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).
One of our co-founders,
Idan, he's good in writing.
I call him the chief writing officer,
and we started a blog in November
24, I think it was the first one, and
we just wrote what we think there.
Honestly, everything that we think
or know, we usually write there.
Uh, and that really kind
of changed, uh, the tone.
I think people search for kind of
thought leadership and opinions and
what's coming and makes sense, and
that what brought us Visa, for example.
They read our blog.
So I think that really helped,
uh, organically for us, and
it's relatively small market.
It's, you know, uh, dozens of, of,
of agent developers who are already
funded and not indie developers.
So everybody at this moment is
searching for, uh, tell, tell them
what to think, what to do, how to act,
and, and some thought leaderships.
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 future, and now
here's your host, Scot Wingo.
Welcome to the Retailgentic Podcast.
One of our most popular topics on
the Substack is agentic payments.
In this episode of the podcast, I'm
excited to introduce you to Eyal
Karni, CEO and co-founder of Nakuda,
which you can visit at nakuda.ai.
Nakuda is one of a couple startups working
on innovative agentic payment solutions.
This solution is designed for developers
of agents to use so that they can
manage consumers' payment mechanisms.
According to Crunchbase, they
recently raised five million
dollars from one of the top consumer
venture capitalists called Madrona.
That round also included the
strategics Visa Ventures and Amex
Ventures, which is a big sign of,
uh, approval when the two big payment
companies invest in your company.
In this interview, we go over Eyal's
background, the Nicuda founding
history, and details on how their
adaptive payment solution works.
I hope you enjoy the interview.
Today on the podcast, excited to
welcome Eyal Karmi, CEO of Nicuda.
Eyal, welcome to the show.
Thank you.
Thank you for having me.
Nice to meet you also, Scott.
Yeah.
Yeah.
So far, uh, on the podcast, we've had
some folks that kind of pontificate
about consumer trends, and I thought
it would be fun to pivot a little
bit and jump into talking to a fellow
entrepreneur about what they're building.
And, uh, you and I connected on LinkedIn,
and you have started a business that, that
I think is gonna be interesting for us to
dig into, and I'm excited to learn more.
Before we do, maybe run us through
a little bit of your background.
I studied statistics, and I worked in
the Central Bank of Israel for, uh,
almost four years in, uh, cross-border
payments, actually And after that,
uh, I started my own company in,
uh, with two other co-founders
in AI and agriculture, actually.
So imaging, uh, fields, m- mainly in the
Midwest, and analyze the image and, and
give an agricultural report of the field.
Uh, I'm an old, uh, deep
learning, uh, era guy.
Uh- The OG AI.
Exactly.
Before it overloaded the term yet again.
Exactly.
Yeah, so, uh, done that
for, uh, almost seven years.
I, I also really love, uh, cryptography
and encryption and et cetera.
Uh, in 2020, discovered
Ethereum and crypto.
Uh, the world's computer narrative,
uh, back then, uh, blew my mind.
Decided to have to do
something in that industry.
So I left, uh, happily, uh, my company and
joined another startup named Solidus Labs.
Uh, Solidus Labs do market surveillance,
surveilling crypto trade, and trying
to find, uh, suspicious patterns,
uh, mainly for compliance reasons.
Uh, w- me-- I kept my, kept my
pulse on AI, and me and my other
co-founders relatively early on
agents, and we, we did some, uh…
We messed around a bit
with, uh, with LangChain.
Was a-- it still is an open
source, uh, community-driven,
uh, framework for agents.
Uh, it's got much bigger now.
And we did, uh, an agent that can order
a pizza for you, uh, through Slack.
Joe's Pizza, actually.
It fa- it failed miserably.
It failed to order a pizza,
or you, like, you tried to
commercialize it and it failed?
Uh-
It, uh, it wasn't intent to commercialize.
It was more of an experience.
It failed miserably because,
uh, we can talk about it, but
first, uh, they didn't let us in.
Like the bot detection kill you.
And even if you go after that
and try to do the payment, my,
my debit card got canceled.
So, so, so, uh, we realized there's an
inherent friction, and we extrapolated
that, uh, in the future, we're going to
delegate a lot of actions on the online
to, to those agents, and a lot of that
will be transactions and payments, and
it's not the-- it's not built for that.
Everybody assumes that the human comes and
they can press buy now and authenticate
and, and have the authorization.
So, uh, that's how we basically
dove into the rabbit hole.
That's kind of was the light bulb
moment, and we started Mekuda,
uh, almost a year ago, actually.
About 10 months ago.
Cool.
And you mentioned other co-founders.
