Innovating Leadership:
Co-Creating Our Future

Resources:

To learn more about the Career Choice program Sandy discussed, check https://www.aboutamazon.com/news/workplace/career-choice-free-education-for-amazon-employees.

For more about the robotics developments Scott talked about, read the stories at https://www.aboutamazon.com/news/tag/robotics.

Our host Maureen Metcalf posts a newsletter every week on LinkedIn. You can subscribe here. 

Maureen’s latest book is Innovative Leadership & Followership in the Age of AI. You’ll find details about it at https://bit.ly/LeaderInAI, or check out the Kindle version at https://amzn.to/44buVz8. The audiobook version is now available at  https://amzn.to/4dTCleZ.

Her other 10 books are available on Amazon here.

Other episodes you’ll enjoy:

• With Great Success Comes Great Responsibility with Holly Sullivan
• Building the Systems (& People) That Sustain Growth with Carla Morelli
• AI, Your Story, & the Professional Identity Crisis with Christopher Washington

Guest(s):

Guest(s) Bio:

This episode brings together two leaders responsible for different (but deeply connected) parts of Amazon’s global operations. Sandy Gordon leads the employee experience systems that support more than one million frontline employees, while Scott Dresser leads the robotics and AI systems that power Amazon’s fulfillment network. Together, they offer complementary perspectives on a single leadership challenge: how organizations can scale technology without losing sight of the people who make that technology successful.

Sandy is Vice President of Global Operations Employee Experience & Relations at Amazon; her organization focuses on employee engagement, workplace experience, leadership communications, career development, and policies that help create a safe, inclusive, and high-performing work environment. She also oversees initiatives such as Amazon’s Career Choice program, which invests in employee education and workforce development.

Scott is Vice President of Amazon Robotics. His organization designs autonomous mobile robots, robotic manipulation systems, AI-driven automation, and intelligent warehouse technologies that improve safety, operational efficiency, and customer delivery performance. He has been instrumental in the development and deployment of technologies including Proteus, Sequoia, Sparrow, Robin, and Amazon’s next generation of AI-enabled robotics.

Transcript
(auto-generated)

Maureen: [00:00:00] Hi, I’m Maureen Metcalf. Welcome to Innovating Leadership: Co-creating Our Future. Prime Day is Amazon’s most visible test of technology and network design, but underneath it is an equally demanding test of the human system. 

Our guest today is Sandy Gordon, vice president of employee experience at Amazon, who leads the work of preparing that workforce. Sandy, welcome. 

Sandy: Well, hi, Maureen, and it’s wonderful to talk to you as well. And I’m looking forward to it.  

Maureen: Amazon’s supporting more than a million employees delivering this Prime Day. What does it actually take organizationally to bring a workforce of that size to peak readiness on a fixed day?  

Sandy: Leading the employee experience for the Amazon team is truly a privilege, and one that I think my 30-plus years in professional life have brought me to. 

It’s a space where it is all [00:01:00] about listening to our people and taking advantage of every day to build and earn their trust. We do that by truly engaging them in the work that’s in front of us. Some of our best ideas… You know, you hear about Amazon and the desire to be leading in innovation, and our ability to move at lightning speed and working our way back from the customer because we’re customer-obsessed. 

Well, all of that starts with creating a work environment that engages and inspires and energizes our people. And then our people help us be able to deliver to our end-state customers quite frankly. And so they identify the opportunities. They help us work through the challenges. 

And so yeah, it takes systems. It takes intention. We have to be intentional with our actions. We have to make sure that we are always looking for ways to continue to hear and act, uh, respond to the feedback we receive from them. And Prime Day it’s a moment for us where it pulls it all [00:02:00] together, and we get to celebrate through these four days with our people because that’s really what it is. If you go into the centers now, there’s a lot of activity happening, but it’s happening with a lot of energy, food, music… just kind of knowing that we’re putting a lot of smiles out there for customers.  

Maureen: I love the framing of we’re putting smiles out there for customers.  

Sandy: It is true. Here’s one piece that helps in that. Our Amazon employees all have free Prime membership. And so the smiles that they’re bringing to customers, they’re receiving it because they’re Prime members themselves. They too are partaking in all of the wonderful Prime deals that are out there. Part of our fun is even asking, “What’s in your basket? What’s in your cart?” to understand- … you know, what’s of interest to them. It, it tells you a lot about people, what’s going on when you hear that. So it’s an interesting way to think about it.  

Maureen: So Sandy, what’s in your cart?  

Sandy: So like I shouldn’t have brought that up. 

