Retailgentic | Consumer Behavior & Retail Trend

In this episode of the Retailgentic podcast, we sit down with Karl Haller, Partner at IBM Consulting and leader of their Consumer Industry Center of Competency. Karl brings deep insight into the retail, CPG, and fashion sectors, sharing how AI is reshaping enterprise operations, from store-level automation to the rise of agentic shopping and AI-native brands.

We explore:
  • How IBM uses AI internally to save billions
  • The future of AI-native startups with 5x leaner orgs
  • What retailers need to know about agent-based shopping
  • How personalization and dynamic pricing could transform the consumer experience

📍 Timestamps:
00:00 – Intro & Karl’s background
01:10 – IBM's early AI work and internal cost savings
05:00 – AI’s shift from hype to workflow transformation
07:15 – Retail use cases: contact centers, operations, and automation
10:55 – The agentic shopping shift: consumers vs. retailers
16:00 – Why gift-giving may tip agentic shopping into the mainstream
18:30 – ChatGPT’s hosted checkout and new marketplace dynamics
22:00 – The rise of AI-native brands and single-person unicorns
30:00 – Store-level orchestration and event-triggered workflows
36:00 – Personalization in stores—like having Anna Wintour’s whisperer
39:00 – Dynamic pricing and loyalty-based personalization
44:00 – Final thoughts: the slow shift until the tipping point hits

đź§  Want to stay ahead in AI commerce? Subscribe and follow along: 
đź“° Subscribe to the free Substack: retailgentic.ai
📺 Watch episodes on YouTube: youtube.com/@retailgentic
🎧 Listen wherever you get your podcasts: Spotify, Apple, etc.

Creators and Guests

Host
Scot Wingo
CEO and Co-founder, ReFiBuy.ai
KH
Guest
Karl Haller

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).

The way I imagine it, I go back
to, uh, um, Devil Wears Prada, and

there was the scene at the, at the
Met Ball where, um, the Anna Wintour

character, um, you know, someone would
come and whisper in her ear who the

person was she was about to talk to.

And that's, that's essentially what
we're talking about putting in place.

You know, you're gonna then…

Your best sales associates in a high,
in a premium luxury space, they, they do

that already for their best customers.

But what it'll give you the opportunity
to do is ha- get more personalization, you

know, when you're not, uh, you know, uh,
you know, whatever, a billionaire spending

$50,000 at a single s- single store visit.

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 checkout?

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.

This week, I'm excited to
welcome Carl Haller to the pod.

Carl is a partner at IBM Consulting
and works in their Consumer Center

of Competency, where he focuses on
retail, CPG, and fashion verticals.

Carl's been working on AI at IBM for
over nine years, and before that, had

15 years of experience in the industry.

In his current role, Carl is frequently
meeting with the C-suite of top retail

and CPG companies to discuss AI trends
and strategies, so he's a great person to

check in on what he's seeing and what is
top of mind with retail executives today.

In this discussion, Carl shares that his
clients are moving from AI experimentation

to workflow transformation.

Carl also believes we are
going to see the rise of a new

category called AI native brands.

Everyone remembers digitally native brands
will get ready for AI native brands.

Well, not only are they gonna be digitally
native, but because they're also AI

native, they're gonna be hyper-efficient.

So they're gonna be able to
hit 100 million in revenues or

more with 30 people or less.

So that will give them some interesting
advantages we'll talk about Finally,

we go over the agentic shopping
inflecti- inflection point, and

where will AI take store retail,
and the advent of personalized

pricing.

I hope you enjoy this conversation
with Carl as much as I do.

Carl, welcome to the
Retail Agentic Podcast.

Excited to have you on.

Thank you.

I'm excited to be here.

Uh, you and I have gotten to know each
other well on LinkedIn and have a couple--

had a couple convos, and I've, I've really
enjoyed your content on, on LinkedIn

that you've been, you've been sharing,
and we've had a couple back and forths.

So I thought it would be good to get
you on, and you have a broad spectrum,

um, through your position there at IBM
Consulting, uh, of what's going on.

But let, let's start
back at the beginning.

Let-- Give us a little biography of, uh,
where you're from and, uh, where are you

right now, and tell us what you do at IBM.

Um, so I am a, uh, a, I was originally
a consultant back in the Internet 1.0

era.

Um, I was then working, uh, at an ad
agency and at a series of retailers and

fashion companies, uh, throughout most
of the 2000s and probably half the 2010s.

And then I've been, uh, at IBM Consulting
since 2016, and it's very interesting

because when I joined in late 2016 and
into 2017 and 2018, we were talking

to all of our clients about AI.

And we were the only ones talking
to our clients about AI then.

Um, and then, you know, I'll say,
gosh, around the time COVID hit, that--

those conversations all died out.

And then two and a half years later,
you know, 2022, 2023, it's now pretty

much all we're talking about again.

