Topics
Artificial Intelligence and Business Strategy
In collaboration with
BCG
In collaboration with
EBay is familiar as an e-commerce site that facilitates transactions between buyers and sellers. But as eBay’s first chief AI officer, Nitzan Mekel-Bobrov is focused on the role artificial intelligence technology can play in enhancing the user experience for everyone who engages with the platform.
In this episode of the Me, Myself, and AI podcast, Nitzan shares examples of the AI tools eBay is building, such as a 3D visualization tool for sellers create their own models, and intent detection tools to enhance customer service. He also discusses his academic background in biology and neuroscience, his purposeful progression from health care to financial services to online travel and finally to e-commerce, and the challenges of scaling up AI capabilities organizationwide to drive transformational value.
Read more about our show and follow along with the series at https://dev03.mitsmr.io/aipodcast.
Subscribe to Me, Myself, and AI on Apple Podcasts, Spotify, or Google Podcasts.
Give your feedback in this two-question survey.
Transcript
Sam Ransbotham: Launching a new AI initiative is quite different from launching a new technology platform like an ERP. Find out the key differences on today’s episode.
Nitzan Mekel-Bobrov: I’m Nitzan Mekel-Bobrov from eBay, and you’re listening to Me, Myself, and AI.
Sam Ransbotham: Welcome to Me, Myself, and AI, a podcast on artificial intelligence in business. Each episode, we introduce you to someone innovating with AI. I’m Sam Ransbotham, professor of information systems at Boston College. I’m also the guest editor for the AI and Business Strategy Big Ideas program at MIT Sloan Management Review.
Shervin Khodabandeh: And I’m Shervin Khodabandeh, senior partner with BCG, and I also colead BCG’s AI practice in North America. Together, MIT SMR and BCG have been researching AI for six years now, interviewing hundreds of practitioners and surveying thousands of companies on what it takes to build and deploy and scale AI capabilities across the organization and really transform the way organizations operate.
Sam Ransbotham: Shervin and I are excited today to be talking with Nitzan Mekel-Bobrov, the chief AI officer at eBay. Nitzan, thanks for taking the time to talk with us. Welcome.
Nitzan Mekel-Bobrov: Thank you so much. I’m excited to be here.
Sam Ransbotham: You’ve got a relatively new position, within the last year or so. Can you tell us what your role is at eBay?
Nitzan Mekel-Bobrov: Sure. Before me, my predecessors, so to speak, have all been chief AI scientists, and the change to a chief AI officer was actually a strategic one, with the recognition that AI is more than just the machine learning models — that AI is the engineering that it takes to “productionize” those models at scale and, of course, the business impact and business use cases. We really think of AI as an end-to-end experience.
Sam Ransbotham: All right; so eBay’s pretty excited about this. What is eBay hoping to gain from this?
Nitzan Mekel-Bobrov: I think [it’s] opportunities in every facet of our business, as probably most companies our size would say as well. I think, for us, probably what we’re most excited about, and the reason that I joined eBay late last year — what made me excited — is the ability to create AI-led tools, essentially, that we put into the hands of our customers, both buyers and sellers, to create their own experiences that they share with each other. It’s not just about us using AI to build things, but rather us building AI tooling to enable our buyers and sellers to build things, and I think that’s really exciting.
Sam Ransbotham: Give us an example of one of those.
Nitzan Mekel-Bobrov: Most recently, we rolled out our new 3D visualization experience, which is using computer vision on the back end to do the processing and some of the rendering. It’s not about us using this technology to create 3D visualizations, but rather, we’re enabling, through our mobile app, our sellers to create visualizations of their own items at scale, and in a sort of super easy way that doesn’t require professional equipment.
Sam Ransbotham: What do you think that eBay’s doing uniquely in artificial intelligence?
Nitzan Mekel-Bobrov: I’ll start with the approach, and then I’ll get into an example or two. From an approach perspective, I think we’re going at it in a unique way because we’re a two-sided marketplace. We worry about buyers and sellers, and so much of our attention is on building capabilities for our customers to use versus building experiences ourselves directly. Building tools for our customers to build experiences I do think is unique.
It’s unique also from a technology perspective because there’s a different level of resilience that’s needed, and you have to test a far greater number of ways in which it can fail when you’re not actually building the experience yourself; you’re putting tools in the hands of others to do. So it’s almost like we are software as a service within an e-commerce company, right? So that’s sort of one aspect of it.
