0:00

you have crappy data, you're going to

0:02

have crappy AI, right? So that's a

0:04

lot of companies I know are trying

0:06

to move fast with AI, but their

0:08

data really is not a good place.

0:11

So because we had a good data

0:13

foundation, it allowed us to really take

0:15

advantage of all the innovation that was

0:17

coming from AI and Gen AI and

0:19

eventually agentic AI. Welcome

0:22

to Technovation. I'm your host, Peter Hai.

0:24

My guest today is Shamim Mohamed.

0:26

Shamim is the Executive Vice President and

0:28

Chief Information and Technology Officer of

0:31

CarMax, the largest used vehicle retailer based

0:33

in the United States that earns

0:35

more than $26 .5 billion in annual

0:37

revenue. Shamim's been with the

0:39

company for more than 12 .5 years. He

0:41

took on the CIO role nearly 11

0:43

years ago and added the CTO role

0:45

to his responsibilities six and a half

0:47

years ago. Across his journey, he's

0:49

helped CarMax become a digital -first company.

0:51

He was one of the earliest to

0:53

implement a product operating model. And he

0:55

was also among the earliest adopters of generative

0:57

AI. I look forward to hearing

0:59

more about the progress he and his team continue

1:01

to drive for the company and its customers. Shamim

1:04

is also a board member at

1:06

UNFI for more than three years, joining

1:08

an elite group of tech and digital chiefs

1:10

who joined the boards of publicly traded companies. Shamim,

1:13

welcome back to Technovation. It's great to speak with you

1:15

today. Well, great to see you, Peter. Looking

1:19

forward to the conversation. Yeah, I always look

1:21

forward to our conversation, Shamim. Thank you so

1:23

much. And speaking of which,

1:25

you and I thankfully spend a fair amount

1:27

of time given the proximity between our

1:29

offices. But it's been I think it's roughly

1:31

nine years since we've been on the

1:33

record together. And I've been

1:35

fascinated across the interim

1:37

between then and now in our off

1:39

the record conversations to hear so much about

1:41

what has transformed across. CarMax and among

1:43

the ways in which it's been transformed has

1:45

certainly been digitally and through better use

1:47

of data and technology. I'm looking forward to

1:49

getting into all of the above. But

1:51

I wonder here at the outset if there

1:53

are a few things you would list

1:55

as some of the key attributes that

1:57

are difference makers across the

1:59

past nine -ish years since we last

2:01

went on the record that you would underscore.

2:04

Yeah, it's hard to believe it's been almost a

2:06

decade. So

2:08

over the past decade, What

2:11

has changed is, as you

2:13

know, we've gone through a

2:15

pandemic, global pandemic, and the

2:17

consumers, the world actually changed

2:19

drastically over the last nine

2:21

or so years. So

2:23

from a Carmex perspective, I

2:26

think one of the biggest changes

2:28

that has happened is that

2:30

we have become a digital first

2:32

technology company. And

2:34

also we became

2:36

the most comprehensive. omnichannel

2:39

retailer in our industry. What

2:42

I mean by that is

2:44

where customers can buy or

2:46

sell 100 % online or

2:49

100 % in store or something

2:51

in between. So those

2:53

are two big things that have changed.

2:55

What has not changed in the

2:57

last nine years is we still remain

3:00

A very focused culture

3:02

on collaboration and high

3:04

engagement and just a

3:07

great place to work. And

3:10

we also remain one of

3:12

the most customer -centric companies

3:14

in our industry. And

3:16

one thing that hasn't changed, not just last

3:18

decade, but the last 30 years since our

3:20

company was founded. was that

3:22

we established a company on the

3:24

principles of integrity and transparency. And

3:26

that has not changed over the

3:28

last nine years. That's

3:31

a great overview. Thank you for that, Shamim. Something

3:33

else that has changed is

3:35

your purview. It's grown. When we

3:37

last spoke, you were the

3:39

chief information officer of the company.

3:41

You're now the chief information

3:43

and technology officer. And I'd love

3:46

to understand a bit more what

3:48

the additional CTO said of your

3:50

responsibilities and more generally speaking,

3:52

what your role, how it is.

3:54

is different relative to nearly a

3:56

decade past. Yeah, Peter, the role

3:58

of technology has always been

4:00

important for Carmex from the beginning

4:02

of the company. However, I would

4:04

say over the last eight, nine

4:07

years, the role of technology has

4:09

even become more important, right?