No, really. Uh, [00:03:00] I, I am, uh, when I look at my cart right now, I’ve been doing a lot of purchases for my father. I’m taking care of an aging parent and trying to help- 

Maureen: okay …  

Sandy: him maintain his independence. And so the things that are in there are everyday essentials that we need, and I need to make sure that he’s able to maintain that independence, either through groceries that can get there at lightning speed for the caregivers who are fixing a meal for the day or cleaning utensils. 

But the fun things (I think every cart should have something fun) is a tablet. He’s technology challenged and, um, we got the simplest tablet that we could find so that he can talk with his grandchildren and see his grandchildren. And they are looking forward to teaching him all about it this weekend. 

Um- So it’s gonna be a fun time. But that is really what I mean when I talk about the power of our employees being Prime members and understanding the connection that they make for employee understanding that. Because being on that side, of a customer, with my [00:04:00] Prime membership, I know how important it is for it to be reliable, and the speed that it gets there when I need it, we need it then. It’s just real relative to my own personal situation, and that’s magnified a million times over for our employee base and our customers. 

Maureen: And knowing that your dad will be cared for is one of those things that matter. 

Sandy: It does. It does. There are countless stories out there, where the Amazon experience has been fabulous. Someone will engage with a customer, see that a customer needs help, or picks up something in the lawn as they’re walking to the door… 

And so it’s really about when you asked about what it takes for a million plus people employed and how do you help make sure that they’re connected and a part of this, it really is about making sure they felt seen, they feel seen and heard, and you’re not lost in the sheer size of our organization. 

And so in each of our sites, we are deliberate about ensuring that there are mechanisms for our employees to [00:05:00] speak, to share. They are not only encouraged, but they’re empowered to take kind of control of, “Hey, this is going on in my site. We think this is a better way to do it. Here’s some information,” or, “This is something that’s not meeting our standards, not raising the bar for us because we’re constantly looking to evolve and improve.” 

And we take that feedback very seriously. We hear it, we act on it, and then we respond by telling all, “Hey, here’s what we hear.” The voice of our associates matters. And that’s a big part of how you connect to a million people and ensure that we’re all on the same place and focused on the same mission. 

Maureen: So let’s pivot slightly. I wanna talk about the billion-dollar commitment to Career Choice by 2030. ‘Cause I hear you saying people feel heard, and one of the things I think many people want, younger and older, is to own their career and their career success. [00:06:00] And making that investment part of the broader two and a half billion dollar Future Ready pledge to help 50 million people for the future of work. 

That’s more than just Amazon, people, unless you’re scaling up by a- … a factor of 50, and I’m, I know Bezos has big dreams, but- Yeah. Like- … the rest of you may not have that.  

Sandy: Got it. No, for certain. What we have, and we’ve made this commitment to $2.5 billion investment in upskilling. And that is by the year 2030, at least 50 million people. 

And you’re right, that goes far beyond our Amazon, employees, and reaches into the communities which we serve. With a billion dollars of that directly committed to Career Choice, which is a program we’ve had for quite some time. And in our Career Choice program one of the things I’m most proud of it is it removes one of the greatest barriers for people in education, and that is the fee, the cost. 

And so we offer [00:07:00] prepaid tuition for our employees. And that is everything from a GED to an undergraduate degree. And that allows them to pursue education that can help them either inside of our walls or, quite frankly, outside of our walls. You know? It’s, it’s nothing for us to encounter an employee who’s partaking in the Career Choice and may be in the medical field as a nurse. 

That’s a wonderful thing if that’s their passion and that’s their calling, and our ability to help that. But then equally that very same work can lead them to careers and further advancing their careers inside of Amazon. And then we In addition to the educational programs, we have robotics apprenticeship programs. You can use Career Choice for getting your certification in certain fields that are calling for that. That is all about allowing our people to advance and progress. And as you said it: the younger people are very much into it, but not only just your young people, right? 

And there are people who have hit another part in their life, and as technology continues to [00:08:00] advance and the world continues to change, there’s an interest and a curiosity at continuing to learn and grow and maybe take a change and turn their profession in a different way. This is what affords them to do that at Amazon. 

As it relates to the other $1.5 billion and reaching out into our communities, that really is about e- everything from your family members to local schools. So we think about elementary, middle, high schools introducing and beginning to build kind of this base knowledge and understanding around technology and how it can help you advance. 

it’s evolving and so, there’s never too soon to become aware and familiar with it and comfortable with it. And When you have that level of understanding then, um, makes it not so intimidating. You know, I’ve been in profession now 30 plus years, and I started my career on the front line in aviation. 

