Host Mike Graen sits down with Sarjoun Skaff - CEO of Bossa Nova Robotics - to disucss the story of how the Bossa Nova Shelf Scanning Robot came to be, along with topics including:

• Challenges of implementing robots in retail stores.
• Using robots in retail, focusing on stocking and customer service.
• Supply chain automation and integration.

Got it. Got it. But talk to us a capital versus an

OP X, I would think most retailers are in the business of

selling stuff, not maintaining robots, not doing software

updates on robots, not putting guidances and all that kind of

stuff. So isn't it more attractive to most retailers to

say I just want to buy this is a service?

Yes, I'm sure that yes, this is the case, in

particular, for mid market companies who have less of an

experience doing that. So in that case, it probably makes

sense to do less, but you could offer support, either yourself,

or through a third party that you train. And that industry,

the retail industry is set up like like that, for instance,

the checkout counters, do get serviced by third party service

providers. So and those service providers are getting more

sophisticated and are servicing robots. So that's another that's

another option. And the developer - ourselves -

developer of robots, you're still on the hook for

availability. So you need to man the the call center the

operations room, make sure when there is a problem, that you are

the first line of defense, you are triaging problems, you are

dispatching a solute servicing, field servicing, for instance,

you are escalating to engineering for debugging.

You're not off the hook. It's not a fire and forget type of

sale. But the point was not to tie your capital with hardware.

A startup needs every dollar to pay people. And that's the

better way if you can pull it off. It's a much better way of

doing.

It makes sense. It makes sense. So as as with

anything, let me just kind of two parallel paths. You started

in this space? You said in 2005? Roughly, yeah, yes. So here's

what's interesting. I'll just throw this out here. In 2005,

you started exploring RFID at retail or I'm sorry?

Sorry. 2013 was the retail starts?

Yeah, but Okay, cool. So 2013. In 2005, I got

involved with RFID at retail. And we started and then stopped

and then started and then stopped and started to stop

about five or six times with some of the retailers out there.

For different reasons. You know, there were some reasons as the

costs were too high, this technology didn't work, right.

It wasn't ready yet, etc. Some was exactly what you said it was

really cool technology looking for a business problem. Instead

of here's the business problem here. What's this technology.

But now clearly RFID is mainstream. I think this year,

we did over 35 million tags, and I think the glide path is going

to continue to explode is become the new technology platform, if

you will, especially for items such as apparel and general

merchandise and things like that. Here's the question. Why

don't we since you started in 2013, why don't we have robots

in every one of the stores in the US what, what had been some

of the major challenges because because number one, they don't

get tired. Number two, they can do multiple things. They can do

multiple functions in a store, not just one. So you are sort of

sweating the asset. It seems like the hourly rate is

certainly very reasonable compared to a human doing it.

We're not displacing people, because we have all these other

tasks that we need people to do that we don't want to have them

take time doing those robots. Why are we have robots

everywhere?

I was in the business of answering this

question. I ... It's such a difficult answer. Your

experience with RFID showed a 15 year sales. True Do you know any

startup that can sustain such a sales cycle?ke

Not at all? Not at all.

It's kiss of death? Can't do it. Time is

money, you can't do it. So it's the incumbents, the big service

providers who can sustain it from other from other revenue

and maybe it may be innovate on the side. There may be a more

fundamental characteristic of the retail industry. It's just

my personal opinion, the one that's maybe jaded from my

experience, but ... I see as a handful of innovators in the

industry. And by and large, and everybody else are slow

followers, not even fast followers, slow followers. But

by and large, my view is that read it are our operators, some

of it? Operator then merchants, right supply chain operators and

product merchants. Yep. You're not technologists, right? And

you keep highlighting on your show. And research continues to

prove how big of a pain point out of stock is up. And after 10

years of trying to do it firsthand, I am not so sure they

are that interested in solving it?

Could be. Could be.

Could be. It's quantified. They know it. They

have a way of managing it that mostly works right now. And if

it's working, if it's not broken, don't fix it type of

thing. You can make it better. Yes. You can make more money.

Yes. Oh, there is another problem. Sorry, is proving that

you can sell more with the technology is very hard.

Hmm. Interesting.

Very hard. It's very easy. If you replace labor.

That's a very easy calculation for any CFO. The upside, the top

line impact is very hard to calculate. Even though it's real

interest, let's say you introduce a technology that

should increase on shelter availability, and therefore

sales. You try it out? You see an increase in sales, right? How

do you isolate that increase to that one variable that is there

all but it could be a seasonal seasonality difference. It could

be a fluctuation in the weather, it could be so many different

parameters. That calculation, I have viewed to be similar to the

way drug discovery is proven. It's just very, very hard. It's

a statistic. It's an exercise in statistics. That actually takes

scale. You can't do it in a small sample. IQ scale, and it

takes art. And proving that is a big barrier to adoption.

Yep. Great point. I think you're right. I think I'm

not trying this class. Because there's no such thing as a 'a

retailer.' They're all different. Every store is

different. Every product is different and center. But in

general, there are operators, right? We put stuff on the

shelf, and we take the customers money. That's what we do. Right?

