With guests from established enterprise companies to start-ups barely out of infancy to everyone in between - you’ll get the inside scoop on what’s Up Next in Commerce. New episodes come out every Tuesday and Thursday.

Subscribe to get notified of new episodes and our Up Next in Commerce weekly newsletter. 

EPISODE 93

The Big, Bold Future of A.I. in Retail and Ecommerce

With Alex Spinelli, CTO and EVP of Product, Technology, and Operations at LivePerson

Or listen in your favorite podcast app

Apple Podcasts  /  Google Podcasts Spotify

Interacting with customers requires a level of finesse and talent that is beautiful when done well, and a tough sight when done poorly. There is give and take, and you have to flow through various movements and ups and downs to reach a satisfying end result. It’s like a dance. A tango if you will. At least, that’s how the folks over at LivePerson see it.

Alex Spinelli is the CTO and EVP of product, technology, and operations at LivePerson, and on this episode of Up Next in Commerce, he broke down what that dance should look like, and how A.I. is taking the lead.

As Alex explains, LivePerson is a set of tools, technologies and platforms that enable businesses to have conversations with customers through messaging channels, and to detect where customers may be getting stuck or frustrated. Then, with a small immediate intervention, LivePerson’s A.I. routes that customer to a human who can make the buying process easier. It is a way to get to a better end result more often, and it works. Businesses using LivePerson have seen double-digit-percentage-point improvement in conversions and higher NPS scores than ever. But the power of A.I. doesn’t end there, and Alex dives deep into where we are headed with A.I. as a tool in retail, including the blended in-person and virtual experiences that seem to be overlapping more than ever before. And Alex gets into the nitty-gritty of the ethics behind A.I. and how everyone will have to be more involved going forward when it comes to defining their limits, wants, and needs. Enjoy this episode!

Main Takeaways:

  • Joining Forces: The future of A.I. in the ecommerce space is in the way brands can join together an A.I. experience with a human-based one. The way brands should be looking at A.I. is as a conversation-starter and a tool that can solve transactional problems, but when a deeper conversation is needed, it should be able to usher customers through a seamless transition to a real person who can build a relationship, form trust, solve problems, and ensure that the customer experience is a good one the has a positive end result.
  • Let’s Get Ethical!: With any new technology, there are ethical questions that have to be addressed. This is especially true when dealing with A.I. Not only do you have to take into account the repercussions that A.I. will have on the labor force, but you also have to consider how A.I. is being trained, what kind of biases are being programmed into the model, and how and when to start and stop collecting data to build bigger and better A.I. models.
  • Blend It Up: As we move further into the fourth industrial revolution, we are beginning to see more blending of virtual, digital, and physical experiences. Conversational technology will begin to follow us into physical stores and A.I., along with more targeting-types of technology, will be used in and out of stores.

For an in-depth look at this episode, check out the full transcript below. Quotes have been edited for clarity and length.

Key Quotes:

“Smart speakers are very transactional, right?… I ask for something, I get the result, I ask for something, I get the result. And what we’re trying to do here, because we’re working for all different businesses and companies, is allow you to have a full-fledged conversation to build a relationship with the things that are important to you in your life, your bank, your healthcare company, your insurance company — Industries that tech and AI have tended to ignore.”

“If you look at banks, telcos, all these big, stodgy, old businesses — or at least we used to think of them that way — they lost their differentiation. A banking app is a banking app is a banking app. They all look the same, feel the same, act the same. But not all banks are the same. They have different values, they have different missions, and without being able to talk and have a conversation, you don’t get to connect your values to where you’re putting your money. So I think that’s the shift. I think people now care. They’re spending all their dollars in the digital world, by and large. Even the restaurants, who’s delivering? Is it Postmates? Is it Uber? Is it Grubhub? It matters. We want to support the right business who has the values that we share.”

“The whole platform lets you never get stuck. So you can have a conversation, it can be part of an automation, you can be looking at a product, you can be asking questions to a bot about the size, color, compatibility, etc. And then when you get stuck, we can actually detect that and route you immediately to a person, a real human being that can help you. And we call that the tango.” 