4:11

Because as a technology digital first

4:13

company, that technology plays even

4:15

a much more critical role in everything

4:17

we do, like most companies out

4:19

there now. So CIO, in my view,

4:22

is the role where, you

4:24

know, a CIO has to

4:27

run the business of technology

4:29

very well. It's really focused

4:31

on execution. It's focused on

4:33

keeping things secure. It's making

4:35

sure that it's driving efficiency

4:37

throughout the organization. So it's

4:39

really more like execution focus

4:41

around technology. And then CTO,

4:44

you know, just most technology

4:46

companies, they have CTOs that

4:48

are driving products

4:50

and services, digital capabilities,

4:53

creating competitive differentiations. So

4:56

that's the role I play

4:58

also at CarMax. So this is

5:00

why five, six years ago,

5:02

I assumed the role of CTO

5:04

in addition to being the

5:06

CIO. So I had to play

5:08

both roles very well because

5:10

I get to drive execution effectively

5:12

and then also have to

5:14

drive innovation to be successful in

5:16

my role of a technology

5:18

digital company. To

5:20

what extent, Shamim, did that suggest

5:22

a change of your team, given

5:25

your growth of purview? Certainly,

5:27

of course, some members of the team, or

5:29

at least the roles associated with them, the

5:31

reports that the titles that reports of yours

5:33

might have probably are the same. But I'd

5:35

imagine something has changed as well. How would

5:37

you typify some of the org changes relative

5:39

to your own personal purview changes? Yeah,

5:42

the org has evolved, Peter,

5:44

quite a bit since we

5:46

last interviewed together nine years

5:48

ago and grown exponentially. You

5:50

know, the digital talent, the

5:52

data talent, the AI talent,

5:54

just the overall, the engineering

5:56

talent that's driving a lot

5:58

of this innovative products and

6:00

solutions for our industry, for

6:03

our customers, that has grown

6:05

significantly. And over the years

6:07

also, we continue to fine

6:09

tune and tweak. How do

6:11

we evolve as a digital

6:13

-first company? So a

6:15

lot of those talent and

6:17

groups and resources continue to

6:19

develop in my team. And

6:22

as a company also, we

6:24

can learn how to operate

6:26

like a digital -first business,

6:28

digital -first company. So I've been

6:31

playing basically a larger role

6:33

in that digital journey and

6:35

driving innovation through my colleagues

6:37

and partners. So

6:39

from just a role, you know,

6:41

we have a huge product, as

6:43

you know, we have a big

6:45

product organization that has grown and

6:47

all the product engineering and all

6:50

the product, you know, delivery type

6:52

of resources, they all report to

6:54

me. Now, the product management still

6:56

sits outside of technology, but I

6:58

work very closely with my head

7:00

of product for driving those innovations.

7:02

So it has changed and evolved,

7:04

but we remain very cross -functional,

7:06

very collaborative kind of organization. I

7:08

think that is the root of

7:10

a lot of the success we

7:12

have as a company. So I

7:14

think that has remained the same.

7:16

But from my role, I think

7:18

it's just playing a bigger role,

7:20

larger role, a strategic role, driving

7:22

innovation, driving growth for the company

7:24

through technology. That's what I've been

7:26

doing. Yeah, fascinating. You've mentioned now

7:28

appropriately a few times the notion

7:30

of CarMax being a digital -first company.

7:32

I wonder if you could talk

7:34

a bit about What are the

7:36

ingredients to doing that? How

7:38

has that manifested itself? There are several

7:41

companies, not all, of course, but that refer

7:43

to themselves as such. And as an

7:45

organization and as a leader who I know

7:47

has been thinking a lot about how

7:49

to breathe life into this vision, I'd love

7:51

to understand some of the necessary ingredients

7:53

to do so. Yeah,

7:55

there are a few key components

7:57

to be a digital first company,

7:59

Peter. In my view, if

8:01

you're a legacy organization, you have a

8:03

lot of legacy technologies. Although

8:05

you want to be a

8:08

digital company, it's hard to

8:10

do because the digital companies have

8:12

certain characteristics. They're nimble. They

8:14

test and experiment quickly. They're

8:16

able to adapt quickly with what's

8:18

happening in the marketplace. They're

8:20

able to keep up with the

8:23

consumer expectations. They're also able

8:25

to drive innovation, innovative culture, innovative

8:27

tool set for their internal

8:29

associates. So those are some

8:31

of the characteristics I see for

8:34

a digital company. So for

8:36

us, it's a few things.

8:38

First, we continue to transform our...

8:41

core technology foundation. So I call

8:43

it the digital backbone. You

8:45

have to have very robust, scalable,

8:48

adaptable technology foundation. So that's

8:51

today CarMax majority for technologies

8:53

now are running on modern

8:55

cloud -based technology platforms, which gives

8:57

us tremendous ability to innovate

8:59

at a rapid pace. Then

9:01

a lot of the characteristics,

9:03

digital practices that you would

9:05

see from startup environment companies

9:07

or pure technology companies, those

9:09

practices and principles are some

9:11

of the things that we...

9:13

We practice, we adopt it

9:16

over the years and continue

9:18

to fine tune them. The

9:20

product mindset, the product approach

9:22

that we have adopted almost

9:24

a decade ago, we continue

9:26

to evolve that. And that

9:28

product mindset is a big

9:30

part of being a digital

9:32

first company. And ultimately, it's

9:34

all about driving innovative products

9:36

and solutions for our associates

9:39

so they can take care

9:41

for customers better and also

9:43

for our customers so that

9:45

they have the the best,

9:47

most innovative tools so they

9:49

can interact with CarMax in

9:51

the best possible way. So

9:54

Omnichannel, again, was a great example

9:56

of that. And last thing about

9:58

digital first company, in my view,

10:00

is a company that is able

10:02

to leverage and unlock the potential

10:04

of data and the insight and

10:06

being able to take advantage of

10:09

the evolution and exciting thing that's

10:11

happening in AI to bring those

10:13

innovations, leveraging the data platforms they

10:15

built. Fantastic

10:17

insights, Shamim. And I'd love to double click

10:19

on the data aspect of that a little

10:21

bit further, if I may. Can

10:24

you describe some of the building blocks

10:26

you and the team have put in

10:28

place to ensure you have the right

10:30

data strategy and hygiene and governance to

10:32

be situated to derive more value from

10:34

it? I'd love to get to maybe

10:36

in a subsequent portion of our conversation,

10:38

the AI aspect of this. But before

10:40

we do, I think covering the foundational

10:42

elements. are critically important since there will

10:44

be a governor on how far you

10:47

can go with AI if you don't

10:49

have great data practices in place. Let's

10:51

cover some of those data practices as

10:53

you see. You're absolutely right, Peter. I

10:55

think data is the lifeblood, right? People

10:57

of an organization is the most valuable

10:59

asset. I would think the next valuable

11:01

is really the data. So

11:03

CarMax, we had good data.

11:05

We had some of the best

11:07

and most comprehensive data related.

11:10

to buying and selling used cars.