I think about the opportunities that are presenting themselves [00:09:00] now to our employees through these very clearly articulated pathways to progressing their career. I wish that had been as clear back for me. I’m fortunate I had a great career. I’ve had a great career. I’m continuing to have one. 

But it wasn’t as cleanly available. I had to work a little bit more for it, and I think that’s one of the things that I’d like to see us do is how do we make sure it’s so cool? If you’re interested, then this is the way, and here are some options for you to consider.  

Maureen: Your scale allows you to invest in people in ways that my little company d- I can’t do that. 

But you can and you do. ‘Cause there are a lot of people who can, and they don’t.  

Sandy: Well, we have a leadership principle… We have 16 of them, and one of them speaks to our responsibility because of our scale and scope. And quite frankly, it, it is a responsibility that we have. 

Um, you know, Amazon is so vast in all of the industries that we have a presence. And so there’s [00:10:00] tremendous opportunity for our people to have enriching experiences here at Amazon. And so shame on us if we aren’t thinking about how we can help not only our teammates, but how we can help, again, the communities that we’re in. 

We’re serving them, and so let’s serve them even further than just delivering packages, but let’s serve them in other ways that help them grow too.  

Maureen: Let’s talk about safety as a system, not as an initiative. So safety tends to get harder to hold steady when the volume gets the highest, right? 

You got more people, you got more robots, the stress level goes up.  

Sandy: Safety is first, end stop. There’s nothing else to say. And so it’s not acceptable to us that it erodes because of our scale. In fact, we are actually using that innovation in addressing our scale to improve it. 

So the robots that you talk about: when our tech team starts to think about robotics and technology, they start with the customer in mind [00:11:00] first and work backward from the customer. Now, if you’re thinking customer at the very end of this line who receives a package, that’s one thing. But our employees are the customers when we’re thinking about robotics inside of the sites. 

And so with that, they start and think about, “Well, what would help make the employee’s job more safe? How can we do this faster? And how can we do this smarter?” And it is with that in mind that we build and advance a lot of the robotics that’s in our sites. So think about those repetitive movements that today are required or were required of our employees in these sites to move packages from one place to another, the reach, the bend, the pull. 

 Think about lifting heavy items pulling, tugging. Then think about Proteus, who is our newest, autonomous robot that we have introduced into the sites right alongside our employees. So prior to this, most technology that you saw inside of the Amazon sites, there was a [00:12:00] fence and the robots moved withinside of this fence, and there were trained employees who would work and interface with them. With Proteus, Proteus is on the floor working right alongside our people. 

And here’s what it does. Proteus is able to lift up to 900 pounds. So if you think about all the packages that we might be crating for delivery to a particular, area a particular zip code, et cetera, it all goes into a, a cart that could weigh that much. And in the past, employees had to pull and push to get those carts positioned onto the dock, getting ready to be loaded. 

Now, Proteus comes through, picks it up, and carries it over to the area that it’s going to be dropped off, and that frees the employee from the risk of having any type of injury associated with that. And so since 2019, we’ve reduced our recordable injury rate by 43%. And in the last 15 years, this is how we’ve been [00:13:00] introducing these robots. 

For us, the scale is not increasing risk for injury, ’cause we’re absolutely partnering that with the technology being able to come in and address those types of things.  

Maureen: So does Proteus actually look like a humanoid robot? 

You’ve got a broad range of robots, everything that looks like a Roomba to just things that move up and down, and then you’ve got the humanoid ones.  

Sandy: I think Roomba’s a great way to think about what this looks like, right? Proteus is that version of our floor kind of moving robot. And then as it pulls underneath a cart, it has the extender that would allow it to lift up and then proceed on. 

One of the other things that the team has done that I think is really nice is they’ve made it where it interacts. Look, all the safety intelligence that we need for these robotics , it’s built into Proteus. 

That’s what makes this one so special and unique for us. And as such it takes very little training [00:14:00] required for our employees to be able to be in this space, and it alerts our employees to its presence. It kinda gives signals to where it’s gonna move, so you can understand and interpret what’s happening on the floor. 

And so, like I said, as a result of that it’s improved our spaces from a safety perspective.  

Maureen: You’re creating more than 4,000 new jobs at delivery stations in rural communities. 

What’s different about building an employee experience in rural delivery stations versus large urban fulfillment? I imagine there’s a lot of differences that need to be built into the system.  