How many, uh, how many
co-founders do you have there?
So we're a total of three
co-founders, so a CTO and a CPO.
Barak is the CTO, and Idan is the CPO.
Uh, we're a total of, uh, seven people,
four in New York and, uh, three in Israel
And you drew the short stick to be CEO.
Yeah, exactly.
When, uh…
That means once you have customers and
they complain, they're gonna call you.
Yeah.
It's a niche market.
There's not a lot there.
I think all of our customers
has a Slack channel with us.
We have, uh, a lot of them, but,
uh, it's, it's, it's an exciting
moment and it's, it's a good, it's
a good stage, so you can, uh, you
know, be there on the trenches and
iterate on the product and hear
feedback from, from the actual people.
So, uh, we like it.
Yeah.
So you were-- So agents,
let-let's kind of…
I definitely wanna learn more about
Nakuta, but before we dug-dig in, you
were, you were very early in agents, and
agents themselves have gone through a
big transformation where LangChain, you
know, v0 was, you know, an, uh, a g- an
LLM and some tools, and it was just kind
of, it was like giving a kindergartner
a, a machine gun or something.
Probably some bad analogies in
there, but, you know, it was, it
was way overpowered and, and the
GenAI wasn't really, you know…
Because…
And then, you know, if we fast-forward
to today, now the agents, they have
memory, they have retrieval-augmented
generation, so they're…
And then the tool sets they have available
are, are much more expansive too.
Has it been interesting to see that, that
kind of change over such a short time?
Yeah, for sure.
It's the, honestly, it's the
fastest and crazy thing that
I ever saw in, uh, my career.
Uh, we circulated that it will take
time, and we were relatively optimistic.
We said, like, in two
years, in three years.
Uh, apparently it took s-six
months to go from, people then
didn't even say agent, was like AI
automation was the, was the big word.
Mm-hmm.
And you have like nodes and decision,
and inside the node you might want
to have an LLM to do something
non-deterministic, but you will have
like this AI automation flow, and, uh,
six months after, it's like, uh, uh, you
know, blow-blowing up as, as you said.
So i-it is, it is a crazy ride.
It's a very…
It's the fastest moving, uh, industry,
uh, I ever saw in my life, including,
I don't know, crypto or all the
open source communities in some
part that are very fast usually, but
this is, uh, this is something else.
Yeah.
Parts of crypto at least are slow
because you've got the underlying,
you know, blockchain protocol piece.
Like if you're on Ethereum,
Ethereum doesn't change very
quickly because, you know, it, it…
You don't want a financial store or
financial contract or however you're
using Ethereum to, to change very much.
That doesn't, but basically
you can fork whatever was built
until today and add a little bit.
So- Yeah
there's, you know, thousands of indie
dev that every day are, you know, pushing
some new project on top of the stack.
So you're right.
So it's kind of a crazy
open source community.
And in general, you see that in
open source because you don't
need to invent 99% of the things.
You just fork it and you add a little bit.
So you don't need, you know, develop
the entire stack yourself, so.
But this is even crazier.
That's what I think.
Yeah, I agree.
Cool.
So then a year ago you started Nakuda,
and then, um, tell us a little bit
about, um, you know, are you still
doing the original idea, or did you
have to pivot a little bit in there?
We, we did some iterations.
I mean, so there was this light bulb
moment, and, you know, the, the first
natural thing is, uh, we said, "Oh,
we, we didn't-- we need to do like
an agentic checkout or, or, or some
machine endpoint or a route that
will make our agent life easier."
Uh- I think it's a good idea.
Uh, we didn't crack the go-to-market.
Like, uh, you know, start with
one merchant, and he ask you,
"What's, what are agents?"
A.
B, "Who's coming?
How much GMV are they bringing?"
Uh, and, and you, you're s- uh, you
know, it's, it's, it's a-- I think it's
a, it's, it's a big go-to-market, uh,
thing to do there to convince merchants
that already have a stack of things and
a payment processor and everything, uh,
already there, and n- not necessarily,
they're not, you know, technological.
Or not all of them.
Some of them are.
So for us, it was really kind of
we didn't know how to crack it.
Uh, it's a good idea, I think, to
be, like, a big company or, or maybe
the PSPs, you know, the, or the
gateways or, or the ones who are
already integrated will do that.
But, uh, for us, it was, uh, we didn't
know how to crack the, the go-to-market.
But, uh, uh, we, we,
we built a few agents.
We still have our own agents for,
for demos and, and, you know, we
want, we wanna be builders and, you
know, really relate to our clients.