“Explainability is still a big problem with A.I. It’s not a solved problem, and I don’t have a good answer, actually, of when do you have enough [data], when don’t you have enough? I think you need to constantly benchmark, constantly look at your accuracy, and have all the protections in place that you are looking for that bias, you are looking for those negative consequences. And that’s hard work. That’s not like putting some technology gaps in place and a threshold, that’s really having a dialogue internally, asking the questions, turning over rocks, what could be the negative consequences here?”

“Because digital is a necessity not a convenience, we’re starting to see a little breakage of the flat experiences. So I think the big opportunities are around how can we really help people discover… To me, that’s all about actually modeling the behaviors that we would have in real life. We’re going to go back to stores, we’re going to go back to malls, but they’re going to be changed up, they’re going to be very different. I think we’re going to see conversations in the digital world follow us in and try to fill in the gaps and start to really help us in a much more blended way.” 

“I’m hopeful about the future, I actually think A.I. is going to be a powerful tool of change, positive change. I don’t think it’s going to kill everyone’s jobs. I actually think we’re going to find new ways to make it augment and enhance us in ways we don’t even expect. So I guess in the A.I. space, in Big Tech space, I spend a lot of time talking, I hear a lot of fear and the sky is falling. And I guess I don’t think that way. I think I’m pretty uplifted and positive about what the future is to come.”

Mentions:

Bio:

Alex Spinelli leads the technology organization at LivePerson, overseeing all R&D globally. Prior to LivePerson, he was global head of Alexa OS for Amazon.com, running a large, distributed team of developers around the world and leading the development of the core software systems and capabilities that underpin the Alexa platform.

Before joining Amazon, Alex served as CTO of McCann Worldgroup — one of the world’s largest marketing organizations with 24,000 staff across more than 100 countries — and CTO for North America and Asia-Pacific at AXA Technologies, where he built large-scale grid computing for predictive modeling. At Thomson Reuters, he oversaw product and technology for news and media, including real-time, AI-driven news curation for financial professionals.

He has also held executive technology leadership positions at TheStreet.com and MTV Networks as CTO of Comedy Central. Alex started his career as an engineer in New York City, working on real-time data feeds. He has an MBA from Duke University.

Up Next in Commerce is brought to you by Salesforce Commerce Cloud. Respond quickly to changing customer needs with flexible Ecommerce connected to marketing, sales, and service. Deliver intelligent commerce experiences your customers can trust, across every channel. Together, we’re ready for what’s next in commerce. Learn more at salesforce.com/commerce

Transcript:

Stephanie:

Hey, everyone, and welcome back to Up Next in Commerce. This is your host Stephanie Postles, CEO at Mission.org. Today we have Alex Spinelli joining the show. He’s the CTO at LivePerson. Alex, welcome.

Alex:

Thanks for having me. I’m excited to have this conversation.

Stephanie:

Yeah, I’m really excited to have you on. So I was looking through your background. And I was hoping we could kind of start with your days at Alexa, because I feel like there’s probably a lot of good juicy stories there and I want to hear a bit about what was your role there? What did you do there? And then we can jump into the big topic around AI and your current product?

Alex:

Sure, sure. I led what we call the Alexa OS. And what that was, or is today, is really the core software platform that powers the Alexa experience, the brain. It included things like personalization, speaker recognition, so Alexa knows who’s talking to her. And then all of the APIs and technologies, dialogue management, they really power the whole experience and allowed both internal developers at Amazon and skill builders, so skills are like apps for Alexa, to go and build those experiences. So it was really the cloud operating system for Alexa.

Stephanie:

So what drew you to that field and industry?

Alex:

Yeah, so I’ve always been pretty connected to AI, natural language, even going back to, I have a lot of roots in news, something I was pretty passionate about, in news technology. So at Thomson Reuters, for example, where I lead technology for news, both for real time news, algorithmic trading, and then also all the Reuters news properties and journalists, the tools that journalists use, I spent a lot of time trying to understand how do people consume information, how they read information, and how can digital and computers really help us find the most important things, gain insight from information, gain insight from data.