Um, and it's, it's-- I think we're
actually finally getting to the point

where AI is not just a thing, but
it's becoming a piece of everything.

And, you know, that took, that
took digital and the internet a

decade or so, and it's taking,
you know, two, three years for

AI to progress through that pace.

So what I do now, um, I lead our
consumer industry center of competency.

Um, we're a global practice full of people
like myself who came out of industry and,

um, we work with clients in the retail
space, the CPG space, and the fashion

space, um, to transform different parts of
their business with, uh, data information

and next gen technology and capabilities.

Got it.

And when, when people think IBM,
they obviously think of the hardware

company, but this is, this is--
You're not out there, um, selling

ThinkPads, so I, I take it.

We are not.

No.

We're, uh, we, we have a, we have
a, you know, very strong partnership

with the tech side of IBM.

Um, and in fact, you know, the tech
side of IBM and the work IBM corporation

has done transforming our business with
AI is probably one of the, the biggest

selling points we have right now.

Uh, we went public at our investor,
uh, meeting back in January saying that

we were driving about three and a half
billion dollars of annualized run rate

savings largely due to AI and automation.

And so just for perspective, we're
about a $60 billion company, so that's,

you know, 600, 650 basis points of
SG&A savings, which, you know, that,

that, that opens the eyes of any COO,
CFO, CEO that I'm out talking to.

Interesting.

So you're, you're, you're dogfooding this
capability internally to drive automation?

Yes.

We prefer to say drinking our
own champagne, but, you know,

if you go dog food, that's fine.

I'm a software guy, so I, yeah, I
recognize that has a negative connotation.

Um, cool.

So then as a, so as a center of
competency, you're-- are you doing

mostly thought leadership kind of
things and, and helping, um, uh…

Or, or that just, that's the
definition of your vertical?

It's just the definition.

Um, what we have is we have the
majority of our consulting practices

live in the, the what we call the GOs.

So the US has a practice, Canada has
a practice, Europe, all the countries

there, uh, you know, Asia, et cetera.

And then we have small global practices,
um, that get deployed in where there

are the specific level of industry depth
required, you know, industry functional or

industry technical depth that is required
that might not exist in the market.

So even though I'm based in New York,
um, a lot of the work I do ends up being

with some of our smaller markets where
they may not have as strong of either

a retail practice or a CPG practice.

Um, and so we, you know, we're a
group that I'll say it used to be

we always had to have a bag packed.

Um, now we have to be ready
for a 2:00 AM or 3:00 AM client

presentation, you know, on, on Teams.

Yeah.

Okay.

Cool.

And then because you're living in
retail CPG and fashion, um, and, and

you talked about this wave, like when
you guys had Watson, you had a lot

to talk about, and then, then kind of
it went down and now it's come back

with the evolution of the ChatGPT kind
of, you know, new, the GPT-based AIs.

W- as you go out and talk to
clients in those, those buckets,

what are some of the things that
are, are top of mind with people?

It, it really depends on client by client.

There are some clients who are thinking
about what I would call, you know,

everyday productivity tools, thinking
about how they upskill their, you know,

ten thousand, fifty thousand, a hundred
thousand, half a million associates.

Um, there are some people who are very
focused on specific areas of the business,

so maybe it's they wanna see what they
can do in finance, or they wanna see what

they could do in marketing or in store
operations, um, or in, you know, the, the

whole product and merchandising space.

Um, and then some clients, and I'll say
this is a, this is a big shift we've

seen in the last six months, um, clients
moving from, "Hey, I'm gonna create an AI

lab or an AI garage, and let's get, you
know, 50 use cases running, and let's, you

know, spin up a bunch of POCs and MVPs."

And what they've found is not
all of those are going anywhere.

So we, we've seen a real shift
toward, um, you know, 50 POCs

to let's focus on transforming
four key workflows that matter.

So maybe if you're a CPG company, maybe
that's my idea to launch my, um, my sales,

my marketing, and my data and analytics.

And you'll, you'll take a very strategic
consultative approach, um, to how

those might-- how those would change
and build out a more integrated set of

capabilities to use, um, I'll call it,
you know, old-fashioned AI in the AI ML

space, generative AI and agentic AI, um,
in the different processes that make up

kinda one of those broader workflows.

Uh, and that, that's been, that's been
a really noticeable shift among kind of

the more leading edge of our clients.

Got it.

So they used to be kind of throw it
against the wall, let's kind of do

some experimentation that wasn't wildly
successful, but now they've identified

some deeper, deeper challenges and
workflows that they can go and,

and create a lot of automation on.

Yeah.

And they wanna, they, you know, they
wanna drive, you know, top line growth,

they wanna drive margin, and they'd
love to capture some of that, uh,

similar type of, you know, 300 to 600
basis point savings in, uh, G&A cost.