And, and one of the things we
realized is that there was a lot
of mistakes, uh, of the intent.
So I said once, uh, we have like a, a
test store with, with shirts and, and swag
of Nokuda, and, uh, uh, I once said, "Go
get the shirt," or something like that.
I didn't mean buy it.
I mean, get the price.
I didn't say it exactly that,
and it went and bought it.
So I was very surprised.
Uh, it caught me off guard.
So we realized that we are not
as good as saying what we intend,
uh, what we mean as a species in
general, or write it, or say it.
And we realized there's gonna be
a lot of, uh, a lot of friction
there as well, or, or problems.
And, and then we kind of…
It, it drove us to focus
more on authorization.
So I think the payment industry
is very focused, justifiably
obviously, on authentication.
It's an umbrella term, but I'll make
sure that everyone understands, though.
Proving that Eyal is Eyal, and nobody
stole my card, and I'm authenticated.
But then everybody assumed that Eyal
pressed buy now, or Scott pressed buy now.
Like, once you prove that you are
you, they left you alone I think
once you delegate that action
to a software, authentication
is really important, obviously.
But then also even given that Scott
is Scott, what is Scott authorized?
What did you intend to do?
Uh, so you take this non-deterministic
event where I need to describe what I
want you to do, and you need to do a
deterministic action, which is go and buy
something or, or to do some other action,
and you need to interpret it, my intent.
So in our solution, we baked it in
that, uh, we're, we're more focused
on that part of the stack, so
downstream of authentication, which
the authorization part, basically
Got it.
Okay.
And then, um, so one of the things I
think's interesting about agentic payments
is there's a new, new party involved.
So, so you've got the merchant, you've
got the shopper, and then sometimes
you have a third party, which is
the platform the merchant is using.
Sometimes there's a fourth
party, which is the payment
provider the consumer is using.
Um, and then sometimes there's a
fifth party, which is the gateway.
But if we just kind of keep it high
level, we've got the software- I think
every transaction-
Yeah … there's, like,
40 vendors
or something like
that, right?
Yeah.
And there's some banks,
and, like, it gets crazy.
But, like, let's keep it super high level.
You've got the merchant and the shopper.
Well, now we have a new, um, third
party, which is the agent, right?
And so who, who are you-- who uses
your software in this equation?
Is it the shopper, the
agent, or the merchant?
Uh, the agent, the agent
company or developer.
Uh, w- the value proposition, we, we
help them basically store the payment
credentials of their users on the
agent themself and use it, uh, just
in time and onboard their users.
And so that's one layer, and of the SDK.
And the other layer-- And we use
network tokenization, I think from,
from payments to, to help them, uh,
uh, with, uh, with compliance, like
PCI DSS scope and, and other things.
And the other layer is the authorization.
So we record that intent to make sure
that, uh, when you're not present,
the, the agent, uh, uh, has, ha- let's
say, has, knows, knows its guardrails.
And so that-- these are the
two, two layers of the…
Okay.
So there's, there's a user
experience for the consumer, I guess.
And do you guys expose that, or is
it up to the agent builder to, to
kind of provide that interface too?
It's up to the agent builder, who can
customize, uh, almost everything that
I said, or, or you can obfuscate it.
We're focusing more on
browser-based agents, and it
could be an app, a consumer app.
Uh, it doesn't shows as a browser-based,
but sometimes on the back end they
have browser capabilities because the
internet is already built and working and
proliferated and, uh, is already there.
And, uh, there's-- sometimes there's
APIs and programmable way to do the
transactions, but, uh, or the payments,
but, uh, most of the internet do not,
'cause again, it was built for humans.
So, uh, for now we're focusing on, uh,
an agent that can manipulate the browser
and goes exactly like a human, the human
route, goes to the cart and the checkout.
So these are the, I would say,
our, our main focus for now.
Uh, it will, it will grow from there,
but that's where agentic payments
and commerce starts specifically.
So your customers are other developers,
so that's why they're on Slack kinda
chatting with you, so that's good.
And then, um, the, uh…
Yeah, so, so I'm gonna build a
new agent, and it's gonna be,
you know, a New York pizza search
engine or something like that.
And, uh, I, uh, y- you know, the-- I set
up my payments with the agent builder.
Uh, I'm the agent builder.
I work with you to add all this
payment capability so I can focus
on, you know, training my engine on
what the best pizza slice is, and
distribution, and all the other things
that go in- into starting a startup.
And then I have to give the user
the, a gateway to kinda collect
their payment information.
You store it in a tokenized way so that
I don't have any PCI compliance issues.