Alex:

So then I kind of took a little bit of a hiatus from news, and when I joined Amazon, I was leading search. So the whole experience for browsing and discovering the right product for you, and trying to optimize that, make it easier. And one of the things that was really interesting is, I started to see the limitations of these very flat experiences, search pages, web pages and apps. And people started to try to have a conversation with Amazon search. So they’d ask questions in search, is this product compatible with this one? What’s the best gift for my daughter who’s graduating high school? And all these interesting questions, and the experience often fell down. So we actually started looking at what we called query understanding and natural search and all these interesting things where we wanted to help people get answers to their questions and have a dialogue with the search experience.

Alex:

And I thought it was pretty hard. In the sort of traditional, I put in a query, I get a set of results, that interface just didn’t work really well. Alexa at the time was just kind of quirky little device that was just launching at Amazon very early days, I actually had one, I was part of the early beta testing. And I said, “God, I want to be part of that. That actually is starting to recreate the way we’re going to interact with our digital lives and we’re going to use natural language. And I guess the rest is history. I think when I joined there was 20 or 30 people in the team. And again, it was this quirky little device that people were like, “What the hell is this thing? Is this going to be big?” And yeah, in six months, we sold millions of devices, and I was growing a team, and we were adding all kinds of new features and capabilities. And it was pretty much a rocket ship, which was pretty fun.

Stephanie:

That’s awesome. I have Alexas throughout the house. I’ve always wondered though, how to get past that hurdle of, like when you’re talking with someone, you are very free flowing and you’ll ask any kind of questions. And I feel like oftentimes, with Alexa or any speaker that you would talk into, you’re like, “Ah, ah, what can I ask? I don’t know what to say. I don’t know how to phrase it.” And it feels like there’s still a bit of a hurdle with a lot of conversational speakers to get past the getting you help kind of with anything, and being able to query things in a million different ways, so that you’re not like me, where you’re just like, “And I’m stumped and now I’m just going to open up the app on my phone and resort to the old way of doing things.”

Alex:

Yes. So it’s interesting, because that same challenge is what actually led me to LivePerson. So when I met my current boss, he was explaining what LivePerson was doing, which was really a messaging platform for customer service and sales. And he said, “Listen, we’re really interested in taking things to the next level with AI.” And my first response was, “I’m good. I’m at the hottest product on the planet.” And Rob and I had known each other for years from New York and we had conversations earlier. I said, “I’m good.”

Alex:

He said, “Well listen, there’s an opportunity to take what you’re doing at Amazon, creating these natural experiences, but actually democratize it and do it for companies all around the world, large and small, and really help consumers interact in a new way.” And it kind of stuck with me. And I started, we had more and more conversations, and I ended up joining. And I think the key differentiation that you’re seeing is, I think the smart speakers in that whole space, they aren very transactional, right?

Stephanie:

Mm-hmm (affirmative).

Alex:

They’ve kind of started to center around entertainment, home entertainment, smart home. And they are still fundamentally very, very, I ask for something, I get the result, I ask for something, I get the result. And what we’re trying to do here, because we’re working for all different businesses and companies, is allow you to have a full-fledged conversation to build a relationship with the things that are important to you in your life, your bank, your healthcare company, your insurance company — Industries that tech and AI have tended to ignore, like, “Those are big, boring, you can’t change them.” And I think the problem is we’ve kind of leaned into these proxies of relationship building, apps, well, you can’t build a… You and I are having a conversation. I didn’t send you the A.L.E.X. app and say, “Here, you can get any answer to any question, you can click and browse and tap and search, and you’ll get…”

Alex:

You didn’t send me your app. We’re having a dialogue and a conversation. What’s crazy is businesses have put the app, they’ve actually done that crazy thing. They’ve said, “Oh, no, don’t talk to us. Don’t have a conversation with us here, use our app, use our website.” And what we want to do is actually do exactly what you described, have that fluid conversation, build a real relationship. And the key for us, and this is where I think the smart speakers fall down, is humans have to be involved as well. So you can’t get stuck. The AI is not going to be able to solve every dialogue.