11:12

So the problem we had originally, or

11:15

many years ago, was that we

11:17

had all this data, but they're not

11:19

organized, they're not available in a

11:21

way that we could leverage the best

11:23

possible way. So one of the

11:25

first priority we had was to establish

11:28

a good data governance. So part

11:30

of the data governance was really ensure

11:32

the quality of the data. Because

11:34

we can have all the data, but

11:36

if we don't have the good

11:38

quality... And if we don't have the

11:40

right confidence on the data, then

11:43

what the decisions and insights coming out

11:45

of this data, we would not

11:47

be able to make a good decision

11:49

or feel good about it. So

11:51

we put a lot of emphasis on

11:53

the data governance, data quality, and

11:55

data ownership across the company. The data

11:58

is not just a technology thing.

12:00

It's cross -functional. And the data owners,

12:02

in most cases, reside outside of technology.

12:04

So having that sense of ownership

12:06

on the data. And then having good

12:08

data governance really was the foundation

12:11

of what we had to do early

12:13

on. So once you

12:15

got that established, the next thing

12:17

we really focused a lot

12:19

on, making sure the data is

12:21

available and we have the

12:23

right technology to manage those data

12:25

elements, manage the data, and

12:27

make sure those data is available

12:29

across the company. So we

12:32

put a very strong, actually a

12:34

startup company, but now it's

12:36

done really well. We implemented a

12:38

robust master data management tool

12:40

that's cloud -based. So that helped

12:42

us really organize our data in

12:44

a good way. And then

12:46

we moved, we platform our main

12:48

data warehouse. We built, moved

12:51

things in the cloud. So all

12:53

those things helped us to

12:55

provide the scale and this really

12:57

strong analytical capability without being

12:59

constrained by the capacity of the

13:01

technology. So that really helped

13:03

us. From a tool and technology

13:05

perspective, the last thing I

13:07

would say is talent. We

13:10

really upgraded and added a

13:12

lot of talent around data. So

13:14

it's the data scientists, the

13:16

data engineers, the data

13:19

architects, analysts. We built a community

13:21

of strong data leaders across

13:23

the whole company. And that really

13:25

also helped us drive. So

13:27

now we have the quality of

13:29

the data. We have a

13:31

good technology to drive. the

13:33

data, and then we have great

13:35

talent that is really driving a lot

13:37

of the innovation on the data

13:39

platforms we built. Very

13:42

well articulated, Shamim. Thank you so

13:44

much for walking me through the steps

13:46

of that process. Let's now talk

13:48

about then taking advantage of the foundation

13:50

that you've laid. You mentioned a

13:52

bit earlier the power of artificial intelligence,

13:54

and I know something that you

13:56

and the team have been working with

13:58

for quite some time now. And

14:00

talk a bit about how you think

14:02

about Maybe we talk about analytical,

14:04

more traditional artificial intelligence and generative AI

14:06

in two different buckets. Maybe we'd

14:08

begin with the former, if you don't

14:10

mind. Talk a bit about

14:12

the institutionalization of artificial intelligence, again,

14:14

based upon the great foundation that you

14:16

and the team have laid from

14:19

a data perspective. How have you been

14:21

leveraging that? Yeah,

14:23

so Peter, as you mentioned, as you

14:25

know, the good AI needs good

14:27

data. So if you have crappy data,

14:29

you're going to have crappy AI,

14:31

right? So that's a lot of companies

14:33

I know are trying to move

14:35

fast with AI, but their data really

14:37

is not a good place. So

14:39

I would say that focus, because we

14:41

had a good data foundation, it

14:43

allowed us to really take advantage of

14:45

a lot of the innovation that

14:47

was coming from AI and Gen AI

14:49

and eventually agentic AI is what

14:52

I'm most excited about. And I'm sure

14:54

we'll have some time to talk

14:56

about that. In

14:58

our early days, we focused

15:00

a lot on just the

15:02

AI and machine learning. And

15:04

those efforts were focused on

15:06

just improving a lot of

15:08

the operational processes and really

15:11

providing a good experience for

15:13

our customers. For example, I

15:15

think we might have talked

15:17

about this before. We used machine

15:19

learning algorithms based on the

15:22

data to build a nice recommendation

15:24

engine. where customers are

15:26

shopping for cars. So it

15:28

really made the car buying

15:30

process simpler and a little

15:33

bit more easier for our

15:35

customers as they're shopping. We

15:37

used machine learning and AI

15:39

capabilities to drive digital merchandising

15:41

improvements. So like how we

15:44

make sure that with the

15:46

cars we're selling, they're coming

15:48

across as realistic as possible

15:50

with all the, you know.

15:53

glitches or all the uh you

15:55

know just as it is right

15:57

actual cars uh that are showing

15:59

up just like the way it

16:01

would be in a showroom uh

16:03

or in the stores making it

16:05

very real and practical for customers

16:07

so used ai there uh one

16:09

interesting innovation there i remember a

16:11

few years ago our engineers developed

16:13

something called a digital sweeper it's

16:15

like sweeping the floor because we

16:17

take pictures of cars, right?

16:19

Every car we sell, we have to

16:21

uniquely merchandise and we have to uniquely

16:23

take pictures and present it on the

16:25

website. So the customers who are shopping,

16:27

they can actually look at it and

16:29

get a good feel for it. So,

16:32

but problem is, know, we drive

16:34

the car into a studio, then we

16:36

take a picture, this little turntable,

16:38

and sometimes the turntable could be dirty

16:41

because tire marks and things like

16:43

that. So our engineers figure out a

16:45

way to use AI to clean

16:47

the floor. And, you

16:49

know, without changing the car, the car

16:51

actually is the same car, presented

16:53

exactly the way it is, reality in

16:55

real. But they use the digital

16:57

super technology, AI technology to

16:59

get rid of all this thing around

17:01

the car. So the car is the

17:04

car. So actual car with the right

17:06

visualization by using AI technology. And that

17:08

is actually patented. We're able to patent

17:10

that because nobody else has it. So

17:12

this is a simple example of how

17:14

we have been leveraging AI from the

17:17

beginning. Now, I know

17:19

we're to get into GNI.