In terms of specific areas of focus that are maybe unique, we are [doubling down] now on visual experiences. We can say “computer vision,” but I think of it a little bit more broadly than computer vision as an AI approach. I’m thinking of it as intelligent visual experiences that are immersive, interactive, adaptive, etc. 3D is sort of the tip of the iceberg, but as we get deeper into the “metaverse” and deeper and deeper into ways in which our digital platform is more than just a place to do the commerce itself, we think that visual experiences in real time, live interaction between people, [and] ways of visualizing products that feel like they’re in your hands versus just being on the screen is something that will be transformational for eBay.
Sam Ransbotham: You’ve got a very interesting background: Bookings.com, Capital One, Hearst, Boston Scientific. Maybe you can tell us a little bit about how you ended up where you are. And one of the things that I want to harp on, perhaps, is that a lot of our guests tend to focus on the artificial part of artificial intelligence, but your background is actually in the intelligence part, like the actual human intelligence, with your research on brains and how brains evolve, so I think that’s a fascinating, different angle. You’re coming much more from the intelligence part than from the technology part. How did you get to where you are, and how did you learn these lessons?
Nitzan Mekel-Bobrov: It’s interesting. I actually was planning on being a lab biologist or geneticist. The problem was that I was terrible at the bench. Anything that I needed to use my hands for never worked. None of my experiments worked; everything was a flop. So I quickly learned, “Well, I’m good with coding and computer algorithms; maybe I’ll focus on more theoretical aspects of biology.” And then I got immersed in the world of neuroscience and computational genomics, and that’s, by the way, how I got introduced to neural networks. It’s interesting because me and all my peers, that’s how we entered what ended up being machine learning applications in business, and I had no awareness that engineering, essentially, and computer science were playing such a big role on the other side of a somewhat academic fence.
To me, it seemed like a very obvious progression. As I was working on modeling how real brains, so to speak, work and how human intelligence works, moving from there into artificial intelligence seemed like a very natural progression, and I wasn’t alone in this. But then I got introduced to this whole other set of folks coming at it from the other end. … The progression — you’re seeing me jumping around from one industry to another — isn’t accidental. I spent the first half of my career in health care because it was very natural. I did my graduate work in, essentially, machine learning methods for genomic analysis, because it was the human genome era, etc., and so I stayed in health care. But at some point, I really wanted to see how AI can be used in other industries, and so I fairly purposefully moved from financial services to online travel to e-commerce.
Shervin Khodabandeh: As you have traversed the wide array of sectors and industries, what did you find was one of the biggest hurdles in getting AI at scale in these organizations?
Nitzan Mekel-Bobrov: Always people; they always get in the way.
Sam Ransbotham: [Laughs]. Spoken like a true engineer.
Nitzan Mekel-Bobrov: It’s hard to get a whole group of people with different incentives to coordinate together in a way that is needed. Doing AI at scale, and in a way that could drive transformational value, does require a much broader set of players playing nicely together. And while typically everyone is on board that it’s the right answer, the prioritization of that versus the very immediate-term business objectives is what typically ends up faltering.
Sam Ransbotham: Part of your background that I was interested in is that one of your dissertation findings was that the human brain is still evolving and getting smarter. I think you may be in a unique position to compare … but I assume that artificial intelligence is also getting smarter. Tell us a little about what you see happening in terms of these rates. Are the machines getting smarter faster than humans are getting smarter, or are humans still continuing to outpace? I’m kind of curious what your perspective is on those two things.
Nitzan Mekel-Bobrov: You really dusted off some old things when you were looking at my background. I appreciate you digging.
I think there’s two things I would pick up there that I do find interesting on a sort of philosophical level almost. The first one is that when we look back in history, both recent and longer term, we look at these macro changes, and then we assume that because those happen over long periods of time, that they’re almost episodic — like it’s not something that you would observe day to day. But as any geneticist would tell you, evolution is just population genetics happening over a longer time scale. It’s not as if it’s something that is episodic; it’s something that’s actually continuous. In that sense, I think it’s not surprising that humans have continued to evolve and are continuing to evolve. It’s just the nature of biology.