And then you're gonna go and take
care of the authentication kind of
piece you said, uh, authorization,
sorry, the authorization part.
So, you know, is this really kind
of what they're intending to do?
Did Scott really want 100 pizzas
or, you know, did he want 10 slices?
Or what…
Did he want 10 pizzas or 10 slices?
And then, um, then you also, you, you
then go and do the checkout, uh, as well.
You take those payment credentials
and somehow, you know, place them down
at the merchant level to take care of
everything so the agent builder can focus
on the best user experience possible.
Was that, do I-- What
did I get wrong in there?
Uh, you didn't get anything wrong.
Okay.
You got it perfectly right.
For now, we're focusing, so the
merchant side, though, thinks, it
doesn't know it's an agent for now.
Yeah.
It's relatively naive.
It literally go to the
iframe, to the checkout.
Looks like it's right in the
browser, so it just looks like
you're, it's not some centralized
IP or anything like that, yeah.
So that makes it less blockable, right?
Exactly.
I think it will change.
Uh, certainly will change.
Uh, we're in early days, you know.
I think- You know, how many tr- payments
or transactions have an agent's made?
Probably not a lot.
Maybe I did half of it just
from experiment, experiments.
So it's really early days.
I think first transaction happened
in the last few months, right?
So, uh, and, and it will proliferate
from here in a nonlinear pace.
But, uh, for now it's, it's, it's,
it's focused on that, but it will
change a lot, the entire stack in RD.
It's interesting.
So in, you know, I've, I've done…
I know you've read some of the
stuff I've done, and I always
talk about the big guys, right?
But what we don't know is there's
always these new challenger brands
that pop up, like, like ChatGPT itself
is one, and Perplexed is another one.
But everyone I talk to in this
space that's selling to builders of
agents, there, there's this tsunami
of interest in what they're doing.
There's so many people building agents.
Uh, you know, I think like over
half the last Y Combinator class
was all agentic, like 80% agentic.
Um, and what, you know, I know
you probably are under NDA,
but like what do, what do…
You know, do you see that, like
all these people reaching out
to you that are building agents?
And what are some of the…
Are there certain categories?
Like I imagine being in New York,
there's gonna be like fashion ones,
jewelry ones would be very popular.
I do hear a lot about food ones.
So, so is that, is that what you're
seeing or are you seeing stuff that's
even like, you know, more niche?
And what…
Give us a feel for what's going
on out in the, the primordial
soup of agent building.
So I think, uh, you really
named all the categories.
There's the general purpose
one, and, uh, ChatGPT and, and
Perplexity is a good example.
Uh, but so that's, that we
also serve, uh, as well.
Uh, there's also verticals.
So- Uh, I think i- in my mental model,
the, the good ones that I've, I've seen
or, or they are customers, they, they
have a different type of experience.
It's not like a skeuomorphic experience,
they just took, you know, the Amazon
or the e-commerce experience and put
it Control+C, Control+V into a chat,
uh, which the average consumer, I
guess, can digest, but they actually
pushed it into a persona base.
So it's like, uh, you said it, I think,
like a stylist, a fashionista that helps
you with your wardrobe, and by the way,
helps you procure, you know, the items.
It's not the store, it's, it's, it's
your stylist, and it's very stateful.
It knows your size, your taste,
what looks good on you, what…
Like, that thing is, is, is
really, I think, a big moat.
We have another customer who is a, he's
like a mechanic, helps you upgrade and
do modes to your car, uh, and it sources
the parts on the internet as well.
So I think-- And I'm not
saying that, uh, uh…
I think, I think that new experiences for
consumers would, would, some, one of them,
all of them will have breakthroughs 'cause
they're offering something, uh, uh, that
you can't, you know, buy on Amazon, which
is very, very reliable logistics, and you
have the Prime, so it's really hard to…
In one-click.
It's really, you know, the experience,
you, you need to, to win on that.
It's almost perfect.
So I think if you break through from
that then you can deliver to the consumer
something that, you know, rich people
has that in the end of the day, they
hire people to be their stylist, they
hire people to be their travel agent.
Uh, not just, uh, ordering on Expedia,
copy, uh, uh, a trip or, or a flight.
Something that will, you know, go end
to end, do a delightful itinerary trip
for me, uh, uh, for my vacation, and it
knows that I have kids and knows my seat
preferences and very, very stateful.
Stateful with memory, as
you said in the beginning.
Uh, I think these are the ones who
will break on the vertical side I
also don't think we're in eq- any
equilibrium on the general purpose.