17:21

We are also one

17:23

of the early companies to

17:25

start testing and embracing

17:28

the GNI. Our teams

17:30

were using GNI long before the

17:32

chat GPT or GNI became

17:34

a big hot topic, probably six

17:36

to 12 months before it

17:38

happened. And partly because, again,

17:40

as you know, we have a

17:42

product kind of team and

17:44

culture mindset. Our teams are always

17:46

looking for ways. to embrace

17:48

new technologies, new tools, new capabilities

17:50

to drive a certain business

17:53

outcome. So one of the

17:55

team figured out, hey, we have all

17:57

this massive amount of customer reviews,

17:59

massive amount of changes in the cars,

18:01

description and all those things. We

18:03

can use actually a chat GPT to

18:05

organize the content, organize the new

18:08

car information and present it in a

18:10

way that's easier for the customers

18:12

to review and assess and evaluate a

18:14

car. And we had that in

18:16

production, you know, like I said, six

18:18

to 12 months before that GPD

18:20

was a big popular thing, global phenomenon.

18:24

Fascinating, Shamim, and so many things

18:26

I'd like to unpack there. Let's

18:28

talk a little bit further about now,

18:30

especially given the fact that you were six

18:32

or 12 months ahead of the world

18:35

in many cases, in terms of your use

18:37

of it and not unlike the product

18:39

operating model, which, by the way, I also

18:41

would like to cover further in this

18:43

conversation. You have a lead on many. That

18:45

is to say, you've matured further. as

18:47

a result of simply time in addition of

18:49

course to experimentation and validation of value

18:51

and so forth talk a bit about from

18:53

a generative ai and again i would

18:55

like to eventually get into agentic but for

18:57

a generative ai perspective where you've seen

18:59

the most value as well as maybe some

19:01

of the things you're most excited about

19:03

in the use of it yeah i think

19:06

you know as you know we are

19:08

probably still uh in the early stages of

19:10

what ai can do for for the

19:12

world There's a lot of

19:14

excitement, a lot of innovation, a

19:16

lot of hype. It's really

19:18

being able to identify what's real

19:20

and what's not. And more

19:22

importantly, what of the AI capabilities

19:24

available we can use responsibly to drive

19:26

a business outcome, right? Ultimately, if we

19:28

focus on the business outcome, then we

19:30

can find the right tool to do

19:32

it. And AI sometimes may not be

19:34

the right tool, right? That's something to

19:36

keep in mind. In our

19:38

case, Where we have the

19:41

GNI, I found the most

19:43

benefit is really aggregating a

19:45

large amount of data and

19:47

content and being able to

19:49

organize and drive insight from

19:51

it. One area that we

19:53

are taking advantage of is

19:55

really around our internal knowledge

19:57

base, our customer, our information

19:59

about different rules and policies.

20:03

different requirements of use card because

20:05

it can vary from locality to

20:07

locality. So we created an internal

20:10

GNI -based tool called ROADS, R -H -O

20:12

-D -E -S. So that's really empowering

20:14

our associates internally to be able

20:16

to get to information but organized

20:18

in a way that's greatly simplifying

20:20

their job and they're able to

20:22

get to information quickly. And then

20:24

they can help the customer. So

20:27

that's one thing we are very

20:29

excited about. The other one is,

20:31

we call it Sky, S -K -Y -E.

20:33

This is more for the customers,

20:35

where as they're buying or selling

20:37

a car or financing a car,

20:39

this Sky tool allows them to

20:41

really navigate through that whole buying

20:44

and selling process. So the two

20:46

areas that we are very excited

20:48

about, we're experimenting in a

20:50

lot of different areas. I think

20:52

the engineering side, here's a huge opportunity,

20:54

right? As you know, there's always

20:56

going to be shortage of technical talent,

20:59

engineers, developers. So

21:01

using code generation tools or

21:03

AI to generate code that

21:05

our engineers don't need to,

21:07

right? They can still over

21:09

review and they can still

21:11

make sure that it's done

21:13

securely and all those things.

21:16

But ultimately, though, if

21:18

we can take 20, 30 % of

21:20

the coding done through AI, that

21:22

can greatly increase our capacity and

21:24

we can actually drive that newfound

21:26

time. to drive more value and

21:28

that the way we can mostly

21:30

use those very precious and scarce

21:33

developer engineering resources so that's an

21:35

area also i'm very excited about

21:37

very interesting so so uh you

21:39

see this as more of an

21:41

augmenter there are those who pontificate

21:43

that in the shorter even the

21:45

shorter middle medium term this could

21:47

could lead to competition, if you

21:49

will, between artificial intelligence and human

21:52

with AI perhaps winning, meaning less

21:54

necessity for humans. You see this

21:56

rather as something that adds efficiency

21:58

that allows humans to do more

22:00

as a result of the time

22:02

that is freed up. If I

22:04

recap that well. Yes, Peter,

22:06

that's very well said. I don't

22:09

think, mean, some roles might

22:11

change and evolve. And we've seen this

22:13

over the, I don't know, 200 years,

22:15

right? New technology comes, to take everybody's

22:17

job. But in my view, AI is

22:19

just like any other tool we've seen

22:21

in the past. However, it is

22:23

the most transformative tool that we might

22:25

have seen in our lifetime, right? So

22:27

it comes with a lot of possibility.

