So, so that's what we call
canonicalization, getting them

all into the where they belong.

Uh, and that's hard.

Then once you canonicalize,
you have-- You're like, "Okay,

these products are the same."

And what happens is this.

Uh, you have some-- These are attributes,
so now we know these are the same product.

Some of the attributes
are missing, so we know…

So we have two products.

Let's say it's a pair of boots.

Um, and, you know, one of them
says leather and one doesn't.

Well, it, it doesn't have anything.

So, so what do we do with that?

So there's, like, this overlap thing.

There's-- Each one has some pieces
that don't overlap and some that do.

When they overlap, if you're, if you have
a good situation, it'll be like color.

Uh, we have these in four colors,
you know, uh, X, Y, Z, and A, and

th- they all line up and you can just
collapse that one and say, "Okay, good.

Those are all the same color.

We have this.

We can, like, collapse that."

But what happens is, um, they almost
always, um, collide with each other.

So someone will say…

So, so, you know, we do this at scale.

So a good example is someone
would say, "All right, here's

an Under Armour, um, shirt."

And we would say, "We think
this comes in six colors."

And then we would look at all the offers
coming in, and we'd have fifteen colors.

W- well, someone has to do something
with that, and, and that's us.

So we, we could programmatically
do it with some things.

So if it's data, you can kind
of, like, size information.

So everyone's uses different things for
sizes, so we could, like, normalize that.

Category, all this kind of stuff.

Um, but as you get into the some
of these problems, they became very

subjective or non-deterministic in,
in a computer science-y kind of way.

So no algorithm or heuristic would do it.

A human had to look at it.

So colors is a really good example.

Um, even though you, you could
argue, can't the computer look

at the pixels and everything?

N-not really, um, because it's on
a shirt, and then, like, it's kind

of like it's hard to pick out…

You know, these colors
aren't just like red.

It's like, you know, what is the color
of this letter or stuff like that.

And so then the computer couldn't,
like, figure out where on here, where

on the shoe is this orange piece
'cause it's the tongue and the, the,

uh, lower and, and all this jazz.

It, it's not solvable with
computers, Jeff, trust me on

that, and, uh, prior to 2002, '22.

And, uh, uh, okay, so the only
way to solve it was people.

So we had a team of people
in North Carolina, and then

that team grew to be so big.

At, at our peak, we had, like, $20 billion
worth of software, of products going

through ChannelAdvisor annually, which is,
like, you know, a lot, a lot of products.

And we had a team, we had to
out-outsource and outsource again.

Ultimately, it ended up in Bulgaria, and
we had three hundred and fifty to four

hundred Bulgarians working on this 24/7.

And they were organized by category
teams, specialist, escalation paths,

junior, senior, and, you know,
analysts, all this kind of stuff

to kinda try to optimize this.

And it was a huge drag on our
business, but, you know, we were an

$80 million, $100 million business,
and it was, like, $2 or $3 million.

So it was just a cost, and it was
this important we had to do it.

Okay, any questions on that?

So we'll call that the
product catalog problem.

Every…

The other thing is, this is the dirty,
other dirty secret of e-commerce is

we'd go to shows, and I'd talk to other
competitors and friends and people.

I'd say, "Do you have
this catalog problem?"

They're like, "Oh my God, I
have 500 people working on it."

We'd be like, "Whoa."

So every retailer, every marketplace,
every brand, all the retail media network

people, they want cleaner catalogs.

This catalog thing cuts across
every part of e-commerce.

It's, it's pervasive.

All right.

So, um, so I'm reading…

So back to fourth quarter 2024.

I hadn't thought about this thing for
10 years, and I'm reading up on Gen AI,

and Anthropic put out a paper that said,
"Hey, we have this cool new thing called

agentic Agentic AI, and we're gonna…

Here's a white paper.

And, um, and in there they said, "All
right, y- what we found is an interesting

way to use these AIs is you take the
LLM, the gen AI, and it's like the

little brain, and you pair it with
some memory so it can learn, um, after

training, and then you need to give
it a little database so that it has

knowledge after the training cycle too."

And they call that
retrieval-augmented generation, RAG.

It can come in different
formats, but for simplification.

And then the other-- So that's like
three pieces, and then the fourth

piece is you give it some tools
so it could actually go do stuff.

Otherwise, it's just read-only,
and it's just like a chat that

you can just, like, talk to.