And I think to your point, industry has struggled at let's

take the RFID RFID. For years has been it's just a technology

thing. While we're improving your on hand accuracy from 60%.

Up to 95%. You have a so what that's it that's it help. Well,

now we're starting to tie when you improve your on hand

accuracy and heard from a retailer the other day, every

10% of improvement in inventory, accuracy gives you 1% of sales.

So if I'm if I like 50% on any inaccuracy and I go to 100, or

close to 100, that's it's a five point improvement is fails.

There they are storytellers are starting to tie those together.

For every time I don't have product on the shelf. This is

how I disappoint the customer. Right? We're getting there, sir

June and the way you test that is not in a bunch of stores, you

do a scientific test versus control. We're going to do this

in 30 stores. We're going to not do it in 30 stores. And we're

going to measure pre versus post longitudinal the study of here's

the improvement and on shelf availability. And here's the

here's the improvement in sales. Here's the deal. Your technology

is not the solution. Right? Just like RFID is not the solution.

It's a work process that says we're going to do everything we

can to make sure stuff's on the on the shelf for the customer to

buy. We think they will buy more. We think they will be

loyal and come back just like the retail you fired. You get

this bad enough people just say forget it. I'm not coming back.

I got no time to sit around and go to a store where they have

half the stuff not there. I think you're in I think your

intervention was just too early. People are starting it that

that's why RFID is exploding, not because oh well now's now

it's cool before one cold no now people are seeing what the

results of that were so I I believe you were just too early

in the process because let me just call them out. Brain Simbi

and badger are all selling robots and being very, very

successful in engaging retailers and suppliers with this stuff.

To let them know about in store conditions. It's have never been

available before.

Nothing would make me happier than than their

success. And I'm friends with I'm friends with all of them. It

would be wonderful for the industry for robotics for

technology for the customers as well. I think you're right,

there is no single panacea. I heard this the phrase, it's a

multi signal approach, correct. Again, you should throw the

kitchen kitchen sink at the problem, the sales velocity

images that you can get from professional shoppers that are

fulfilling your online order. Robots, if you can cameras, if

you install, all of that helps move the needle.

Yeah, but for example, let's just take a robot

for a second, if I run a robot by a pile of clothes, your

computer vision will never be good enough to go 30 feet 32

It's never gonna happen, right? Try and take and put an RFID tag

on a watermelon, good luck. It's the good luck read it

technically, or try and put it on Pop Tarts, where you're

adding up to four and five cents for Pop Tarts. When I don't

think I need to know that level of accuracy. There's not a value

proposition stuff turned so fast. I shouldn't, I should. By

the way, there's other categories, let's say milk and

bread and yogurt, etc, that even a robot scanning a couple times

a day is not frequently enough, because I gotta be in stock all

the time, I can't be out of stock for a nanosecond, we'll

fix cameras make a lot of sense for that. So a retailer that

figures out this multi sensor platform and figure out which

one of these roles play. But they got to start with a premise

that says we're disappointing our customers, they're going to

start firing us or going spending their money other

places, if we don't have product on the shelf, and it's gonna be

a multi sectoral approach, it is not going to be a single single

single It can't be. And that's, that's hard, right? Because it's

like a big, great big science experiment. So where do you

think the future is? I mean, obviously, you're saying you

know, you're you are currently involved, or at least friends

with all the folks who are doing this, nothing would make you

here. So what do you see is the vision for those? Where's this

thing gonna go?

Yeah, I honestly hope an early adopter scales.

It's the way Walmart did with us. And that will signal the

change to the industry. I do believe others will follow if

that does happen. But somebody has to take the plunge.

So what are the technical ... so that's the

business part, we got to have retailers adopt us and take it

and etc. What are the technical challenges? So we didn't get

into the specifics of how the robot works, just to be

experiment. But this thing is extremely safe. It doesn't run

into customers, it avoids obstacles that does all because

is there anything technically, that still is on your kind of

innovation pipeline, other than RFID and other additional

sensors? Is there anything else that we need to do? That would

make it more reasonable for retail environment?

I think that technology is mature from a

robotics perspective today, we had to build it all ourselves

because we couldn't buy it today. You could correct

actually, there are platforms, their software is mature, and

they are service providers that will do the fleet management

that will help you can you can avoid building you could buy it

you could buy your way to the market doing it and focus on the

business or on the date on the on the sensing aspect. You could

focus on perception. And, and that's great. I think the

technology has a long and promising potential in front of

long future in front of it and promising one which is so far

we've we've we've we've limited ourselves to providing insights

that humans are doing the manipulation. As as, as the

robotics develops, manipulation, I think will be the next

frontier and helping humans by by doing the restocking tasks,

in increasing levels of complexity would would I think

the absolutely fantastic.

Let me make sure I understand, are you saying

robots potentially could stock product in the future?

Yeah, stock and pick? They're already picking

in, in, in, in the supply chain.

Right? Interesting.

And they pick from the shelf.

So here's the big goose egg question. Are you

really trying to put people out of jobs are June? Is that what

your goal is to make robots become replacements for humans

because that's somebody's thinking that already we might

as well throw it out there.