22:29

But ultimately, though, this is meant to

22:31

be, in my view, this is going

22:33

to be making the humans better. It's

22:35

really more augmenting humans. And in some

22:37

cases, it will replace some of the

22:39

tasks that are not value added. It's

22:41

not really accretive. It's not really driving

22:44

a lot of value to the companies

22:46

or individuals. So we can use AI

22:48

to simplify and automate and make those

22:50

things better. But is this going to

22:52

free us our time to do more

22:54

things where we can drive most value

22:56

for the organizations and for ourselves? So,

22:58

yeah, I don't see it taking everybody's

23:00

job. Yeah, that's really well articulated. Thank

23:02

you. You mentioned earlier, and I'd love

23:04

to turn to it now, how excited

23:06

you are about agentic artificial intelligence, certainly

23:08

in many ways the next wave of

23:10

innovation that is as being predicted and

23:12

even in some cases realized. Talk a

23:14

bit about why you are so excited

23:17

about it, Shamim, and again, where you

23:19

foresee some of the applications of it

23:21

in a business like yours. Yeah,

23:23

so Peter, that's where this goes back to

23:25

the last conversation we just had. I

23:28

see the agents. is

23:30

the next evolution of GNAI. So

23:32

you can think about machine learning

23:34

AI, GNAI, agentic

23:36

AI. So agentic AI basically

23:38

is where you build

23:40

those independent, with certain guardrails,

23:42

independent agents to take a part

23:44

of the role that you may

23:46

have, automate certain parts of the

23:48

process or certain use cases so

23:50

that that does the job for

23:52

you, that frees up your time.

23:54

So think about what I call

23:56

a mechanic, for example. So, you

23:58

know, we have technicians, they're working

24:00

a lot of cars. If we

24:02

have an AI agent that's helping

24:05

them, making them better and smarter

24:07

in terms of how to repair or

24:09

recondition a car, that could be one example.

24:11

Or you can think about, know,

24:13

I gave you an example of

24:15

Sky, which is a GNI tool to

24:17

help you buy or sell a

24:19

car. Now we can make it an

24:21

agent so that that shopping agent

24:23

can help you do most of the

24:25

work for you. and come back

24:27

with a recommendation to you as a,

24:29

hey, Peter, based on you

24:32

and your preference, this is the

24:34

car you should buy and these

24:36

are all the things I've taken

24:38

care of for you. So those

24:40

are like a couple of examples,

24:42

but I can see that pretty

24:44

much every role, every business process

24:46

function, there could be some agent

24:48

or small agent that has a

24:50

very specific purpose. It's not this

24:52

agent does everything, right? But the

24:54

agent can take a certain part

24:56

of the function or role or

24:58

task. and then automate and operate

25:00

in a way based on the

25:02

guardrails or the responsibilities we have

25:04

assigned to that particular agent. So

25:06

that's how I view this evolving,

25:08

which is exciting because until now,

25:10

a lot of people are experimenting

25:12

with Gen AI and other things,

25:14

but ultimately I think the agentic

25:16

AI is where we're going to

25:18

see a lot more benefits coming

25:20

out of AI. That's why I'm

25:22

very, very excited about it. I

25:24

can sense your passion for it

25:26

and really interested to continue to

25:28

remain abreast of the progress you

25:30

and the team make relative to

25:32

it as well. I know the

25:34

most recent conversation you and I

25:36

had in your offices in Richmond

25:38

a few weeks back, you talked

25:40

about the increasingly complex partner ecosystem

25:42

that an executive like you must

25:44

manage. And this is a fact

25:46

of life and an essential element

25:48

to managing a great tech and

25:50

digital. uh organization and ultimately a

25:52

a company at the scale of

25:55

yours um talk a bit about

25:57

how you best manage that ecosystem

25:59

by having so many partners obviously

26:01

it introduces complexity but as i

26:03

mentioned it's also a source of

26:05

speed and insight and innovation as

26:07

well. And so one must, in

26:09

order to accomplish the latter, one

26:11

must grapple with and the complexity

26:13

associated with that. Talk a bit

26:15

about how you think about best

26:17

doing so such that you derive

26:19

as much value as possible from

26:21

the ecosystem you curated. Yeah,

26:23

I think at first is

26:25

recognizing that to be successful,

26:28

no company or no team

26:30

can do it alone, right?

26:32

We have to rely on

26:34

the ecosystem. So the

26:36

more effective we can make our

26:38

partner ecosystem, the better we will

26:40

be. So in my view,

26:42

the first thing you're going to do,

26:44

understand, you may have 50 different suppliers

26:46

of technology that company you work with.

26:49

Identify which ones are your most

26:51

important ones and which ones are

26:53

most strategic. And then don't look

26:55

at them as a supplier. Look

26:57

at them as partner, right? They're

26:59

not just the company that trying

27:01

to. get revenue or sell

27:04

you more stuff, but they are your

27:06

partner. You need to co -create, co

27:08

-innovate, co -progress together. We have

27:10

this win -win kind of mindset

27:12

that you can help them, they

27:14

can help you. And you do

27:16

that by investing time with them. Spend

27:19

a lot of time with not

27:21

just the sales team on those

27:23

companies or partners, but also invest,

27:25

find a way to engage with

27:27

their product development, their engineering teams,

27:29

engineering leaders. Because they're the one

27:31

actually working on things that you

27:33

must be most interested in. So

27:35

that's one really important one. The

27:37

other thing is, you know, these

27:40

VC firms, venture capital firms, and

27:42

a lot of the startups, spend time

27:44

with them also. Because I get

27:46

so much energy by talking to a

27:48

founder or a CTO of a

27:50

small startup. They are doing things. They're

27:52

going for the impossible. Some will

27:54

be successful. Some will not be. But

27:56

just talking to them, you can

27:59

see the vision, right? You can understand

28:01

the energy they have. They're going

28:03

after things that are disruptive. So just

28:05

staying connected to that kind of

28:07

group of folks is also very important.