Um, so if you want it to do things
and have agency, thus the word agent,

it has to have a toolset to do that.

And then they invented this thing that
lets them use APIs, basically called

MCP, and that we read a lot about that.

And then I was like,
"Well, that's really cool.

I like that."

And they said, "But even cooler is think
of that as a Lego block, and then we

have come up with six different ways
you can configure these Lego blocks."

And one of them was called an evaluation
network, where you could have three

little LLMs, and they could have
different brains or different memories and

retrieval-augmented generation, and what
happens is they form different opinions.

So you could have them look at
non-deterministic problems and vote.

Well, this really resonated with me
because this is how we would set up some

of the Bulgarians on the Harrier things.

We'd have three or four of them
vote on stuff, and if, if, if it was

unanimous either way, you knew you
had an answer, but if it was like

a tie or close, you'd have to kind
of elevate it again, escalate it.

So-

It's like scientific.

Yes.

So I was like, "Uh, wow, I
wonder if we could solve this

catalog problem with this."

So talked to-- We ended up, um,
building a pretty big team in Ireland at

ChannelAdvisor for this because some of
these problems in, in canonicalization

are similar to internationalization.

And, uh, the people in Ireland are
very good at internationalization

because they know English, but then
a lot of them are also multicultural.

And the programmers there, there's,
there's certain universities that

have taken that as a focus area, and
they have just, you know, degrees in

internationalization So that, that
was-- we, we ended up having a team

there because this was, like, the core
problem that we had at Channelazer.

So talked to one of the superstar
engineers there, and he had left

Rhythm and gone to Micmac, and
at Micmac, he had tried using

retrieval-augmented generation.

And we got to 50% of solving
this problem at Channelazer,

and that was, like, heroic.

Um, they got to 60% at Micmac using
retrieval-augmented generation.

Uh, and I was like, "Huh."

And he said, "I just read this similar
paper, and I'm kinda curious too."

So he went off, and in
60 days, um, I said…

He said, "What should we try?"

And, uh, I said, "Let's try
pickleball paddles," 'cause I had

been reading a lot about, like,
what a messed-up category that was.

And, um, within 60 days, we
had-- he had built the perfect

pickleball, um, paddle catalog.

So it was, like, 200 paddles, um, you
know, rows by 80 perfect attributes.

And then we put a chat in front
of it, and it was the best buyer

experience ever 'cause it knew…

The first thing it knew to ask
you was, "Are you beginner,

intermediate, or advanced?"

And you'd say beginner, and
say, "Oh, here's what you need."

And it would just point you right
to the, like, three paddles.

It was amazing.

And it's because it
had all the attributes.

Okay.

So I was like: Wow, that's pretty magical.

What do we do with this now?

And then right around then, I
had been consuming more content.

I'm a big Perplexity fan, and they
came out with Bion Pro, and I thought I

think I know what's gonna happen here.

I think these guys, ChatGPT,
Perplexity, and these guys have

captured consumers' imagination.

They're gonna need to monetize that
because we can all spend-- see they're

spending hundreds of millions of
dollars on compute GPUs, and the best

way in my world to monetize thing is
a marketplace, and they're gonna…

And I thought-- Or ads, you know,
uh, but I, I don't think they're

gonna do ads because that's their
differentiator, um, is they can't…

You can't compete with Google
and with ads, um, because they

will always win that battle.

So you have to come at them
from a different angle.

So I thought these-- a lot of these
are gonna become marketplaces, and if

that happens, everyone's gonna need
a much cleaner catalog than we even

had to worry about in marketplaces 1.0

because, you know, they-- these
engines crave context and content

like, like you've never seen before.

Uh, and yeah, so we started
building towards that in

basically January of this year.

And so what is the sort of, uh,
ICP that you're working with today?

Yep.

Retailers and brands, um, they
tend to be larger in SKU count.

So if you have two SKUs, you know, we
can help you, but it's not gonna be

like, you know, you can do that yourself.

Um- Yeah.

It's not a super painful pro- pro-
uh, problem for them right now.

Yeah.

You can spend some time and
it's, it, it's not fun, but it's

not like a, you know, months.

Um, and you know, so, so you need kind
of like, you know, 1,000 SKUs, 5,000

or whatever, um, and some complexity.

So the categories we, we're not gonna
do, um, any B2B right now or auto parts.