Sandy: From an employee experience perspective, I’m not sure that there’s that much of a difference. We absolutely have to be creative about how we ensure that we’re able to meet our commitment of same-day and next-day delivery in some of these remote areas, right? 

And We don’t have a one-size-fits-all. That’s another way that we’re able to scale [00:15:00] so well is that you don’t go with the cookie cutter approach. You have to be able to meet the region and the customer needs, you know, as you find them. And so that’s there. But no, I think for the employee perspective, the great part about this is just honestly the fact that we have been able to introduce this service in areas that previously didn’t have access. 

And we’re able to do it with the kind of same level commitment that we’re giving to the vast majority of our more urban areas, and the speed that we’re able to give to them. And think about just that conversation I was having with you earlier about my dad that there’s somebody in a rural community who’s got the same challenge in front of them, and now this might bring a little ease and comfort to their ability to provide for their loved one and care for that. 

In addition to that we’re introducing jobs to these areas, and those jobs are direct to Amazon. But then there are lots of jobs that are not direct to Amazon, that are indirectly [00:16:00] created by our presence, including small and medium-sized businesses. 60% of the product that we offer in store belongs to small and medium-sized businesses. 

And so by using our logistics and our AWS platform, they’re able to have far greater reach outside of just their immediate community. That’s how I think about it. As far as it relates to interacting with our employees, we continue to have those same values and commitments that we do in other places to make sure that they are heard, the experience is something that they can feel proud of, and that it’s engaging. 

Maureen: I love the balance of I get to learn, I get paid a living wage, I have healthcare, and I do something I feel proud of.  

Sandy: Yeah. For sure. And you get healthcare on day one. We have a 401k program. , We’ve got the career choice. You know, the headline always seems to be about the technology, but the magic, quite frankly, [00:17:00] is our people. 

The stories, the exceptional stories about their care and consideration and the heroic efforts to make sure that our promise is met with our customers. It just never ceases to amaze me.  

Maureen: So one last question for you. For anyone listening who wants to take something away from this interview that they can do immediately. Many of our listeners don’t have a million employees. 

Some of them run smaller businesses, or they have 10,000 employees. What can people do immediately that leverages your lessons learned from Amazon?  

Sandy: Maureen, I think the best advice I could have for anyone who’s asking that question is, you have to get as close to where the work occurs to understand where the opportunity exists. 

As large as we are, nothing cuts through like the conversations that we have [00:18:00] inside, be they round tables to employees being able to simply write on a simple whiteboard, “Here is what happened to me today,” and our commitment to responding in 24 hours and really researching and looking that information up, to also the technology that we provide all of our associates with, through our A to Z Program, which allows them, if you can’t write it on the whiteboard and you’re on your way home and you wanna give us a message and send it to us, and that communication. 

Trying to filter out and make sure that the data that we’re seeing matches the anecdotes we hear from. And when they don’t, leaning in and trying to understand that so we are truly responding to the needs of those closest to the work. We’re listening to the creative ideas from those closest to the work. 

That is the best thing you can do. And I know for as advanced as we are and as large as we are, probably an expectation that this is gonna be something really unique and, you know, futuristic. But the truth of the matter is it’s employees working inside of a building [00:19:00] with a common goal. 

And the best thing you can do is make sure that they feel valued and invested in what you’re trying to do. And you can’t do that in some kind of really futuristic way. It is truly relationship oriented.  

Maureen: I love that. Thank you. So invest in their education. 

Invest in their safety. Give them a sense of purpose. Ask for their input, and respond to it respectfully.  

Sandy: Well said. Yes.  

Maureen: Sandy, thank you. Thank you. And I am looking forward to opening my Amazon packages tomorrow.  

Sandy: Well, we appreciate you being one of our customers- Oh, yeah … and 

and thank you for your time today. Truly appreciate us spending some time with you.  

Maureen: Thank you. Same [00:20:00] here. 

Once a year, Amazon’s entire operating system is tested in public, in real time, at scale. Almost no other organization even attempts something like this. 

Prime Day is not simply a sales event. It’s a four-day stress test of fulfillment, transportation, workforce systems, all running at peak load across every region the company serves. Scott Dresser, VP of Amazon Robotics, leads the technology behind the journey. 

So Scott, how do you and your team think about Prime Day actually testing inside Amazon’s operations?  

Gordon: Prime Day is a big day for us. Obviously, it’s a big week for us, actually multiple day event and the planning, the preparation that goes into that, you know, starts back in our other peak events like our holiday season. 