28:09

And then lastly, I would say

28:11

is don't be afraid to switch out

28:13

a technology or a platform as

28:15

your business is evolving. In our case,

28:17

we build an architecture. that we

28:19

can plug in and out different partners

28:21

based on what's happening. Because most

28:23

partners are also not all innovating at

28:26

the same pace. So some are

28:28

falling behind, some are winning. Then make

28:30

sure that you are staying with

28:32

the winning kind of partners and not

28:34

getting stuck with legacy old companies

28:36

that really cannot change. Because for you

28:38

to innovate and for you to

28:40

move at the speed you need to

28:42

move, you have to make sure

28:44

your partners are also moving along. And

28:46

reinforce that with the partners when

28:48

you talk to them. That's so well

28:50

articulated, Shamim. I love that notion.

28:53

I've not heard others say it quite

28:55

like you did, that if you're

28:57

going to continue to innovate, if you're

28:59

going to continue to be a

29:01

leader, you need to surround yourself with

29:03

partners who are doing the same

29:05

thing. And understandably, if you have a

29:07

collection of those, they are not

29:09

all going to be progressing at the

29:11

same rate and so continue to

29:13

evaluate the progress or lack thereof and

29:15

making changes where necessary so that...

29:17

lack of innovation, their lack of changes

29:20

and impeding your progress becomes an

29:22

important set of factors. Very nicely said.

29:24

Thank you so much. I want

29:26

to turn to the product operating model.

29:28

You mentioned it earlier a couple

29:30

of different times appropriately, and you were

29:32

one of the first practitioners I

29:34

can remember explaining this to me. I

29:36

remember where we were. uh

29:38

in an innovation lab of yours uh in

29:40

richmond uh and you shared with me as

29:42

we walked the floor this again nine years

29:44

ago as we walked the floor you pointed

29:47

to cohorts of people that were from multiple

29:49

parts of the organization um and you said

29:51

if you look around them and you heard

29:53

them talk you wouldn't know who the it

29:55

person was who the marketing person was you

29:57

know who was from which discipline because

30:00

they've sort of melded together in

30:02

such an interesting and productive way. Again,

30:04

this is nine years ago. This

30:06

is a set of practices that so

30:08

many of your peers are just

30:11

now introducing into their organizations. And so

30:13

much like the prior conversation about

30:15

AI. You have a lot of turns

30:17

at this. You have evolved tremendously

30:19

through the model. And I'd love to

30:21

understand how it has evolved at

30:24

CarMax. What's different today relative to recent

30:26

years as you continue to progress

30:28

the model that you would underscore as

30:30

areas that your peer group might

30:32

consider striving for as well? Yeah.

30:35

So, yeah. So,

30:37

we still remain

30:39

very committed. And

30:41

we remain a firm believer

30:43

of the product operating model. And

30:45

we are so glad that

30:47

we started the journey many, many

30:50

years ago because it gave

30:52

us so much flexibility. It allowed

30:54

us to be so adaptive

30:56

over the years. So we remain

30:58

very committed. And the core

31:00

principles of the product model still

31:02

remain the same. The core

31:04

foundation is cross -functional team. It's

31:06

a mission base. It's a

31:08

culture of experimentation and test and learn. And

31:11

always looking forward that what's going to

31:13

drive the best business outcome. So those

31:15

principles still remain the same. We still

31:18

have the product teams are still very

31:20

small, seven to nine people. So

31:22

those things remain there. What we

31:24

have learned over the years, though, that

31:26

how the product teams are organized,

31:28

like the individual team, how the group

31:30

and how they report up and

31:33

how they structured. Some of those things

31:35

we had to change over the

31:37

years because our business also changed. right

31:39

before we had all the product

31:41

teams organized you know based on the

31:43

customer shopping journey that was really

31:45

when we talked originally many years ago

31:47

but now they're more focused on

31:50

really three big sort of groups we

31:52

call it a platform we there's

31:54

one group called retail the other group

31:56

called supply so retail mean anything

31:58

to do with any customer associate experience

32:00

so retail then we have supply

32:02

anything you do it buying and procuring

32:04

the cars and anything to do

32:06

with logistics. So moving the inventory and

32:09

getting the inventory, that's the supply.

32:11

And then in the platform, this is

32:13

like a shared set of products

32:15

and services that both of those retail

32:17

and supplier are using. So a

32:19

little different than how they were before,

32:21

but the individual product teams remain

32:23

exactly the same way like they were

32:26

many years ago. The other big

32:28

thing is we've grown. When

32:30

we talked nine years ago, we

32:32

had, maybe a handful of

32:34

product teams, right? We probably had

32:36

25, maybe at most 30.

32:39

Today, we have 70 plus product

32:41

teams. So, and today, everything

32:43

pretty much. Everything that's driving innovation,

32:45

driving customer experiences, driving competitive

32:47

advantage, is done through these

32:49

product teams. So, product principles and

32:52

everything has become even much

32:54

more ingrained in the whole company.

32:56

And it goes back to

32:58

the very first question you had

33:00

is about being a digital

33:03

first business. I believe being a

33:05

product -centric organization is what's enabling

33:07

us to remain a digital

33:09

first business and grow as a

33:11

digital first business. Super interesting.

33:14

Again, appreciate you sharing those perspectives.

33:16

You know, we keep returning

33:18

to the duration of your

33:20

role, or at least the duration

33:22

of our collaboration and friendship.

33:24

But you've been with the company

33:27

for nearly roughly 12 and

33:29

a half years. You became CIO.