Of course, you remember AltaVista and
Lycos and, uh, uh, which were, uh, A-
AOL, uh, huge companies, raised hundreds
of millions of dollars, and, uh, half
of your audience of this podcast doesn't
even know what I'm talking about.
So, uh, we're not at all in equilibrium.
There's so much room in search,
in consumer experiences.
May- maybe ChatGPT is a winner,
it will be here forever, but I, I
don't-- I, I won't crown anybody
else, I would put it like that.
Uh, that's my, my personal
humble opinion, uh, but, uh-
Yeah.
Whenever there's this much disruption,
disruptive creation going on, it's
hard to know what's gonna win.
It's impossible.
You just kind of have to-
Exactly … see where it goes.
Yeah.
Exactly.
Cool.
And I, I noticed you
guys raised some funding.
What, what can you tell us about that?
Yeah.
We raised a, a seed round from, uh, a
VC named Madrona, uh, from Seattle and,
and also Visa and American Express.
Nice.
That's, uh…
I think those, those would be two--
If I was doing a payments company,
those would be two folks that'd
be good to have on the cap table.
Um, Madrona, you're a, you're
a very, uh, humble person.
Madrona's kind of like a tier
one kind of consumer company.
Howard Schultz, uh, was one of
the, the founder of Starbucks.
That's his, his, uh…
I don't know if he's involved with it
day-to-day or anything, but I think
it's some of his capital in there, so.
That's a good one to be associated with.
For sure.
Yeah, we-
But now you have to do a
coffee ordering agent, so yeah.
I, I, I would do it.
Uh, I just-- Actually, I just spoke
with, uh, with someone, uh, on that idea.
Uh- Yeah … I, I-- Well, I use, I
use the Starbucks app, uh, a lot, so
coffee waits for you and you take it.
So why- Yeah, yeah
why should my, uh, you know,
assistant, uh, agent assistant should,
uh, know it or, or, you know, do
that, uh, orchestration for me and
utilize my points or my subscription
or, or something like that, uh-
Mm-hmm
…
for me.
Yeah, your stars, your Starbucks stars.
Exactly.
Cool.
That was, uh-- Were you
guys, um, out fundraising or
they just kind of found you?
How, how hard was the fundraising process?
We, we've, uh-- On our side, we found
there's a lot of, lot of VCs very
interested in this overall space, so.
Well, now.
Yeah.
But, uh, uh- You were early,
yes.
Yeah.
We-- First calls, we would need
to describe what is an agent,
which is, you know, unless…
You don't need to do it today, but,
uh, uh, you, you-- It was-- You need to
convince that there's underwriting, that
there's gonna be a big agent, uh, boom.
Like, that wasn't that obvious.
But, uh, I think what, uh,
maybe it's, uh, it's a good…
What's going to hap-happen for us is,
uh, uh, one, one of our co-founders,
Idan, he's good in writing.
We call him the chief writing officer,
and we started a blog in November
24, I think was the first one.
And we just wrote what we think there.
Honestly, everything that we think
or know, we usually write there.
Uh, and that really kind
of changed, uh, the tone.
I think people search for kind of
thought leadership and opinions
and what's coming, and makes sense.
And that what brought
us Visa, for example.
They read our blog.
So I think that really helped,
uh, organically for us, and
it's relatively small market.
It's, you know, uh, dozens of, of,
of agent developers who already
funded and not indie developers.
So everybody at this moment is
searching for a- Tell, tell them what
to think, what to do, how to act,
and, and some thought leadership.
So I think it's from all the things
that we've done, decks or, or a post
or whatever, the, the blog and, uh,
was the most, with the most ROI.
Very proud of that blog.
Yeah.
Yeah.
It's, uh…
That's how I found you, so it, it is good.
And I'm, uh- See?
You're preaching to the choir.
I'm, uh, uh, yeah.
You're letting all our
secrets out, uh, y'all.
You can't, can't tell
everyone how this works.
No.
We're building the open.
Yeah.
Yeah.
It's, uh, it's-- I'm o- I'm old school.
Building the open's a little scary
'cause we don't want, uh, people
to steal our ideas, but, you know.
I'm, I'm old school as well.
I don't know if you see my gr-gray,
uh, gray hair in here, but yeah.
Ideas are easy, execution's hard.
You just gotta out-execute everybody.
Exactly.
What are the incumbents?
So we've got, you know, the payment
industry's had, you know, many,
many generations of companies.
You've got, uh, I saw
PayPal announced some stuff.
Um, Stripe is kind of poking around here.
H-how do you position them?
Are you, are you kind of,
like, competing with them?
Are you, uh, tangential to
what they're, they're doing?