And if you look back to last year’s holiday event, in January, we all sat down and said, “What went well? What didn’t go well?” [00:21:00] And we worked backwards from there thinking about Prime Day as the next event that we have coming up. 

We make adjustments in our systems, in our robots, uh, make sure that we’ve got them deployed in the right places, that they’re working at the levels they need to, that our employees have the right training to know how to use them and we get ready for our next events. That involves everything from scaling our software systems to making sure that our supplies and spares for our robots are stocked at full so that when we hit our big events , that we’re ready for them.  

Maureen: Most people probably haven’t been inside an Amazon fulfillment center. 

Can you describe the range of robots? 

Gordon:  

Yeah. We’ve got a wide range of robots in our latest generation, 12th generation fulfillment operations. 

So we have the Roomba sized robots. We call them structured field mobile robots, and that’s a fancy way of saying a robot that can move a shelving unit around. And usually that shelving unit is full of inventory or products that our customers are going to use. 

That robot can lift and lower that shelf, and it can bring it over to an employee or in some cases [00:22:00] another robot to pick items off of that or stow items and replenish inventory into that pod. 

There are a number of other mobile robots. You mentioned one really important one, which is Proteus. That’s our autonomous mobile robot. That robot works in free space around people. People, think of our loading docks and areas where we’re moving pallets around, where we’re moving what we call a go cart, which is a fancy cart that contains customer orders. 

Proteus is tasked with taking those from our package sorters where we sort packages to go to different dock doors in our warehouses, and Proteus is tasked to pick up that go cart, bring it to a dock door where somebody will load it onto a tractor trailer. So those are movement robots, mobility robots, things that move goods around a warehouse. 

We have a second class of robots, that are manipulating robots. So these are the arm-based robots, and they can do things like pick up out of a cluttered scene maybe a customer order or an inventory item that you’ve ordered. They can look at it, identify it, and then sort it [00:23:00] and densely pack it into a container like a moving tote, so that it can be on its way for some other outbound task. 

And we have various– they’re all bird names. It’s a bird name theme. So we started with Robin. We have Cardinal. We have Sparrow now as well too. Those are all various different iterations of manipulating robots that can use different technologies, mainly suction, to pick up a package or an inventory item, identify it, and then sort it. 

We have hundreds, in some cases thousands of them in one warehouse. At the same time, we’re employing twenty-five hundred people, so there’s still a lot of people working in and around these robots. 

Maureen: So I’m assuming a robot doesn’t just one-to-one replace a person, especially these Hercules robots. You didn’t have humans carrying around shelves.  

Gordon: No, we didn’t.  

Maureen: How do you redesign the work? ‘Cause it seems like that’s a continual realignment every time you get these manipulating robots that you didn’t used to have. That changes what those 2,500 people do when they show up to work every day.  

Gordon: An interesting [00:24:00] example would be to think about Proteus and how did we get to the innovation that’s Proteus and the job that it does, ’cause if you think back a couple years ago when we were thinking about the challenges and the opportunities that we had, uh, inside our fulfillment operations to go after, we said, “You know what? There’s these 900-pound carts that we wheel around a fulfillment operation every day.” The application here was on a loading dock. 

You’ve got people coming in and out of tractor-trailers. You’ve got other equipment moving around that space, people doing other tasks that are in a very busy and congested area of the fulfillment operation. And we said, “What’s the hard problem that we have to go solve in order to develop a mobile robot that can actually work in that space?” 

And what we come back to a lot of times in robotics is safety. We have to make sure that we can develop a robot that can perceive its environment, including the people that are in that environment, in a way that is verifiably safe, and move in and around them. 

So that was the place that the innovation started, was how do we develop a robot that’s relatively low [00:25:00] cost, that can work around people, but also be useful, right? ‘Cause a robot that works around people but stops all the time and doesn’t get anything done isn’t a particularly useful robot either. 

It has to accomplish its task of moving a cart. So we sent a, this very small team off at the time to go off and innovate on that exact problem. 

We didn’t tell them exactly how to do it, but we knew we had a problem that we wanted them to solve. And they came back and iterated on it several times, and we liked some things, and we didn’t like some things, and then after a couple iterations, we got to something that we said, “You know, that looks a lot like a robot that we could build, and again, most importantly, we think it would be safe around people.” 

Not just be safe, but be accepted around people. It’s one thing to be safe, it’s another thing to be in people’s way or be clumsy or clunky. Uh, we also had to make sure that it understood people’s intent and communicated its intent well. 