When you're doing that G&A cost, is it--
Are you replacing people or augmenting

people or a little bit of both?

You know, it's both.

I think there, it, it's…

You really have to think about
what the job function is.

So right now, one of the clients
I'm working with, we're looking

at contact center, and that's, you
know, that's a widely known use case.

A lot of people are talking about contact
center, and a lot of people have moved to

put virtual agents in the contact center.

And what that ultimately may
mean is it ultimately may mean

you need fewer frontline agents.

Um, but if you're, if you're in
a growth mode, what that means is

you're not having to go out and
hire as many frontline agents for

a job that churns quite heavily.

Um, and you might actually be changing
the profile of what that frontline agent

is doing because that frontline agent now
is not handling, you know, in a retail

sense, I don't know, 30% of a lot of
calls are WISMO, you know, just a where

is my order or how, why did I get charged,
uh, for, for that So as those start to

get off-boarded, you really need a more
consultative agent, um, that is, that is

working with a customer, and maybe that's
resolving a more complex service issue.

Or maybe it's shifting those contact,
those people in the contact center to

serve customers in a place where that
brand doesn't have physical presence.

You know, maybe they don't have their
own stores, maybe they don't have

their own, um, you know, wholesale
accounts that they deal with.

And, you know, it's, it's a way that
you can extend and in- expand and

augment what the contact center does.

So I think, you know, yes, there's, there
are, there are cost savings that people

can put in place by reducing headcount.

However, what we would say is that
might actually be taking a, you know,

a limited, a limited scope or a limited
view of what the real opportunity is.

Cool.

And then, uh, so you've read a bunch of
the content we have on Retail Agentic.

Any, um, uh, is this kind of stuff coming
up with your clients or, or still a

little too early to be on their radar?

The, the idea that the traffic is
shifting, the organic traffic's gonna

be d- decreasing, uh, the GenAI answer
engines are increasingly capturing

consumer share, and therefore they're
starting to pull the transaction

up into the, the GenAI layer.

Again, it really depends on which
company I'm talking with and, and who.

Um, you know, I w- I was on
with a client this morning.

Um, he's a senior executive in a
specialty retail, uh, company, and he's

focused on, I'd say the intersection
of stores and customer engagement.

Um, and he was not fully up to
speed with what's happening on the

kind of agentic retail, agentic
commerce, AI-enhanced shopping side.

Um, and so we were just talking
about that, and, you know, his eyes

are, were gigantic, and he quickly
sees what the possibilities are.

Um, but that's not his focus area.

Now, if I was talking to the person
that is, you know, in charge of SEO

at that same company, they're probably
much more up to speed on that.

I do think there is still a gap
in what people are doing about it.

Um, I think there's a lot of
experimentation and a lot of, like,

the ad dollars are probably the, the
first things that start to shift.

Um- Outside of a handful of companies,
I'm hard-pressed to think of ones

that are either developing their own
agents to shop or embracing, you know,

third-party agents to come in and shop,
and who are starting to think about, you

know, whether it's my, my, my site, uh,
design and content and the way that is

displayed for agents, um, or the, the
nature of the content itself and whether

it is made to be consumed by agents.

You know, there, there's, there
are a lot of, there are a lot

of, uh, complex discussions and
for most it's still very early.

Do you guys, um, have a point of view
on where this is going or you're--

that's not really how you work?

You don't kind of put out thought
leadership pieces, you, you more

work one-on-one with clients?

So, uh, uh, we do, we do a bit of each.

So, um, we have, uh, I think what is
now considered the top or second, number

one or number two ranked, um, thought
leadership organization, uh, which is

the IBM Institute for Business Value.

And I'll say the thing that-- A-
and so we put out a lot of, um, a

lot of thought leadership pieces.

Lately, we've probably had a new
piece every two or three weeks on

how AI is going to transform X.

The, it could be, it could be your, um,
digital product development process, it

could be marketing content creation, it
could be finance, it could be, you know,

automation, it could be how, um, you
know, edge compute workloads are deployed.

You know, it, it really ranges.

Um, we've been running a, uh, CEO study
now for 25 years, and one of the things

that's, that's interesting in that is
some of what I was talking to about

CEOs are, are, uh, seeing that the, the
challenge of enterprise, that, that,

that enterprise AI is b- is challenging.

We actually had an interesting
data point on that and, and it's

not specifically on agentic, so
I'll come back to that in a second.

Um, a year ago when we asked CEOs
where they were with AI, two-thirds

of them thought they would be
beyond the pilot phase this year.

Then when we went back and interviewed
them this year, and this is, you

know, 2,000 or 3,000 CEOs that we
talked to, so it's not, you know,

three people and their brother.

Um- 60% of them are still in
the pilot phase this year.

You know, so it's-- we- we're
seeing that the pace of deploying

AI at scale is difficult for a lot
of our larger clients, you know.