What’s happened in the most recent history of the human species is that … the variation in signals and the speed of that variation being introduced is just accelerating massively, and because of that, we’re able to adapt faster and faster and faster and in that sense become “smarter” — maybe “more adaptive” is a better way [to say this]. And I think it’s the same with what’s happening with technology now. There’s a lot of discussion about children being exposed to digital technologies, etc., and the rate at which technology is changing, so the signals that people are getting — the variation in signals — is just getting more and more and more, and so people’s brains are becoming increasingly adaptive.
There’s somewhat of an analogy with AI there, where, really, the amount of data — the variation in signals that we’re feeding our models — is continuously growing exponentially, and so of course the models are becoming more and more flexible and adaptive as well.
Shervin Khodabandeh: Nitzan, you’ve been pretty vocal about AI not being just a fancy technology or a series of fancy algorithms just because they’re cool and awesome.
Sam Ransbotham: Which they are.
Shervin Khodabandeh: There has to be a purpose for it; there has to be a real need for it. Comment more about that meta framing of AI strategy and its alignment with business and corporate strategy at eBay or, in general, the purpose of AI, because I know you’ve been pretty vocal about that.
Nitzan Mekel-Bobrov: Yeah. I’ve probably been pretty vocal because I learned it the hard way, getting somewhat burned in [the] earlier years of my career coming into a new company and thinking I knew it all — that I knew it better than everyone because I understood the technology and what’s going on under the hood. And I learned the hard way, as I say, that it’s a lot more nuanced than brute-forcing a technology onto a theoretical use case that you might think of. It’s really important to understand the business context in which these technologies are deployed and understand it deeply.
So, for example, early on in my career, when I was in financial services, we were using AI for a lot of automation — workflow automation. There’s huge amounts of savings in this — hundreds of millions of dollars a year. And to me, it felt obvious that certain applications — for example, with speech recognition and intent detection, etc., in the call center — would be an ideal fit. But it’s only after I actually shadowed a number of agents and spent probably about a good month deep-diving into their workflows that I understood that there’s so much complexity there; that it’s really about the interaction between the human intelligence and the machine intelligence, and it was making assumptions that just weren’t going to bear out in real life. So it’s that move from what works in the lab to what works in real life that is really critical.
And then, of course, thinking about what’s important to the business is not just a matter of what’s important today but really what’s in the DNA of the company, because it takes time to not just build but deploy and get these things up and running. I don’t know any one of my peers who’s ever been able to get anything up and running at scale in a matter of probably less than a year [before] you see real impact, and honestly, it’s often quite a bit longer than that. It’s a marathon, it’s not a sprint, and so you really have to be conscious of what will be the business strategy two, three, four years down the road, not just what is on the executive’s mind today.
Shervin Khodabandeh: I want to pick up on something you said about automation, which is one of the larger themes that AI’s being used [for], but I think you also alluded to it. I think it’s unfortunate that when most people think about AI, they tend to think of it as the extreme case of “It’s going to replace humans,” and you were talking about the importance of what I call the middle ground, where human and AI work together, so that human-AI interaction is key. Perhaps that’s also one of the reasons it’s so hard to scale, because you’ve got to figure out how humans will work differently with AI. Can you share some stories or insights about how you’ve done that? Because it requires changing the mindset of humans and what they normally do.
Nitzan Mekel-Bobrov: I can give you a couple of examples that I’m seeing at eBay, but honestly, a lot of what I’m seeing at eBay, I’ve seen before. The transformation stage that we’re on now is one that virtually every company is on, and no one is fully there yet. So if you think on the back-office side, specifically customer service, I think that anxiety is typically most acute there, because frankly they’ve seen it before with other technologies; this isn’t their first rodeo. And in truth, over the course of the maturation of the technology, there are individual roles that are no longer needed. That is true. But it’s not that humans aren’t needed; it’s just that the role that they do changes.
For example, on the customer service side, what we are doing at eBay — we tried for the past few years to inject AI in different places of the flow. And it was very challenging, both from a tech-debt perspective, because there’s just a lot of tech debt in different places that made it hard to do that. To make it more concrete, I’ll give a real example: intent detection.
A customer calls, and as they’re talking to you, there is a model that’s picking up on what they think that the customer’s trying to answer, so it’s supposed to help the agent go to the right pages for help, surface the right information. Think about that poor agent, though: He’s talking to a customer who, typically, when they call, it’s not because they’re happy; when a customer calls customer service, it’s because they’re frustrated. They’re trying to help them while at the same time getting these messages on the screen, and that ability to multitask and pay attention to this thing that’s flashing at them with intents can be more of a distraction than a help. What we’re doing now is undertaking a more end-to-end approach, where we’re really replacing or transforming customer service on our back end so that both the systems that the agents are using and the systems that the models are running on are designed together from the get-go — so there’s a much better interplay.