So that’s where we started, and that was kind of the, the Proteus journey, and we iterated on that multiple times, and we ended up with the robot that we’ve scaled today. There was a bunch of work that we had to go do to make sure that people understood what we were trying to do, why we were trying to do it, demonstrating that Proteus was going [00:26:00] to be safe around them. 

We took the opportunity to retrain them and upskill them into some other tasks and other jobs around the fulfillment operation. And now we have Proteus in about 25 fulfillment operations across North America.  

Maureen: One of the recurring challenges boards and leadership teams face is moving from piloting to deployment to scaling, and you’re scaling across thousands of locations. You’re rolling it out in Europe soon. What has to be in place in the system to scale across thousands of sites and millions of employees and keep people feeling engaged and like they’re not threatened by this robot’s gonna knock me over or I’m gonna lose my job? 

Gordon: Yeah, Maureen, it’s a good question, and I think it’s one that is sort of the special sauce that we’ve worked on here at Amazon Robotics. We often talk a lot about that upfront innovation, that, that business challenge that we have and how we drive upfront innovation. 

And while that’s [00:27:00] necessary as part of the development process, it is not going to get you to the point that you mentioned earlier. And oftentimes, what is missed, I think, in that journey is the continuous innovation you have to have from that point on to getting a product to scale. 

One of the things I’ve learned by coming here at Amazon is the appreciation for what it means to get to true scale. We’ve deployed a million of these structured field robots. That is true scale, right? 

I mean, we have thousands of these across 300 fulfillment operations worldwide, and we learned a lot of lessons in that journey, and we applied some of those to Proteus. It cuts across so many different elements of the technology. It has to be reliable. A system that is going to run around people, if it’s broken all the time, isn’t gonna work very well, and it’s certainly not gonna be accepted by the operations team. So we spend a lot of time , in that second phase of development innovating on areas to make the system more reliable and our robots highly available. 

We run them through life tests. We expect them to last for 10 years. We test them that way, right? We make sure that they last through many, many cycles of lift and lowers and many, many miles, hundreds of [00:28:00] thousands of miles of travel on these systems. 

So as we’re innovating on the product side, we’re often involving our internal supply chain teams to make sure that the electronics, the hardware that we’re building our systems from are available, that we have multiple sources for them so we don’t run risk down the road there, that we have a scalable and manufacturable process. 

The robot has to be buildable. If we can’t build them easily and reliably, then we don’t have a product. And then finally, we have to make sure that we can deploy them in the field. Think about receiving a giant erector set in the field, and you have to assemble them from many thousands of pieces and it takes days to build them. 

That’s historically what most companies do to build them. And we said, “You know what? That’s how the industry does it, but that’s not gonna work for us when we build thousands of these things. So how do we think about the product design to make that easier?” And we asked the teams, ” What’s the right way to think about this?” 

And they said, “You know what? If I was buying a washer you’d wheel it into my house, drop it on the floor, level it, plug it in, and turn it on. Maybe you run through a test and make sure that it works, and that’s about what we would do.” We said, ” that’s a good North Star. We’re gonna use that.” 

We wanna actually [00:29:00] deploy them as an appliance. And we’re not quite there yet, but we deploy them in three big modules that seamlessly come together in the field in a few minutes. And then in about an hour or two, we can plug them into the infrastructure that they need, test them, and they’re up and running, right? 

So those are the kinds of things that, to scale at the level we’re talking about, thousands of robots, across our worldwide operations, you have to do to make them successful. Otherwise, you just don’t get the benefits of that early on innovation that we talked about. 

Maureen: Then there’s the human-robot interface too- lots of humans need to make sense of what is changing, adapt their own behavior, and hopefully are excited about it, not resentful and angry. ‘Cause we know humans if not clearly understanding the value proposition can be difficult.  

Gordon: Yes. That’s 100% the case. And, Proteus is a good example of that, I think, where we brought Proteus in to do a job that people still do today, but to a lesser extent. That is an important part of our employee [00:30:00] communications and setting up our expectations to let them understand why we believe this is good for them. 

First and foremost, safety, better work environment, making sure they aren’t moving 900-pound carts around an operation which we think there are better opportunities for our employees to take advantage of. And then we go through a bunch of really interesting programs. 

We’ve got Career Choice. Another one that is really near and dear to my heart, which is more of an on-the-job training program, is our robotics apprenticeship program. If you think about a robot sort of like your car, well things break. 