And we're often dealing with certainly
clients largely in the hundreds of

millions up to the billions, tens of
billions, and even hundreds of billions.

Um, and it's not really the tech,
it's much more the organizational--

the organizations they have in place,
the embedded processes and having to

reinvent how a process works, um, and,
um, just the, the ways of working that,

that exist that tend to slow things
down and make something that, you know,

could take six weeks take 12 months.

Yeah.

And all this stuff is based on data, and
just getting the data together is a huge

front lift on, on all these projects.

Yeah.

And, and it's often these
different pools and, yeah.

Yeah.

And I think that's probably the one, you
know, the, the, the one building block

that I'll call, say, the smart IBMers
tell me that you need to have in place, is

you, you have to have the data in place.

Um, now, and we've got a
bunch of tools to do that.

But if you come back to the, like, the,
the, the, the, that intersection of

retail and agentic commerce or agentic
shopping, um, I think we're going to

be in a situation not dissimilar to
the late nine-- mid-late '90s when I

was kind of first starting out, um,
where consumers are getting ahead of

the big companies that are out there.

Um, I was, I was going back and looking
at some of this data, you know, and

Forrester had a report that from '96
to '99, e-commerce grew from about a

billion and a half to about $20 billion.

Um, you know, so whatever, 130-ish
percent annual growth rate over

those, over those three years.

Um, we're, you know, this year
we're probably starting from

an even-- from a smaller base.

I don't know if we're gonna hit a billion
and a half in agentic commerce, and I

don't even know exactly what that means.

You know, is it the search volume?

Is it, do you have to
actually click and buy?

But, you know, the numbers I've seen, um,
in ter- um Our, you know, the growth is,

you know, either, either triple digits
or quadruple digits, you know, right now

year over year off of a very small base.

But I think we're gonna hit, with
all the releases out, that are out

just in the last out and coming
out, I think we're gonna hit a big

tipping point, uh, this, this holiday.

I think gift-giving is ripe
for letting an agent go and do

some of the shopping for you.

You know, if you know-- Like, like if
I knew what your three of your hobbies

were, I could go search for, you know,
something that might align with that,

um, and come back with, with some
great options, and maybe I just go

through and click to buy right on that.

And that, that agent is just
out there doing that legwork.

Um, and I think we'll, we'll get a lot
of the, you know, a lot of the early

AI adopters and then probably the first
wave of fast followers to start…

We're not gonna turn all of our
shopping into agentic shopping.

Some of us like shopping.

Um, many people like, you
know, we still go to stores.

Stores are still, you know, they're,
they're, you know, whatever, 80-ish

percent of the sales and roughly that
amount of the shopping involvement.

But it's gonna-- We will see agentic or
a- I, I've been calling it AI-enhanced

shopping, um, to, to kind of include
the search in with the agentic and

I, I think we're gonna see, you know,
a, a significant tipping point this

year, and then probably another, the
bigger tipping point, uh, next holiday.

And then I think in 2027,
it's like, it's mainstream.

Yeah, I agree.

Yeah, I think that's a good timeline.

The big, the big outlier for me is
we've got ChatGPT-5 that everyone…

You know, there's always people that,
that have played with it and are amazed,

so they always tease us with that.

Uh, and that's, the-- as of th- this
recording, they're saying that's coming

in August, so that'll be a big one.

And then, um, you know, I, I think…

So it leaked that ChatGPT is gonna have
a checkout, and it's gonna be interesting

to see there where, you know, these--
In my, in my vernacular, there, there's

a world where they actually become
marketplaces where the transaction

happens on the answer engine, which is,
it, it's an interesting user experience.

It's a better consumer experience
because it all is just right there

and encapsulated and, and less chance
for failure with an agen- agent

having to go do the checkout for you.

Um, but it creates complexity on the
retailer side, because now you've got

this pool of inventory you've either
committed or sent in some, some digital

fashion to a number of answer engines
that are gonna be checking out for you.

So, um, it's gonna be…

Uh, for me, it's a replay of
marketplaces all over again.

It's kind of, uh, funny to
see that, that play out again.

I think I, I saw your analogy on that.

I can't remember when you put it
out, but I think that is a very apt

analogy for how some of this plays out.

And, and I agree with you.

It, it offers a lot to the consumer.

Um, one of the things that, and you
see this with marketplaces as well,

where the rub comes is on returns.

You know, on the, the-- It's the credit
card payment and who's actually getting

the payment, who's capturing the, the
information from the customer, who's

your transaction actually with, um,
and then what happens if something goes

wrong, if a delivery is late, um, or
if, if you have to return something.

That's, that's where the…

You know, I would, I would even say
a lot of, a lot of retailers still

have challenges on that with the
marketplaces they operate today.

Yeah.

Maybe there'll be some ChatGPT trucks
that come and pick it up or with robots.

Humanoid robots will just
come grab that stuff.