How do you, how do you answer that
question when, when folks ask you?
A small startup can compete with,
with these, uh, big incumbents.
But, uh, I think, uh- Stripe was
two guys in Ireland at some point, right?
Yeah, exactly.
Uh, but, uh, I think-- So first of
all, this wave nobody sleeps on.
Like, incumbents are not, uh, you
know, everybody f- even from, from
the market, you saw the announcement.
Obviously, you, you said it's Stripe
and PayPal, but even, you know,
Visa, MasterCard, uh, uh- Companies
that you wouldn't imagine that
are, you know, front-running the
market and leading it, uh, in terms
of, uh, again, of the size of it.
Like, let's be honest, how much
volume or GMV went through agent
relative to, to e-commerce?
It's very, very nascent.
So that was surprising.
So nobody sleeps on this because,
uh, right, they, they've been through
e-commerce and the mobile, and now
the third wave here, they, I think
they got the, the message there.
Very impressive, uh, products and
demos, especially from PayPal.
Uh, very, very impressed by
what, what, what they're doing.
I think everybody's throwing spaghetti
on the wall and see what sticks,
uh, you know, uh, at this stage.
Uh, I think in general, uh, maybe PayPal
a bit different 'cause they're already--
they have also a big consumer wallet, just
they're, they're placed a bit different.
It's, it's funny because
it's historical reasons.
It, it, it what, it what was maybe like,
uh, on their, uh, you know, uh, was a,
was a bit, uh, uh, historical reasons,
but, uh, now it's kind of coming back.
But the rest, uh, has already
integrated with the merchants, and
I think that's their natural place
and that's their main customer.
Yeah.
But in a way that limits
their TAM versus you.
By being neutral, you can
work with any checkout.
But if I'm- Exactly … Stripe or
PayPal, I'm now dealing only, I can
only let you shop at the universe of
checkouts I have, which, you know, is-
Yeah
…
even though those are big companies, it's
still not as big as like, you know, Visa
and Mastercard and Amex, or you get 100%.
Like you can, you know,
that's the Rosetta Stone.
Like as, as long as you have
that, then you can be anywhere.
Pe- people sometimes,
like it's very fragmented.
There are people who would only
pay with their PayPal and that's
it, so they need that option.
And there are people who really
like Stripe link on the merchant,
the, the card and file directory
side, so they want just that.
So I think there's a room to be
on the other side of that from,
from the shopper, from the buyer.
Uh, but buyers like optionality, so as
you exactly said it, it's fragmented.
There's room for more natural play.
You play with all of them, so that's one.
The second thing is, uh, uh,
it's very orthogonal to that.
There, there, there wasn't a need to
have, you know, on the shopper side,
you are the one with the agency.
You press buy now, you chose the
card, uh, and you went into a
checkout and, and you dealt with that.
So, uh- I, I think, I think it's
different, but, uh, you know, they're
all big companies and they're smart
and they can do, I guess, everything.
But, uh, we see it as a, as a
different direction versus a…
Yeah.
I think as a startup, your, your strategic
values, you're not balkanized, right?
So, like, PayPal's not gonna really
help someone check out on Stripe,
and Stripe's not gonna really
help someone check out on PayPal.
And so on and so forth, yeah.
Shopify is not gonna really be excited
for you to check out on BigCommerce or
Salesforce Commerce Cloud or, you know.
So, so everyone's kind of balkanized, and
it makes it very hard for them to have
a, the broadest possible user experience.
Shopify is closing their
garden, so it's only Shop Pay.
Uh, Amazon obviously is just Amazon Pay.
That's already closed.
Uh, so you have these gardens,
you have a bit more open, but then
you have, I know, Checkout.com,
Adyen, Stripe, like PayPal.
You have a lot.
I also think like, uh, as a small
startup, uh, sure, you can think
about, uh, what the, you know, the
chessboard of, of the world, but,
uh, just do a good product, uh- Our
customers has two main pro- problems.
We felt that ourself, which is, uh,
you know, just going to a website is
hard for a bot, uh, completing that
transaction is hard for a bot, and
getting higher approval rate is very hard
because you're a hosted agent, and you
look, you look very bad to the system.
It's, you look like a bot that
stole a credit card, basically.
So derive from that, reverse
that, that's the problem.
Solve that.
Uh, do the chessboard, uh,
in, I think, later stage.
You mentioned earlier you're, you're kind
of interested in crypto, and there's been,
since that big Visa, that, that kind of
week when we had kind of Visa, MasterCard,
and PayPal announce, there's been a lot
of interesting stable coin stuff going on.