You need preventative maintenance. You gotta have a brake job every now and then. Maybe the wheels need to be replaced. We need trained and skilled, folks in our operations who can do those jobs. And I was just actually just talking to a woman today who had gone through that program, and she is now a certified robotics maintenance technician, and that is what her new job is. 

She, she was an operator on the floor, and now what she does is she works in our maintenance areas, and when robots come in and they need preventative maintenance, she can clean the cameras, change the wheels. There are some other wear items that we expect to wear out over time on [00:31:00] our robots, and she can replace those. 

And if there are more complex issues, she’s built some of the problem-solving skills to help us diagnose those failures, and report them back to our teams in case we need to fix them. We believe that while robots are changing the nature of work, there’s great opportunity here to upskill and provide a range of new roles to people. 

Maureen: How do you design robots and the overall delivery system so that going faster doesn’t come at the cost of safety or sustainability? 

Gordon: Yeah, I think that’s one of the things that we’ve really doubled down on in the last few years because we also have found that speed and sustainability in particular often go hand in hand. 

If you think about your experience on amazon.com, customers love speed. They like to be able to click on something and have it show up in two hours. It’s an insatiable appetite for faster products and oftentimes it’s a function of what you’re ordering, right? 

If you’re ordering food or groceries or everyday essentials, sometimes you need that stuff faster. So there’s always these cases where people need things faster. So we’re always thinking about how do we make that [00:32:00] experience better. The only way we can get that to you in 30 minutes is if that inventory is relatively close to your house, right? 

If you live someplace and it’s gonna take me an hour to drive it there, we’re probably not gonna get it there in 30 minutes, right? That ability to place inventory closer to our customers is really important. It drives speed, but you can also imagine it drives sustainability: less transportation costs, less fuel costs, less energy in getting that to your doorstep. 

So that is a place we have spent a lot of time thinking about how can robotics enable that? How can the systems that we build on top of robotics enable that? Our storage systems, so the ones that have our customer inventory in them, are much more dense than our manual buildings, right? 

If you think of a, a robotic storage field, it can be built very densely because the robotics, they actually travel underneath the shelving pods that we were talking about. We don’t need the big aisleways. We can build them all the way up to the ceiling, or we can take advantage of the three dimensions in a building and make sure they’re built very high. 

We can use robotics to pull those down. And y-you ask, “Okay, why is that important?” Well, [00:33:00] dense buildings means we can put more inventory in spaces that are smaller, that are closer to our customers, that typically are in suburban or urban areas, and that, that’s great for some of these smaller buildings , smaller from a square footage perspective, but we can still provide the product density and the product selection that our customers love for a lot of the products, maybe not all of them, but the ones that you’re gonna order most commonly really close to your house. 

So  

Maureen: I’m imagining all of the modeling that has to go into designing where a fulfillment center goes, what’s the shape of the building, what configuration of robots, what kind of humans, what products are ordered with what frequency. 

These are just incredibly complex math problems. 

Gordon: Yeah, I can, I could talk a little bit about that. Understanding where we have customer demand, and where building placement needs to happen, that is, like you said, a challenging problem. 

Once we’ve built the network and deployed the buildings, how we decide what inventory goes where, so when it comes [00:34:00] into our network, how we distribute it, those are all problems today that we’re using generative AI in some form or another to go help us with. 

This is a very, very rich area for artificial intelligence, and particularly with the advances in generative AI over the last several years, the ability to optimize some of those things and get that placement challenge right, is an important part of that.  

Maureen: It seems like the real achievement isn’t any one piece of technology, but as you’ve said the ability to leverage all of these tools and humans together to deliver my eco-friendly bug stuff tomorrow before my trip.  

Gordon: We’ll make sure the robots are getting it there. But there are single robots that have value as sort of standalone systems but they are more the thing of the past than the thing of the future. 

I often tell people if I had the opportunity to rename the organization, I’d call it Amazon Robotic Systems, not Amazon Robotics, because while robots are a really important part of what we do, the systems that [00:35:00] they are part of are the thing that delivers value. 

I don’t think of my job as delivering a robot to my operators who are running the buildings. I think about delivering a system that usually is comprised of automation and robotics, but includes a lot of other things. It includes safety systems. It includes software. 

It includes workstations where people can work. It includes a network and IT infrastructure that, that needs to be available in order to make it work. And all of those things have to work together in order for our operators to be successful. The robots are the ones that are the most visible and front and center, and people are very enamored with them. 

But it is all the pieces that have to work together in order to deliver product value and value back to the business and solve that scale problem that we talked about earlier. So you’re spot on.  

Maureen: And we talk about even leadership as a system. There were humans, but there are also governance processes. 