33:31

nearly 11 years ago, six

33:33

and a half years since adding

33:35

the CTO responsibilities. It

33:38

strikes me that part of

33:40

the richness of your team's accomplishments

33:42

are aided by the fact

33:44

that you have been with the

33:46

organization through so many different

33:49

cycles. You have introduced technology

33:51

that has been state of the art,

33:53

seen it through all the way to

33:55

the point of retiring it in your

33:57

setting, and then replacing it with the

33:59

next facet of technology that is state

34:01

of the art. You have had uh

34:03

as time has gone on a long -term

34:05

set of a long -term perspective to

34:07

the decisions that you have made different

34:09

from peers of yours that stay two

34:12

years here you know a year and

34:14

a half there two years there etc

34:16

you have seen so much within the

34:18

setting even of course as your responsibilities

34:20

have grown so it's not the same

34:22

job as we've highlighted uh um across

34:24

this entire entire duration but i wonder

34:26

you know in addition to all the

34:28

exciting things you've talked about, in addition

34:30

to, of course, the expansion of your

34:32

responsibilities, which I have to imagine adds

34:35

a degree of newness and freshness and

34:37

added excitement as that has happened. What

34:39

has kept you with CarMax so long?

34:41

How has the organization benefited from having

34:43

such a strong leader as you with

34:45

your hand on the steering wheel for

34:47

such a long time? Yeah,

34:49

that's an excellent question, Peter.

34:51

I believe... There's a single answer

34:54

to that. The reason I

34:56

remain at CarMax is because of

34:58

the culture, the people. And

35:01

let me elaborate that a

35:03

little bit. I think the culture

35:05

that CarMax has is really

35:07

highly collaborative. That is something I

35:09

thrive on. I like to

35:11

be part of a team where

35:13

people are generally driving in

35:15

the same direction, but they're very

35:18

collaborative and they're very... open

35:20

to getting and providing

35:22

feedback and then working together

35:25

to improve as an

35:27

organization. So that's really what

35:29

the CarMax culture is.

35:31

People first, you

35:33

are encouraged to take steps,

35:35

you're encouraged to learn new

35:38

things, encouraged to take chances,

35:40

and it's okay to occasionally

35:42

not be successful and learn

35:44

from it. So because of

35:46

the culture of that... sense

35:48

of collaboration and mission -based organization,

35:50

we are very different today.

35:52

If you think about even

35:54

30 years ago versus even

35:57

nine years ago, we are

35:59

a completely different organization that

36:01

is a digital first, that

36:03

is still a market industry

36:05

leader in our space, especially

36:08

I believe we are the most

36:10

innovative company in our industry still.

36:12

And we are still the largest

36:14

seller of used cars in the

36:16

US. So all those things are

36:18

still true, but the culture piece

36:20

is really what's helping us to

36:22

remain operating in that mode. And

36:24

to me, it doesn't feel like

36:26

I've been here almost 13 years.

36:28

It feels like I just started

36:30

because I see the opportunities are

36:33

immense in terms of the company

36:35

where we're going. Huge marketplace, lots

36:38

of... opportunities to improve

36:40

and continue to be the most customer

36:42

-centric company in our industry. I mean,

36:44

those things, I mean, excites me, you

36:46

know, because I see the possibilities are

36:48

immense because, you know, it just, I

36:50

feel actually more excited today than I

36:52

did probably when I joined the company

36:54

12 and a half years ago, because

36:56

that time it's a new, a little

36:58

bit nervousness about the joining a new

37:00

company, new industry. But now I'm like,

37:02

I know, I see the possibility, I

37:04

see the opportunities. And all those things

37:06

excite me a lot. So that's what it's giving me

37:08

a culture, the people that I work with day in

37:11

and day out. Remarkable.

37:13

Thank you so much for those

37:15

reflections, Shamim. I wanted to also ask

37:17

you, you have gained entry to

37:19

the board level. You're a technology leader

37:22

who's joined the board of a

37:24

consequential organization. A little more than

37:26

three years ago, you joined the board

37:28

of United Natural Foods. And I wonder

37:30

if you could reflect upon your time

37:32

as a board member. um advantages that

37:34

you that have accrued to you uh

37:36

as a result of of joining it

37:38

um and any sort of further reflections

37:40

you might have about what a did

37:42

tech and digital leader adds to a

37:44

board like the one that you're a

37:46

part of as well yeah the board

37:48

joining the board has been one of

37:51

the most rewarding experience for me One

37:53

of the things that drives me is

37:55

the ability to learn and make an

37:57

impact. I've been with CarMax for a

37:59

while and I know our business very

38:01

well and still very exciting changes happening

38:03

in our business and those are great.

38:06

However, being on a board of a

38:08

company which is a completely different

38:11

industry, very complex business, very large business,

38:13

has given me the opportunity to

38:15

learn even more. It's challenging me to

38:17

think a little bit differently. It's

38:19

benefiting in two ways. One, I'm

38:21

learning. things from the other board

38:23

and other business and with other

38:25

board members. There are amazing board

38:27

members, learning from them that I

38:30

can bring back to CarMax. At

38:32

the same time, all my experience

38:34

at CarMax and before, I'm able

38:36

to add and bring to the

38:38

boardroom at the other company. So

38:40

this is like benefiting both organizations

38:42

because I'm the common element in

38:44

both of them. You know,

38:46

the reason it's important and the reason

38:48

I believe I was

38:50

selected to be on this, I

38:52

have the opportunity to be

38:54

on this board is because of

38:56

my digital experience and transformational

38:59

experience. Every company, every company that

39:01

wants to remain relevant, wants

39:03

to drive shareholder value, they must

39:05

depend on technology. This is

39:07

so important now than ever. So

39:09

having a digital leader on

39:11

the board makes it a little

39:13

bit... easier to have the

39:15

right conversation about technology, about innovation,

39:18

and driving competitive differentiation through

39:20

technology. The other element I think

39:22

is important also, the operational

39:24

risk. A lot of the risk

39:26

may come from technology. It

39:28

could be cyber risk. It could

39:30

be business continued risk. So

39:32

as a CIO sort of role

39:34

in my Carmack, that's the

39:37

thing that I'm very focused on.

39:39

So I'm able to bring

39:41

that perspective, experience, and help. the

39:43

other board members who may

39:45

not be as digitally savvy to

39:47

bring that, just elevate the

39:49

overall digital competency in the boardroom.