Alex:

So the way we look at the world is the AI as a kind of concierge in many ways, and begins and initiates the dialogue and conversation for simple things like play music, do this, do that, AI can do it. But then when you really need to have a more meaningful conversation, we want to connect you with the right person. And Alexa can’t do that, because just the scale wouldn’t work. It’s just for Amazon, where I think when you start thinking about democratizing AI, we can actually start to do that and make it a useful tool, not just for the consumer, but also for the employees of the business.

Stephanie:

Yep. I mean, now it seems like it’s the perfect time too, because I think through the past five years or so, and it seems like we’ve kind of gone through a period where everything had to be optimized, you don’t want to have support centers, you’ve got bots everywhere, you can do drop shipping now, you don’t need brands, you don’t need… Just white label products, we went through this phase, and now we’re kind of coming out on the other side where people are like, “I don’t really want to talk just to a bot, I want to talk to a person. If I instantly want to call, I want to be able to have someone there. And it seems like now consumers’ expectations have changed where it’s like we’re a little bit there, we were getting used to just, “Okay, I’ll just talk to the chat and see if it fixes it.” And now it seems like expectations are so much higher than they even were just a couple years ago.

Alex:

I think part of that is this digital was growing due to convenience, right?

Stephanie:

Mm-hmm (affirmative).

Alex:

We were buying large bulk things. We were buying simple things. We were buying more toothpaste, more batteries, more this, more that. And as we started to need to use digital, and now in the pandemic, obviously, need to, have to, no other way, for all the things in our life. Yeah, we want to actually connect those things to values, our values, right? So the brands matter. The business matters, what does that brand stand for? What are the values that they stand behind? So I do think you’re right, I think the need for developing a real relationship is important.

Alex:

And if you look at, actually it’s interesting, banks, telcos, all these kind of big, stodgy, old businesses, or at least we used to think of them that way, or kind of were perceived that way, they kind of lost their differentiation, right? A banking app is a banking app is a banking app, they all look the same, feel the same, act the same. But not all banks are the same, they have different values, they have different missions.

Alex:

And without being able to talk and have a conversation, you don’t get to connect your values to where you’re putting your money. So I think that’s the shift. I think people now care. They’re spending all their dollars in the digital world, by and large. Even the restaurants, who’s delivering? Is it Postmates? Is it Uber? Is it Grubhub? It matters. We want to support the right business who has the values that we share. So yeah, I think it’s really important. A connection is super important.

Stephanie:

Yep. I also think a lot about retail. A lot of people probably do miss those experiences of going in stores and having someone there to ask questions to, and now that just needs to be mimicked more in the digital space where people are like, “Well, I can’t,” or maybe they can start to now, but for a while there, you couldn’t go in and have your normal conversations and ask where things were. I mean, I go all the time, and I’ll be like, “What kind of wine do you like? Just tell me what you like. I’ll buy whatever you tell me because I don’t know.” And I miss that. And I was looking for that. But oftentimes it was lacking in the digital world. So-

Alex:

Yeah, so I think developing the tools to allow the… Brands are made of people, and enabling people to actually come through the digital world and connect is exactly what you’re saying. Yeah, we thrive that human experience. I mean, we desire that human experience.

Stephanie:

Yeah. So tell me a little bit deeper about what is LivePerson? Because I’m hearing it is like it’s essentially conversational AI for any industry, it’s not just focused on commerce, it can be banking, it can be anywhere, is that the right way to think about it? Or maybe I’ll let you describe it better.

Alex:

Yeah, so at our core, at our roots, it’s a set of tools, technologies, platforms, that enable you as a business to actually have conversations with your customers through messaging channels. So this is the way we’ve all started to interact with one another. My daughter and I don’t talk on the phone as much, it’s all messaging, but I can talk to her all day long, right? Because it’s asynchronous, it’s on my time, it’s on her time, she can be in class, I can be in this meeting, we can start a conversation and continue it. So the core offering is letting businesses do that. So giving those interactions back to the consumer on their schedule. And then we start to layer on a lot of intelligence.