We'll see, we'll see how they
do in the physical goods space.

Um, I think a lot of people
get very surprised when they go

from the digital world into the
world where you've got stuff.

Yeah.

Yeah, it gets complicated really fast.

And that, uh, whatever the th-
whatever the pendant thing is

that they're launching, uh, I
think will be an interesting…

It will be interesting to see you
know, how quickly we're ready for

an adoption, how quickly we're
ready to adopt a, a new device.

Yeah, it's gonna be, uh, I've had
a Alexa Plus for a while, and it's

actually surprisingly good with
the, the new, the LLM, the gen

AI-based, um, conversational thing.

Uh, it's better at services,
which is interesting, than…

The e-commerce is about the same, but
the services has had a significant

upgrade where you can have it
book reservations for you and call

an Uber and that kind of stuff.

That's pretty neat.

You mu- you must not have teen,
teen children in the house.

Uh, I do, yes, but they don't have--
They don't know we have this show.

Oh, okay.

Well, yeah, I was gonna
say, they w- they will now.

Yeah.

Thankfully, they don't, they don't really
watch this pod, so we're, we're safe.

Um, when-- do you think, uh, for the
clients that talk to you about this, do

you have any recommendations for them?

Are you, are you kind of saying a
wait and see, or are you kind of,

you fall into the you should lean
into this, or, or client by cl-client

kind of depends on where they are?

I really think it is
somewhat client by client.

Um, you know, I was talking with a, uh,
grocery retailer, uh, a few weeks back,

and, uh, the, the woman I was meeting
with there, you know, was all up to speed

on this, super excited about it, but
she's like, "Yeah, but it's grocery."

Like it's, it's, it might not be the
thing that we're gonna embrace right

away, especially given that we're still
seeing, you know, frankly, we-- grocery is

where we started to see a little bit of a
shift back post-COVID into stores again.

Um, and I think everyone thought that
we were gonna kind of hold and expand on

the, uh, you know, keep growing on the
grocery e-com, and we've seen, frankly,

a bit of a, bit of a pullback in that.

Um, so you know, that might not be ready.

I think the, the-- a business that is
dealing with, um, a consumer base that

are full of early adopters, you know,
then I would think you might lean into it.

I also think it depends on what
your legacy is, what your legacy

stack is, and how easy it is, it's
gonna, it's gonna, um, how easy

it's gonna be to build that up.

I-- One of the things that, uh, we have
not really seen yet, but I anticipate we

will, especially as consumers adopt more
quickly than large, uh, large retailers

and brands, I think we will see a surge
of, um, what I'm, what I'm calling AI NBs.

So AI native brands that are rethinking
every aspect of their business AI first.

So they're, they're AI first from the
shopping side and the, you know, the

marketing, sales, and service side,
like the digital native brands were.

But they're also AI first in product
development and in merchandising

and in sourcing and supply chain,
and you know, they might have a

finance team of two at $100 million.

You know, they-- 'cause they, they just,
they're gonna, they're gonna take a lot

of those back office functions where you
need to get dozens of people at that,

at, at that size and be able to do it
in a more, um, you know, automated,

a automated agentic workflow manner.

You know, for whether that's finance or HR
or indirect procurement or legal services.

And I think that's gonna give
them a real cost advantage in

the way the DNVBs never had.

Yeah, the, the-- So the most famous
DNVBs were, like, Bonobos and Casper

and, and, and these kind of folks.

And, you know, they all-- they also
had this pattern where they all shot

to 100 million, maybe 200 million, but
then they flattened out because the,

the direct-to-consumer piece ran out
of gas when they ran out of dollars

they could spend at Google and Meta.

So something has to happen on
the distribution side, too, even

if you have a cost advantage.

You know, maybe you get to 100
million more efficiently, um, and,

you know, and you have better pricing
because of that or something, a

little bit more of an edge there.

But at some point, there's this
distribution problem in our world where

we just, you know, the consumers are
locked up in, in a couple of places.

So, uh, I'm, I'm cautiously optimistic
that gen AI creates a wedge.

We've got a new set of consumers and, and
maybe that it creates more competition,

which is always good, and maybe we'll see
some of those not only use the advantage

from the, the AI cost savings, um, but
also they can, they can punch into the

consumer and get to a bigger scale.

And it, and it'll be interesting to
see what happens to the companies that

grew in the, you know, all the, the
companies that were born in the '90s

and early 2000s and then grew since
then that, you know, are now kind

of legacy from a tech perspective.

You know, the, the mag- Most of
the Magnificent Seven, many of the

Magnificent Seven, you know, were
founded between '95 and 2005, um,

and, and kind of rode that first wave.

And whether that's on the, you know, the
tech side or whether that is on the, um,

on the actual retail and product side,
I-- it, it's, you know, anytime the

consumer is gonna get ahead of where the
industry is, it creates an opportunity

to, um, you know, disrupt, uh, what's
out there and start to go stake a claim

for share in a, in a different way.