Um, we'll eventually get there, but right,
right now it's like some mostly consumer

stuff minus, um, we're probably not gonna
do anything in like the old media, book,

music, video, video game categories.

But we do a lot of fashion, beauty,
um, uh, home improvement, those,

those kind of classic categories.

And so h-how do you see…

So you mentioned B2B and auto are
categories that you're not touching yet.

What else do you sort of see as being
on the, on the horizon in the future

once you've sort of like solved this,
this problem for retailers and brands?

So the customers we have are already
asking us for a lot of other things

because they really like this capability.

So we, we kinda have to-- It, it's a
super target-rich environment, and I, I

don't know what direction we're going.

The honest answer is I don't know yet.

But, um, we just kinda, you know…

As an entrepreneur, I have to
fight this instinct to have a

little bit of entrepreneurial ADHD.

So I like to go, you know, but I--
we, we need to build this first.

So, so we, we have the discipline, we're
gonna build this first, and then we'll…

Along the way, we're, we're looking
at those breadcrumbs to kinda figure

out what's next, but I don't know yet.

This is happening so fast that, that
it is mind-- You know, I thought

we'd have two or three years before
this marketplace thing happened.

I thought, "Oh, this will happen half
as fast as the eBay to Amazon thing."

But then it happened in seven months.

Um, and you know, so…

And we got very lucky because the feed
that ChatGPT wants is basically what we

built because, you know, uh, yeah, because
they, I think they basically kinda- They

came to the same conclusion that says,
"We want as much, you know, we want way

more attributes than you're gonna give
anyone else, like in a Google feed."

So, so they have this like super
wide data feed that they want,

and that's what we basically are
set up to help people create.

An-another one of your thesises that
you just shared was the LLMs aren't

going to monetize through ads, they're
gonna monetize through marketplaces,

which certainly appears to be happening
at, at least as the first priority.

Do you think that they will?

And what would…

If yes, what kind of timeline
would you be looking at there?

So ChatGPT is.

So I, I view what checkout is,
is they're built a marketplace.

Um, now it's a very unique marketplace.

Um, Sam Altman calls it an affiliate, and
I think it's just a nomenclature thing.

In my, my purest e-commerce view
of, of affiliate, the transaction

happens on the retailer site.

I, I think he's calling it an
affiliate because it, the, the

retailer's still merchant of record,
um, but they don't leave ChatGPT.

So, so in my mind, that's a marketplace.

When you don't leave the, the, the
site that the retailer is the, the

retailer is not on that site and you
check out up here, that's a marketplace.

I'm, I'm not married to that definition.

Marketplace/affiliate.

Uh, I just, I…

Affiliate I think confuses it, like
I've got all this history of affiliate.

Um, it's-

Sure.

It's, it's own

industry-

It's its own thing … for people

like you and me.

Yeah.

Yeah.

Yeah, yeah.

Yeah.

Uh, and again, I think we're gonna see
everyone follow this pretty quickly.

And I think there's gonna
be several modes of this.

So the one that Google's doing is more
what-- A lot of people wanna argue is

what ChatGPT is doing agentic or not.

We, I, that's like a silly argument.

I, I don't really care.

Uh, it's just a great buyer experience.

But I think what Google's gonna do is
gonna be interesting, where they're

actually gonna have an agent go fill
out the checkout, and, you know,

that's gonna be a different mode.

But it's gonna be a…

In my mind, it's still a marketplace.

It is an agentic marketplace because you
never leave the merchant site as a human.

The human never leaves the merchant site.

The agent goes to it for you.

It, it ne- it feels like
you're still on Google

You know, I th- I think a concern
that retailers had initially is

that the traffic is dropping off.

There's now an intermediary sitting
between us and the consumer.

Is this, um, is this gonna challenge
our business model and, uh, uh, is

this gonna cause sort of radical
commoditization amongst retailers?

I personally have a number of
arguments as to I don't think it

will cause commoditization, but,
um, I'm curious what you think.

I think you can't take-- I think a lot
of that is colored by our experiences

with kind of the duopoly of, or
triopoly of Facebook, um, Amazon, and

Google, uh, where, where merchants,
uh, that you and I talk to have been

of all size, especially Amazon sellers.

I-- You and I know a
lot of Amazon sellers.

They're like, uh, you know, they, they,
they're-- they have-- They get punched

in the face like five times a day.

That's a rough living.