Even though humans put me out of job I'm not

looking for revenge. No, I don't see a scenario with robotics

particular in particular, where they are going to take jobs.

There are such a lack of manpower that and the technology

is is so, nascent in particular, when it comes with manipulating

the environment that we are going to be as roboticist

building tools that empower and, and augment humans for a very,

very long time before we get to replacing them. There is,

especially in retail, where you have such a big revolving door

of people that that come and work and then leave and then

come back. You're just not taking jobs away. It's not true.

Yep, you're freeing up autonomous or monotonous -

sorry - monotonous tasks, letting robots do that you still

need people. Guess what, I don't know that a, a robot will ever

replace a human when it comes to working with a customer to solve

a problem in the store. I just ... can you show me where the

the dawn soap is? Yeah, you might be able to guide me there.

But I'm not sure really answering and having a

discussion about that is something but picking and

stocking and scanning for outs and all that kind of stuff.

Here's the other environment. And this is the game changer I

haven't figured out how to break yet. But there's an opportunity

out there, which is the retailer is not only the one who has a

vested interest to make sure the stuffs on the shelf; the CPG

suppliers - Procter and Gamble, Kimberly Clark - have a very

strong interest to do that. And they get that data and insights

from a third party provider like an Acosta or Anderson, who goes

in and does surveys and takes pictures. But why do we have

them doing that? Let's tell him ahead of time you're going into

the store, we scanned it an hour ago, and these are the five

things you got to go fix let's, let's use AI and technology to

enable them to get more work done. Rather than data

collected, it's just doesn't ...

There are also things humans cannot do, for

example, looking at the shelf and figuring out which products

are misplaced, is completely hard. Whereas with a picture,

you could completely map out what should go where and have a

have a guide, provide guidance as to how to fix things.

100 percent. Well, strategy and I want to kind of

wrap up this conversation with a question. This one to me is one,

I always ask people and it's interesting where people take

it. And I've asked you a lot of questions. What is on your mind

that I didn't ask; what questions should I have asked

that I didn't ask? Because I'd be very interested to know, man

what's going on the top of that brain that I never kind of

unpack? What's your thoughts in terms of what I should have

asked?

What I'm currently thinking about. In

various domains is integration. We've worked, we've talked about

our points solutions. But in reality, for a product to go

from the factory, to your home, is a long supply chain, wherever

they are points of automation that are happening along the

way. But they are typically done from independent companies and

with little talking in between. And therefore it's it's it's a

high friction environment to automate more and more. But if

there is a way to integrate throughout the supply chain,

make it make solutions work well together, and you optimize the

integrated whole wrapped with very good software that does it

all. I think that would be a great service to the entire

industry. And I haven't seen it yet, but I hope it comes to

bear.

Well, let me throw one more complicating factor,

which is actually in the enabler to that. Have you ever heard of

a serialized g 10 before?

Yes.

Okay. So we usually think about UPC and I got 10. In

the serialized world I've got UPC one serial number one, UPC

one, serial number two, I've got unique serial numbers for every

selling unit. So I've got I've got 10 of these cups. Each one

of them have the same UPC, but I have a unique serial number with

them. I have the ability - and the future is going to be for me

getting item serialization throughout the industry. So I

literally can say, I just received a bunch of shirts from

Nike, and that says that Nike on them, but they're not Nike

product, because here's all the unique serial numbers that are

valid. So these are counterfeit, or claims, and I think you're

right. The data throughout the supply chain from the time a

tree is cut down until a mother is diapering their child in

their home. That supply chain flow can only be automated if we

share data back throughout the supply chain. But that is the

hardest part because the retailer their systems and the

distribution people have their systems and the suppliers have

their systems and raw material have their systems. There are

systems in place that allow us to share that information, but

we really aren't doing it yet. And that's, that's a big

opportunity. So...

We scratched the surface a little bit when we

were doing robot robotics, our robots, our shelf scanning

robots were getting in the same aisle as the floor scrubbing

robots. And really, with good integration, that should not

happen. Just one example?

Great point. Great point. Sarjoun, I can't tell you

thank you enough. You've been a friend and a person who I've

worked with, frankly, every time I talk to you, my head hurts

because you're so smart at this stuff, you know, this stuff? I

can, I really hope it will put your LinkedIn profile into this

bio, so people can reach out to you if they have any questions.

But I hope you stay in this industry. Don't give up on this.

Because I think if I know anything, knowing what we've

done in RFID for literally 20 years and how we're starting to

get the traction we need. I think you were just too early. I

think you were an early person out there. And everybody had a

skeptical reason why not, et cetera. I think there's going to

be a light at the end of the tunnel where I don't know that

we'll have robots in every store. But the innovative

retailers that want to be out five years from now, but 50

years from now, are going to have to invest in this to figure

out how to get out A) mechanized work that's monotonous and

tedious. And you can't get people to do or do very well.

But B) just make sure that they've got stuff available on

the shelf for customers. That's that, to me, is the big part. So

thank you very, very much for your time and ...

Mike, you've been an incredible support through

the years. You've been selfless as well. I have to say to

everyone, you're, you're you're genuine and I count you as one

of my friends. Thank you.

Very good. Take care.