Yeah.

Let, let's, let's peel the onion on
that AI-native brand thought experiment.

Is it, um…

In my startup world, there's all this
talk of about a single entrepreneur

or single-person unicorn, right?

Which is like this…

And I think Sam Altman quoted that,
like, in the next five years, we'll

have one person running a unicorn, which
would be a company with a billion-dollar

valuation, which implies $100 million
kind of in a, in a software world.

But even, you know, if we, if we take
that to a, a, a brand, you know, is

it Is it your, each of your functional
kind of things, your, your, your

merchandising, your, your product
development, you're managing a team of

agents and, uh, you know, or is it…

W- what do you think that looks like?

I, uh, so, uh, it, it's interesting.

I was with, um, a couple weeks ago, I was
with a handful of, you know, I'll call

other, you know, senior execs working
in and around the fashion apparel space.

Um, and where we got to was it's
somewhere between 20 and 30 people

to have a business that's doing,
you know, $100 to $200 million.

Um, and that probably needed to be
75 to 100 people, um, previously.

It's, it's especially if
you're making your own goods.

Um, you know, and, and not, I
don't, and I don't actually mean

producing with factories, I mean
sor- you know, doing the design, the

technical design, and sourcing that.

Um, so I do think there is a,
like, that- that's where that

cost advantage comes into play.

So kind of a three to five X multiplier.

Yeah.

And it, and it won't be-- You won't
get all of that because the people

will need to know what they're
doing You know, you have to…

It's like, it's like if all the agents
were really smart interns who, you

know, who kind of knew their function,
but you still had to manage them,

and they occasionally make mistakes.

Um, so you, you kind of-- I always say,
like, you need, you need to know enough

about what the output is, whether that's,
you know, whether that's a, um, a prompt

output or whether that's an agentic
output to know when to call BS on the BS.

So, so these 20 to 30 people are kind
of o- orchestration and oversight,

and maybe they're-- And then, you
know, GenAI is not good at creative

things, so they're, they're probably
on the creative side as well.

So the product idea, the
design, and that kind of thing.

And then they're, they're super
augmented by the more junior functions

that are, are agents doing that on
their behalf, ex-executing on the

design and the plan and the build
and the merchandising and all that.

Yeah.

And what you might have, you know, you,
you'd have someone, like in, if you just

take in product design, you'd probably
still have someone who is in charge of,

um, you know, delivering on the vision
of the brand with actual products.

And maybe that's one or two people,
but y- and you would have agents that

are out doing actual design work,
and then you're, you're reviewing

and curating and making changes.

You might have agents that are
creating the, you know, what are

called the tech packs, the technical
specifications around that.

You might have agents looking for sourcing
and production and then working, you

know, doing, uh, the negotiation back
and forth on production and cost of

goods, agents overseeing and monitoring
that production, um, and then monitoring

the inbound, uh, sourcing of that.

You know, reviewing, you could have
an agent reviewing the, um, you know,

reviewing digital samples to make sure
everything is up to spec, uh, and in

alignment with what you were setting out.

You know, and then pulling that in,
you probably have, you know, again,

someone, someone on top of most of those
functions that knows what they're doing.

But the, the teams of people would
be replaced with teams of agents.

Yeah.

I, I have a small investing fund,
and sometimes people in their

deck, they'll include an org
chart that has agents in it now.

It says there's, like, the human
side and then agent side or,

or they'll be interspersed with
different colors and things.

It's pretty, pretty strange world
that, you know, we're, we're starting

to think about these digital workers
and, and they're on the org chart,

and a lot of them have names and
functions, and because they need a

little bit of a job description to stay
in their lane, and that helps with the

prompting and, and all this other stuff.

It's, it's, it's a really-- If, if
a-- And I find when you talk to people

about this that haven't thought about
it, it's kind of, they, it's super

mind-blowing that this is happening.

I, I was with a client in Canada, uh,
back in June, and we were talking about…

This was, like, to their senior
executive team, and we were like,

you know- Some of you will be
doing reviews for agents next year.

You know, you'll be sit- you'll be,
you know, not sitting down across the

table and doing the annual review,
but you have to think about what

are your HR policies, you know, how
do you do a review for an agent?

And is that a continuous thing?

Is it a pulse thing?

And it starts-- then you rethink,
well, gosh, how should we be

doing it for people as well?

And you, you-- it starts to, uh,
as we were talking about kind of in

the contact center, it changes the
nature of how the work is done in a

more, uh, significant way than just,
okay, now I have an agent doing this

How about, um…

So that, that, that's,
that's super interesting.

The, uh, IBM's known for-- Whenever I
go to the NRF Big Show, you guys have

a huge booth there, and, you know, you
obviously do a lot of the-- That may

be more on the hardware side, but, um,
how about at the store level stuff?