There’s decision velocity. There are all kinds of things that in a bad system, smart people can behave badly.  

Gordon: Yes. 

Maureen: What’s one leadership [00:36:00] choice you believe senior leaders or boards should make with regard to transitioning from piloting to scale to delivering value? So how do you make it practical for anyone who’s listened to this conversation?  

Gordon: It’s a really tough question. 

I think it goes back to what, what we talked about earlier is going back and making sure… I see a lot of robotic systems out there today, and I see a lot of robotic systems that don’t succeed, and I think they don’t succeed because they don’t understand the hard challenge that they have to solve. 

Oftentimes, that is the failure to understand how robots and people need to work together, and oftentimes that comes back to safety and making sure that you’re safe around people. We obviously talk to a lot of startups and new technology companies, and we’re interested in the technologies, and the one thing that I commonly see missed in, on all of this is people are very focused on the technology, and some of it’s brilliant, and some of it is doing some amazing [00:37:00] things. 

But then when it gets to, well, how do people work in and around that? The conversation comes to a quick end because teams haven’t thought through what that looks like. If you don’t think about that from the get-go, you’re not gonna get a successful result in the end because it’s very hard to build that in once you’ve got an initial product. 

Maureen: And my assumption is Amazon from hire date selects people who are gonna be comfortable working with technology. So not only do I work with a gen one robot, but I can also work with a gen 12 robot. And when you get to gen 20, I will have updated my behavior to evolve along with the technology? 

Gordon: Yeah, certainly we expect to have and build teams who have leaders and leadership that understand what that journey looks like. I am constantly in the process of trying to figure out how I outdate the technology that I just deployed, right? I mean, that is the constant state that we are in at Amazon, and we talk a lot about that, being willing to disrupt your own ideas and to put new product [00:38:00] out that is going to be even better than what you have out there, because if you don’t, someone will. 

And I think we’ve been exceptionally good at that, right? We have a lot of technology, but we’ve got a lot of technology that’s been out there in older generation buildings, and we’ve got new stuff coming down the road all the time, and you have to have leaders who are willing to take that risk and not get too married to the technology that they built in the last iteration because something new is gonna come down the road, and that’s the opportunity, and there’s just so much opportunity here to do that. 

I think folks get that at Amazon.  

Maureen: Scott, thank you. This has been really interesting to hear how you think about robotics as part of the overall system. 

It’s not like you get to be done because you introduced the new Proteus and tested it today.  

Gordon: Yeah. I w- I sometimes I wish it was, but that’s also where the fun in the job is as well too, is getting an opportunity to get out there, talk to our employees, and find out what’s gonna make it better and do another iteration. 

Maureen: Well, you know, one of the measures of success is on this busy day you’re able to take time to talk to me. [00:39:00]  

Gordon: I appreciate you saying that actually. 

And that is, that is very true. I was just talking to one of, uh, one of my, one of my employees earlier today, and I was relating story to him about my very first holiday peak at Amazon. We had just deployed our first robotic system at Amazon, out at a small warehouse on the West Coast. 

It was peak holiday season, and we’re very excited that we were able to integrate and innovate our technology and integrate it into Amazon. Uh, and I happened to be on a business trip out there at that time, and the warehouse broke badly, and it broke badly because of our technology. And I spent three days in the warehouse. 

I spent one night overnight there and, and a lot of sleepless nights debugging and trying to figure out what the problem was, and we did finally resolve it. But it taught me that at scale, those types of things matter. The ability to diagnose, recover quickly, get the system back up and running, so that our operators can continue to work, is one of those tough lessons that Amazon taught me early on and has fared very well. 

And maybe is the reason I’m actually s- able to sit here with you guys today because we’ve taken a [00:40:00] lot of those early lessons to heart, and made sure that our systems are supportable by our field teams and very resilient to the types of different failures that they can see. 

Maureen: It’s a brilliant example ’cause I’ve worked with technology rollouts, and boy, the amount of sleepless nights people have, not only on the first site, but 10 or 20 or 50 sites in. 

Gordon: Yes. Right. And it’s not to say we’re perfect, but I will tell you one of the things we’re really good at doing is we’re really good at finding problems and healing. 

Our systems , we’ve designed them to be resilient, and boy, that investment has paid off, many, many times over, and that’s one of the reasons why we can continue to run this large network.  

Maureen: So Scott, again, thank you. Thank you for being a an exceptional example of how these systems can work at scale. 

Gordon: Thanks, Maureen. Thanks for having me here.