39:51

So I believe that, and

39:53

I strongly believe that all boards

39:56

should have very strong digital

39:58

leadership because most of the disruptions

40:00

in any industry is coming

40:02

from the digital. So

40:04

not having the digital competency at

40:06

the right level could be dangerous

40:08

for any company. Nicely

40:11

said. Thank you so much. I want

40:13

to also ask you, Shamim, we

40:15

talked about numerous trends through this conversation,

40:17

certainly, and areas that excite you,

40:19

areas you and the team are investing

40:21

in. You mentioned that part of

40:23

the culture of CarMax, in fact, is

40:25

one of experimentation, of leaning towards

40:27

trends as they emerge, teasing them out

40:29

a little bit, understanding the application

40:31

or lack thereof to an organization like

40:33

yours and potentially helping drive better

40:35

outcomes, reach a destination intended by the

40:37

strategy more readily, et cetera. I

40:39

wonder, are there additional trends beyond the

40:41

ones we've covered that excite you

40:43

as you look to the future? So,

40:47

Peter, one thing that I told

40:49

you 90 years ago is still

40:51

very relevant, which is the most

40:53

important thing we all should be

40:55

doing is positioning our companies for

40:57

rapid change. I mean, that is

40:59

probably more true today than it

41:01

was 90 years ago. So very,

41:03

very, very, very, very focused on

41:05

improving the organizational agility. through

41:07

tools and processes and technology. That's something that

41:09

has to happen. Now, once a

41:11

company feels like it's fairly nimble,

41:13

then they can start tackling some

41:15

of the big things like new

41:17

things are coming out, especially agentic

41:20

AI, to me, is the next

41:22

game changer for our world, actually.

41:24

So AI, gen AI, great, but

41:26

the agentic AI is where the

41:28

true value of the next AI

41:30

we're going to see, because a

41:32

lot of companies have not seen

41:34

the results. from Gen .AI just

41:36

because it's still kind of experimentation,

41:38

but agentic AI is where I

41:40

believe we'll see a lot of

41:42

value across the board. That's a

41:44

little bit more practical and a

41:46

little bit more usable as it

41:48

relates to driving business outcomes. So

41:50

that's what I'm most excited about

41:52

besides keeping our business nimble. I

41:54

mean, that is really a priority

41:56

for me still. Wonderful.

41:58

Yeah, it should be my... A

42:00

reference to nine years ago and the

42:02

remarkable insight you provided, which I also

42:04

highlighted in my most recent book, Getting

42:06

to Nimble, this insight that you offered,

42:08

which was very inspiring for me about

42:10

the necessity to make sure that, I

42:12

think the way I'm paraphrasing, but the

42:14

trend in two years or three years

42:16

that may be most important might be

42:19

something unknowable today. And therefore, setting up

42:21

a Nimble organization to be able to

42:23

seize those opportunities becomes that much more

42:25

important. So really appreciate you. The insights

42:27

you provided there, in addition to some

42:29

of the more specifics of areas that

42:31

you you're excited about investing in. Well,

42:33

speaking about things you read, in addition

42:35

to things you watch and listen to,

42:37

I wonder anything you've recently read or

42:39

watched and or listened to that you

42:42

would recommend to peers. Absolutely,

42:44

Peter. One of the

42:46

books I still refer to is the

42:48

book by Peter Hai called Getting to

42:50

Nimble. I highly recommend it.

42:52

I was speaking at a conference

42:54

just a couple of weeks ago in

42:56

Texas. I actually bought a

42:59

few copies of Getting to Nimble. I gave

43:01

it to the audience and they loved

43:03

it because I think people just forget sometimes

43:05

how important it is. They're

43:07

chasing the new technology, new tools, but

43:09

they forget that there are some basic things,

43:11

the framework you provide. very nicely in

43:13

the book is really people need to pay

43:15

attention to. And they can be so

43:17

much better if they focus on some of

43:20

those elements. So I definitely recommend that

43:22

book. The other book that I like by

43:24

another Peter, his name is Peter Weil. And

43:26

some friends from MIT, they wrote

43:29

a book called Future Ready. You

43:31

know, it's a good compliment to

43:33

your book, Getting to Nimble. So

43:35

I would recommend Future Ready by

43:37

Peter Weil and a few folks

43:39

from MIT. I think it's published

43:42

by Harvard Press. And the last

43:44

book that I actually have bought

43:46

a bunch of copies and gave

43:48

it to a lot of my

43:50

leaders to read is a book

43:52

called Think Like a Rocket Scientist

43:55

by an author named Ozan Varol,

43:57

V -A -R -O -L. And

43:59

he himself was a rocket scientist.

44:01

but here some principles in the

44:03

book. That's really great. I mean,

44:05

I of, I was excited to see

44:07

those articulated in that way because it talks

44:09

about embracing uncertainty, launch and

44:11

iterate, moonshot thinking, you

44:13

know, some of those things are I

44:15

think, really all leaders can benefit from.

44:17

So those are the three books I

44:20

would uh recommend uh anyone um in

44:22

in a cio ctu role any leadership

44:24

role. Thank you so much me Thank

44:26

you for including there, but also thank

44:28

you so much for the other I'm

44:30

a big fan of peter book and

44:32

look forward to check me out the

44:34

thing like rocket scientist well. It's very,

44:36

very interesting. well as always, shamim uh

44:38

a pleasure it is to to with

44:40

you, to hear more about the evolution

44:42

of what nine years ago was a

44:44

remarkable organization, all the more so under

44:46

your study leadership at CarMax. Wonderful to

44:48

hear more about how the organization continues

44:50

to evolve, to innovate, as well as

44:52

some of the areas that you're most

44:54

excited about as you look to the

44:57

future. Thank you as always for a

44:59

stimulating conversation, Shamim. Thank you, Peter. Looking

45:01

to talking to you again soon.