Have you had conversations with,
with folks that have stores

where this is gonna be applied?

'Cause it seem, it seems like it'd
be harder on one level because--

But, you know, on other levels-

I, I'll say it's harder mostly in
that there's more complexity involved.

Because, you know, let's
say you've got 1,000 stores.

Now you, you have to think
about the workflows that are

happening 1,000 times over.

Like, in HR, you don't usually
have 1,000 HR departments.

But when you have 1,000 stores, you have
to, you have to think about not just

the overall workflow that's happening,
but the tasks that are happening in

each store and how you are observing
things, you know, activity that's

happening in the store, and then how
you're triggering, um, how you're

triggering either tasks or jobs that
off of those things that are happening.

Um, and maybe that's coming through
computer vision or maybe that's

coming from IoT or RFID devices.

You know, it could be information from
seeing how many cars are coming into

the parking lot or how many people are
backed up at the checkout, um, or a signal

that, you know, uh, a, um, you know, a
freezer is running three degrees high.

And, you know, but then that can
actually start to trigger a workflow,

um, that some of that workflow might be
agentic and some of it might be human.

Um, but the, the opportunity for, um,
improvements in effectiveness is gigantic

because that's probably, in terms of
staff, most retailers probably run, you

know, 90% to 95% of their overall employee
count, you know, works in a store or maybe

in a DC, um, not in the corporate office.

So to be able to make them more
effective, you're gonna deliver better

service, and frankly, you're gonna give
customers a reason to shop the store.

Um, I don't actually think
this is about cost savings.

I think this is about being able to do all
the things you need to do in a store to

manage the product, manage the operations,
you know, actually serve customers,

help them when they come in, um, but do
that in a way that is, is more efficient

and effective based on events that are
happening rather than based on, you

know, the routines and daily schedules.

That, that take place today.

But that is a, that is a big area
that we're working on right now.

Yeah, 'cause anything that has a human in
the process will have a human that's sick

that day, and they don't get the alert
the freezer's about to break or something,

or yeah, and then, then things go awry.

When you're dealing with in it in like an
event orchestration or an event automation

way, if someone doesn't pick up that,
that job or that process, it'll bounce.

Like, so you, you can, you can take…

The, the thing we talk to a lot of
retailers about is being able to take

their, um, their SOP, their standard
operating procedures, and how can

you actually put those into code.

Um, now the end job might
still be done by a person.

It's…

We're not really great at pulling
things off the back room, you know,

out of a box in the back room and
loading them onto the, to the right

shelf in grocery or, you know, folding
sweaters just the right way in apparel.

Um, but the…

But if you can start to, uh, if you
can better organize and orchestrate

and automate the work that needs
to happen, um, you, you've got a

lot of, uh, you know, labor hours
that you can use more productively.

Do you see…

I was talking to someone recently who
said they think this leads to a really

interesting in-store, you know…

So, so the thought, the thought
pattern goes, people are going to…

They're, you know, the more AI
you use, you're gonna crave some

human context on other ones, so
you'll go to the store for that.

And then this will empower the store,
and I think they call this retailing,

where the store associate will now
kind of know more about you, right?

Where, um, s- because, uh, the AI should
be able to have personalization down to

the store and individual associate level.

Do, do you think that's gonna be a thing?

I think so.

I think if you're…

Let's assume you're permi- you know,
let's assume everything is permissioned.

Um, you know, you're, you're in the
loyalty program, you've accepted that,

um, you know, location, uh, location
management and, um, you know, all your,

your push notifications and everything.

So you then should be known when you
walk in Um, and the way I imagine it, I

go back to, uh, um, Devil Wears Prada,
and there was the scene at the, at the

Met Ball where, um, the Anna Wintour
character, um, you know, someone would

come and whisper in her ear who the
person was she was about to talk to.

And that's, that's essentially what
we're talking about putting in place.

You know, you're gonna then-- Your
best sales associates in a high, in

a premium luxury space, they, they do
that already for their best customers.

But what it'll give you the opportunity
to do is ha- get more personalization, you

know, when you're not, uh, you know, uh,
you know, whatever, a billionaire spending

$50,000 at a single, single store visit.

You know, for someone that came in and
maybe hasn't been in in three months and

you don't have your own sales associate,
you'll be able to come in and get a

personalized touch based on what they're
seeing in your profile and the con-

maybe even something about the context of
your visit because they understood what

you were looking at online or that you
do- that you, um, got a, uh, something

in the, the email or the physical mail
recently, um, and that you had responded

to an offer and you've got three things
in your on, in your digital cart,

uh, and two things on your wish list.

So, you know, they can take all of that
information and then think about what

those next best actions are of how to
best serve you In the moment of the

transaction and, and how to, how to
do that with a journey, but also even

more importantly, in the li- in the,
that stage in the lifespan of your

relationship with that store or brand.