And we also have a lot of our designers — much more creative than our AI folks, in many ways — helping think through what that interplay should look like, doing a lot of user research testing with agents on how they would interface with customers and technology at the same time.
Sam Ransbotham: The verb you used, inject, was interesting. You use that as saying that it didn’t work to inject; it didn’t work to Band-Aid it on or just to paste it on.
Shervin Khodabandeh: Well, another thing I wanted to pick up on, Sam, is, Nitzan, I think you made a comment like, “This is not their first rodeo with technology; they’ve seen other technologies.” And I’m interested in your views on how AI might be … or how AI is different compared to all the prior, let’s say, technologies that came and transformed functions or processes, and whether you think there are any misconceptions there — like people anchoring [on], maybe, “Oh well, this is just another ERP technology or like we did with some other technologies; it’s the same.” And I wonder whether you agree with that or whether you think there’s a misconception to treat it the same, and what the difference might be with anything else people might be referencing or anchoring on based on their prior experiences.
Nitzan Mekel-Bobrov: Yeah, I think any specific piece of AI technology is comparable to an ERP system or some other version, but as a paradigm, it’s a much bigger thing. I think the better analogy is the digital transformation. The change from brick-and-mortar to digital is probably more at the level of [an] analogy. Even where we are now, I would say, is a paradigm shift akin to the digital paradigm shift much more than just the introduction of some new ERP or CRM, etc.
Sam Ransbotham: Nitzan, we have a new segment where we ask our guests a series of rapid-fire questions. Just answer with the first response that comes to your mind. What’s your proudest moment with artificial intelligence?
Nitzan Mekel-Bobrov: Probably on the health care side. When I was in the Boston Scientific days, we rolled out a feature that was predictive, basically using signals from a pacemaker to predict heart failure. On average, it was 30 days in advance of when physicians would detect it otherwise, which is a huge lifesaving benefit.
Sam Ransbotham: What worries you about artificial intelligence?
Nitzan Mekel-Bobrov: Misuse. Misuse by bad actors, whether it’s in military operations or in other types of nefarious activities. If you think of deepfakes as an example, just as it becomes more and more accessible and easy for everyone to use, I do have concerns.
Sam Ransbotham: What’s your favorite activity that involves no technology at all?
Nitzan Mekel-Bobrov: Oh, crap. I was about to say “photography,” and then I realized that’s not even a good example.
Sam Ransbotham: No, actually we can count that. I mean it’s not strictly artificial intelligence, so we’ll give you credit for that.
Sam Ransbotham: What did you want to be when you were a child? What did you want be when you grew up? AI engineer at eBay?
Nitzan Mekel-Bobrov: I actually started as a creative writing major, so I guess writer was my initial passion.
Sam Ransbotham: What’s your greatest wish for what we’re going to do with artificial intelligence in the future?
Nitzan Mekel-Bobrov: I think, making it easier for almost anyone to make a living pursuing their passion.
Sam Ransbotham: Nitzan, great meeting and talking with you today. I think one thing that is going to resonate with a lot of our listeners is this idea that AI implementations are different than existing implementations like ERPs; that that siloed approach that you might take toward a monolithic technology is very different when you involve lots of users, and particularly when you involve your two platforms, with both your customers and your sellers. Thanks for taking the time to talk with us. We really appreciate it; thank you.
Shervin Khodabandeh: Thank you for being with us. It’s been quite insightful, and we really, really appreciate it.
Nitzan Mekel-Bobrov: Thank you, Shervin.
Sam Ransbotham: Thank you for joining us today. Next time, we’ll talk with Helen Lee, technical fellow and regional director at Boeing. Please join us.
Allison Ryder: Thanks for listening to Me, Myself, and AI. We believe, like you, that the conversation about AI implementation doesn’t start and stop with this podcast. That’s why we’ve created a group on LinkedIn specifically for leaders like you. It’s called AI for Leaders, and if you join us, you can chat with show creators and hosts, ask your own questions, share insights, and gain access to valuable resources about AI implementation from MIT SMR and BCG. You can access it by visiting mitsmr.com/AIforLeaders. We’ll put that link in the show notes, and we hope to see you there.