Do you think, do you think agentic
shopping is gonna be an area
where stable coins are used or…?
It doesn't depends on the agentic,
it depends on the consumer.
So as a, I really believe in the ethos.
I love cryptography, so as a,
as an insider, I say it, uh,
but we need to be real here.
Uh, just two things here.
First, small startups create products,
not standard, so I would focus on that.
Uh, there, a lot of these issues
are already solved on a standard
perspective, not on product
perspective, so I'll focus on that.
The second thing is, uh, sure, I
would love to support stable coins.
I love it.
I, I personally love it, but the
general public in, we're US market
focus, uh, do not hold stable coins,
and definitely do not pay with them
on e-commerce or day-to-day thing
because, uh, same reason they don't
do it with debit, uh, in the US.
There are different markets
out there, I realize that.
Uh, so I don't want
real-time, uh, settlement.
I don't want the dollar, uh,
go out from my, uh, hosted or
unhosted wallet in real time.
I want consumer protection, and I
enjoy when it's, you know, settled,
at least for me, end of the month
or, or something very comfortable.
So, and I also enjoy my
points and my cashback.
So I think there's a lot of layers
here from the consumer side and the
experience that needs to be solved.
I see a future when all of that is
solved, don't get me wrong, but, uh,
when the rubber meets the road, uh, the,
like, doing a startup was already risky.
To, you, you need to underwrite that AI,
and they will do all our shopping, right?
And, and agents, it's a,
it's a big underwrite.
You wanna underwrite it prolificating to
the general public stable coins as well.
So two, two ifs, right?
It's, uh, it's non-linear.
So-
So if and when the general public in
the US will start to buy with and hold
stable coins or that will be obscured
for them, but it will be in the back
end of, of banks and, and then maybe
credit cards or schemes, you know,
uh, et cetera, uh, we'll support it.
It's, uh Pretty much agnostic to it.
Uh, it's actually, there's more
possibilities and it's funner to,
uh, for me it's, from a product
perspective, it's even, uh, like,
uh, it's more, uh, it's more fun.
Uh, but I think there's a lot of
layers of the description that
are missing on the consumer.
Are you talking cross-border payments
or, you know, if I'm a e-commerce
seller in Argentina or Zimbabwe
and I wanna hold dollars, uh, sure.
I 100% see the use case and, uh, sure.
But, uh, uh, uh, our, our ideal
customer who lives in the US and
have dollars and credit cards, we're
still a long way from that, I think.
Yeah, I think it's gonna be cross-border
trade where it really makes a lot of
sense to have that kind of, you know…
That's where the fees really bite you on.
They, they kinda get you, like,
three different ways on the, you
know, on, on the fee structure.
So the stable coins will
help with that a lot.
But, but even that, let's
also be real on the fees.
Sure, the, if you, if you from a wallet
to wallet, it's, it's very cheap or zero.
But if I live in Argentina, maybe I, I
need to, to do an offering to my bank.
I need to pay salaries, and not
everyone accepts, uh, uh, USDC or…
So I need to pay for, for, for,
for merchants, for, for, for stuff.
So the, the offering will, will,
you know, it's even more costly
than, than the interchange fees.
So until everything is very holistic
and very prolificate and very safe and
very consumer kind of, uh, uh, forward
with, with, you know, what, at least
on par with what we have with the,
uh, with credit cards and, and debts
programs, uh, I think it's a, it's
a future that is work in progress.
Got you.
Cool.
Um, I put out a bunch of
thought leadership around the
levels of agentic shopping.
I, I don't know if you had a chance to
look at that or not, but did you, if
you did, did you agree with kind of like
the, the five levels of agentic shopping?
Yeah.
I, I do.
Um, I, I have a…
My mental model is, uh, uh, ver-
very similar, even more reductive.
Like, what we see now is like
an AI-enhanced search, but the
checkout is still human checkout.
What we start to see these days, and
again, it's very early, is the agentic
part, which it's still an AI-enhanced
search, so you adjudicate what you want,
uh, using, uh, let's say Perplexity or
OpenAI or, or Menus or, or other tool.
But then also the orchestration,
the actual checkout is, is agentic.
Uh, and the third prong
is, is more to the future.
It's a bit more sci-fi, but once both
sides, buyer and seller, are both agentic,
there's a lot of things that I think,
uh, uh- are going to be unlocked that
are not possible in today's commerce.
So, uh, again, it's all downstream from
search, and you can double-click there.
But, uh, and I think all of them will
coexist, by the way, all of these stages.
It's not sequential or
something like that.