How about, I notice you also seem to
be thinking a lot about pricing and

dynamic pricing as it relates to this.

What, what, uh…

You had a interesting thing where,
I forget the name of it, but

where, you know, you had like a,
some s- some moisturizer was like

a different price for everybody.

How, uh, it gets kind of insane.

What, you know, will the, will the
consumer tolerate that and, and where

do you think that takes us, this
like real-time pricing mechanism

that, that's gonna be available?

I think, I think the most practical
thing that will happen is that

there will still be listed prices.

Um, and I think most states require at
least shelf, shelf listing of prices.

I think one or two states might still
require, uh, individual prices on items.

Um, but I think the personalization
that will take place will

be hidden behind the scenes.

So there will be-- Right now a lot
of places have loyalty pricing.

You know, almost every grocer
has the price that you get when

you're in the loyalty club.

What I believe will happen
is that price you will get in

the loyalty club might vary.

That might not be the same
3% off on these, you know,

whatever, 50 items this week.

It might be, um, you know, you might be
a customer they're trying to get to buy

more s- more of one type of product,
so they incentivize you to do that.

And if they don't need to give
me that discount, they might

not give me that discount.

Um, I thought it was really interesting
what, uh, I think I remember the

article you were talking about.

It was on, uh, what Delta is doing.

So Delta is saying, I think
they're gonna have 20% of their

prices will be individually
driven right now, rather than…

Like right now it's based off of
day of week, time of day, I think

some- something about the device.

But if you and I were sitting next
to each other, I don't think we would

get, you know, using a similar device
in a test and control manner, I don't

think we would get a different price.

But, you know, that, the chance
of us getting a different price,

um, is going way up just based on
what they know about us, kind of

quote unquote, "In their file."

Um, and I think that's what'll happen
for, again, it'll be for loyalty club

members because truthfully, consumers
have too much backlash that there's gonna

be surge pricing, you know, at their
grocery store or at the McDonald's and

they think, they think like the price is
gonna change from the time you're in the

back of the line to the front of the line.

Um, and they don't realize, one,
that's terribly hard to execute, and

two Retailers don't wanna do that.

Like, like they just
don't want that headache.

Yeah.

I've, uh, I went to a vis- uh, I've got
a kid going to school, and we went to

one of the department stores, and they
had all the digital tags everywhere.

Uh, and I was thinking, "Sure, sure is
easy for them to update those prices now."

Yeah.

And they will, and they'll update them.

They'll, uh, what they'll be able to
do is they'll be able to update them

nightly, um, and with much more frequency,
and they'll be able to update them

store by store, which most do not do.

Most ev- are, are zone priced
right now, which kind of

goes back to whatever your…

You know, back when you used to
get circulars in the Sunday paper.

You know, the zone would be based on
that, that trading area, and it was

usually driven around newspaper or
media, um, because that's where you

would control what the prices are.

Um, but now you would be able to
do that at a s- do, do that l-

pricing at a store-by-store level.

And if you, let's say, you know, you've
got, um, you know, meat that you see

your, your normal demand is saying you're
not gonna sell out of the meat before it

goes bad, you can actually trigger, just
for one store, the lowering of a price.

It's, it's…

I mean, it really, it goes back to the,
the, uh, the Kmart blue light special.

You know, you'll be able to offer
people store-specific local prices

that are complete surprise and
delight, um, that, you know, give…

It's, again, it's one of those reasons
people love to go into the stores.

Well, I, I, uh, I know you're busy.

You've got a lot of customer
proposals to get done.

Um, so, so I appreciate you taking time.

Anything, any last
thoughts before we wrap up?

The one thing that I keep coming back to
is Consumer behavior changes very slowly

right up until the point that it doesn't.

And, and I think on one hand we
see gradual change, and then we

see this big inflection point.

Um, and having lived through kinda what
happened with e-com, I think that's what

we're headed to over the next, I'll say,
two to, two to three years, um, in the,

you know, AI-enhanced shopping space.

Awesome.

That's a good thought to leave it on.

I, I, I 100% agree, so
we're aligned on that.

Um, thanks for joining us today
and sharing your thoughts.

If, uh, if folks in the, uh, listening
or viewing audience want to learn more

about your thought leadership, where's
the best place for them to follow you?

Uh, you can find me on LinkedIn.

Um, it's Karl Haller.

K- Karl with a K, H-A-L-L-E-R.

Um, and, uh, my company
has a very complex URL.

It's IBM.com.

Thanks, Karl.

We really appreciate you taking time,
and we'll put, uh, links to that in

the show notes, uh, so that people know
how to follow you over on LinkedIn.

But we won't put IBM in there.

I think people can figure that one out.

Yeah.

And, and your spellcheck
will never get it right.

All right.

Thanks very much.

Thank you.

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

to the future of AI shopping agents.