And I think we have a new competitor
that realizes they're not gonna win by

doing that, and the way they win is by
not doing that, and, and by changing

the game a lot, and not doing ads, not
punching their partner in the face,

and that's how they're gonna win.

Now, everyone can point--
everyone says Facebook said

that, too, and then they changed.

Yes.

Um, but again, I think the
people have learned from that.

And, you know, Sam Altman has been--
He's, you know-- He, he's got a very

long history of this and watching
it, and he's very, you know, uh, he's

probably one of the smarter people
that's ever tackled this problem.

So I think it's gonna
be really interesting.

I think there's a good shot that
it doesn't happen like that.

And even if it does, wouldn't you rather
have four bullies instead of three?

Yeah, no, I, I think so.

I mean- … every, every revolution
in retail has, uh, been a

reaction to a set of issues that
someone new was looking to solve.

So Amazon being a great example of
just creating a completely different

experience that solved for, for problems
that people, like, didn't necessarily

immediately understand were problems.

But over the last f- 20 years,
like, the, there has an…

There's no grand master plan at Amazon,
like, "Let's make merchants' lives as

difficult as possible," or, "Let's,
like, let's make this e- experience for

consumers super opaque so they don't
know that they got a good deal on,

on Prime Big Deal Days or whatever."

I don't…

There's no evil master plan, but it
is compounding decisions that are

made o- off of a, um, a way of doing
business that ultimately there is a,

there is a opposite response out there
in the market and, and that is, you

know, the, the free, the free market.

Yeah.

I'd rather have four people
fighting for my business.

The, the other signal, the merchant
of record signal is very unique.

Um, so, so being the merchant of
record now makes it infinitely

easier to do loyalty programs.

And- Yes.

I love that … this is, this is…

So another way to kinda get
at, to, you know, let's pretend

we're in this debate, you know.

Well, how do you…

All right, so I would say to you, Mr.

Retailer, "How do you avoid,
um, you know, commoditization

in, in r- in physical retail?"

And one of the answers is loyalty, right?

The loyalty program.

So, so building a relationship
with your customer.

So number one, by being merchant
of record, it is for the first time

your customer, and which is very
powerful in, in our marketplace world.

You know, I can point to 200 marketplaces
where you're not, you're not the merchant

of record, and you're just borrowing that
customer, and you can't re-market to them.

So, so this is super unique.

I think Jet did it, but I can't…

It no longer exists, so I
can't double 100% verify that.

I think they came close to it.

They, they would give it to you, but you
weren't technically merchant of record.

They were more merchant-friendly.

They were the most merchant-friendly.

And, uh, so I think we'll see by next
year this checkout will be enhanced

to have a loyalty thing and a bunch
of other stuff that merchants want.

And, and that's how you
fight commoditization.

Absolutely.

Yes.

I, I, I completely agree.

I think that the loyalty programs are
gonna have a real moment and come back

and be more exciting and dynamic than
ever, and it will be a significant way for

retailers to add value beyond, like, l-
lowest priced leather boot is one thing.

But, like, if I'm, if I'm loyal to a
certain retailer because of a, a program

or because of other, other factors
as well, like delivery time, ease of

returns, um, there, there, there's so
many non-price factors that I think

that retailers can differentiate on.

I don't…

I, I think that there's, um, just
a, just a need for retailers to, to

go back to their reason for being
and understand, like, what, what,

what are we offering that's unique?

It doesn't have to be price.

That, you know, the store experience
will, you know, a lot of people like

having the abi- ability to return to
the store and, you know, and that,

that harmonized or omni-channel
or whatever you wanna call it.

I, I think you'll be able to…

Because you're gonna get that
customer, you're gonna be able

to tie it to those things.

Whereas in the non-merchant of
record, you, you have to go through

all this weird, you know, reading of
tea leaves to figure out, is this a

new customer, and what's going on?

Yeah, absolutely.

Well, we've, uh, it's happened again.

We've used up a perfectly good hour
of our, uh, listen- listeners' time,

one of my favorite catchphrases
from the Jason and Scot Show.

Thank you for, um, allowing me
to flip the tables on you, Scott,

and, and dig into your history
and what RefiBuy is all about.

Um, I'm sure that listeners will
appreciate, uh, understanding you and

what you're up to a little bit better.

Sure.

Thanks, Kerry.

This was a lot of fun flipping the table.

Follow, rate, and share to stay up
to date with RetailGenic, your guide

to the future of AI shopping agents.