Yeah.
Well, there'll be a day where
there's autonomous vehicles, and then
there's taxis with humans driving,
and then there's humans driving.
Yeah.
Here, you go to Austin.
I think, uh, yesterday, uh, was the first,
uh, robotaxi, uh, release, so, so yeah.
Yeah, absolutely.
Same, same metaphor.
Um, I had Roy Rubin on, and he
was the founder of Magento, and
he was kind of thinking out…
Uh, like you, I'm a founder, so I'm kinda
trying to think six months to a year.
Um, and it's fun to think
further out, but we don't have
a lot of luxury time for that.
But, but he, he spends time thinking,
'cause he's a, doing a lot of VC
stuff, he's thinking 10 years out.
And he made the kind of statement
that, you know, we're probably not
gonna re- need websites in, by 2030.
Uh, and I think, you know, he
said there'll be a billion-dollar
direct-to-consumer brand built
that doesn't even have a website.
Um, you know, does that pencil for you?
And what, what happens to, you know…
Are these websites like, uh, do
they just kind of, they're there
and we don't really look at them?
Or are they just like a bunch of
machine language and the agents are
talking to each other through them?
Like, what do you think the world looks
like in, when we get to that level?
In general, I agree with Roy.
He's super smart, and I was on a
conversation with him when he said
the same thing, and, uh, I, I, I,
I directionally agree with him.
Um, that being said, uh, I still
want, like, a delightful experience,
and I think, uh, brands Would
not let so much obfuscation.
So I think there's
somewhat of an equilibrium.
It's true that, I think you said 2030.
Sure.
It's hard to tell, but I, I guess, but,
uh, the intent would look different.
Uh, it might look way more headless,
uh, uh, for sure, but, uh, uh, I'm still
searching for a delightful experience.
It m- maybe it's something like,
uh, Shopify exa- did an example
of a, HTML snippets where you
have parts of the UI or, or the
storefront within the chat, like in
an iframe or, or something like that.
And, and that might be the, the storefront
where the brand controls its destiny.
But I also think, I agree with, with
you and Ryan, it's an opportunity.
Like, if you don't have a strong brand
or do you, you're an API or something
more headless like that, uh, it's
your opportunity to, to shine, to…
You have these great channels
and, and you're, you're fit
for, for, for that channel.
So, uh, uh, maybe what was
your disadvantage is now
an advantage in that world.
So, uh, we'll see.
So directionally, totally agree.
I don't know if we don't have any
websites in 2030, uh, but, uh,
we'll, we'll-- It will be very
different and probably much less.
Yeah.
Having-- I've been at this e-commerce
thing for a while, and I think
people are gonna be s-surprised
how fast this one changes.
'Cause we, we thought e-commerce itself
would be faster, but it had a lot of
infrastructure to build out and things.
But it's all in place now and, and
we're gonna go up this curve way faster
than I think people are, are expecting.
Yeah.
So it took us 30 years, uh, for
e-commerce, uh, so this might take And
then you say five, 10, two, I don't know.
It will be faster, but, uh,
three time faster is still 10
years, still a long time, right?
So
It'll go very quickly.
Well, awesome.
Well, I appreciate you taking time out
of your busy day building a company.
Congrats on, uh, being a founder,
raising a great round from a
great investor and, and, uh,
getting those strategics on board.
You're, you're making this look easy
for all the other founders out there.
You're making us look bad, buddy.
Oh, you see…
Again, you see my gray hair, it's, uh-
It's not, it's
not easy.
Yeah.
Um, congrats and, you know,
we'd love to check in.
If you have any news, let us
know for our, our Substack.
And then, um, you know, if you
send out any releases or anything.
Um, and maybe like in six or 12
months we'll have you back on and
get kind of an update because you've
got your finger on the pulse of
a piece that, that I don't see.
I'm talking more to retailers and brands.
Um, so it's interesting to see what all
the, the startups are building out there
and, and you have a front row seat.
Cool.
Uh, thank you very much for having
me and we'll definitely, uh,
check in, uh, in a few months.
That will be very fun.
Yeah.
And if, folks, you mentioned,
um, you know, that you guys put
some thought leadership out.
Where, where should they go for that?
We have our Substack.
I can send you the link if you
have a, a way to put it, but
it's on our Substack, Nucuda.
And on our website there's
also a link for it, nucuda.ai.
And we'll put it in the show notes.
You all have a great day, and thanks
for coming on the RetailGenic Podcast.
Thank you very much.
Follow, rate, and share to stay up
to date with RetailGenic, your guide
to the future of AI shopping agents.