Alex:

So a lot of those conversations can be led by an AI to gather information, to do the simple things, to actually help you with, what’s your name, what’s your account number, what’s your size, reset my password, pay my bill, lots of things that really become kind of very rote. And then you start to really get more and more advanced in enabling you to shop, enabling you to buy, enabling you to transact. And the whole platform lets you never get stuck. So you can have a conversation, it can be part of an automation, you can be looking at a product, you can be asking questions to a bot about the size, color, compatibility, etc. And then when you get stuck, we can actually detect that and route you immediately to a person, a real human being that can help you. And we call that the tango. So it’s this-

Stephanie:

I like that.

Alex:

… beautiful dance that allows us to go back and forth. And that’s really, I think, where we excel. And then just from a technical perspective, we wrap all of that with a set of analytics and tools that even if you’re a small business, you can use to look at the health of those conversations, how’s it going? Is it making you money? Is it costing you more? How’s your customer satisfaction? And those kinds of things. So it’s a pretty full suite of tools to build a new kind of customer experience.

Stephanie:

That’s awesome. So what kind of results do you see? Especially around commerce, when it comes to, like you said, you’re doing the tango, you’re sending them over to a customer service person. What would you see, otherwise? I’m sure losing that customer and not converting to a sale, are there any metrics that you guys have that you can share?

Alex:

Yeah, so it definitely is industry dependent and customer dependent. And we tend not to share direct customer numbers. But this is why I joined, the results are crazy. And so when Rob and I were talking about me joining LivePerson, he said that, “We’re kind of onto something where we see costs go down, customer satisfaction go up, NPS go up, conversion rates go up, and agent turnover, or sales agent turnover go down.” And I said, “There’s no way all those metrics can be moving in the positive direction.” Usually, there’s trade offs. But right now, that’s what we’re seeing.

Alex:

So we do see conversion rates for conversations to be often double digit percentages better than experiences that didn’t have. So if you were interacting in an app or a website, and we detect that you might be stuck, you might be jumping back and forth between pages, we’ll actually offer like, “Hey, it looks like you might be having… Do you have a question? Do you have a problem?” And then we’ll have that dialogue and that conversation.

Alex:

And that might be a tangoed conversation mixed and matched between an AI and a human. And we see conversion rates of those dialogues, again, double digit percent. There’s a large big box retailer whose conversion rates typically exceed over 15% when a conversation is initiated. And a typical conversion rate on web shopping is single digits, mid to 5%, 6%. So significant increases when you actually connect and have a dialogue are pretty common for us.

Stephanie:

Wow, that’s cool. So if I’m a brand-

Alex:

Yeah, it’s pretty powerful.

Stephanie:

Yeah, I mean, it sounds amazing. It sounds like, why wouldn’t someone use something like this? If I’m a brand, and I’m thinking about setting this up, would you be tapping into my customer support people who are trained my way and then you’re like, you train the AI, you got the questions in there, the answers, you kind of map all that out, you’ve got your database, and then you’re constantly learning, I’m assuming, from what people are saying and what’s actually helpful. And then, when you go into the tango mode, it goes over to your customer service people, or how does that work?

Alex:

So our tooling, mostly, is used directly by the brand. So you’re a brand, our technology sits inside, I mean it’s SaaS base technology, but it sits inside your contact center. Actually, the way we typically will train AI, it’s actually pretty cool, you have human conversations first, and you don’t need many. So you actually start to have human to human conversations. And then just in a few weeks, we can actually collect enough data to go and build the best intents. So intent is, as you’re having a natural language discussion and an AI is detecting what you need. So an intent is the thing that you want, I want to pay a bill, I want to buy that product. Is this product compatible? Does this come with batteries? Whatever have you. Those are all intents.

Alex:

So those intents are basically derived from your real customer conversations. So the accuracy ends up being very high. And we’ve actually built the whole series of proprietary data models that are very industry specific. So in retail, in airlines, in banking, in insurance, we can actually have some really high accurate recognition. And again, those intents can be recognized for human conversations, so that we can tell the agent exactly what’s going on and what this person needs. And then they’re also used to go and build those AI driven experiences.

Alex:

And the goal is, can we take all the mundane repeatable stuff away from the agents? So the agents are really closing the sale, they’re really helping tough problems. And this is why you see agent satisfaction go up, because they’re not doing the rote, same, same, same, same conversations, all that’s done by the AI. And then the agents actually having a kind of much more high bandwidth interaction.

Stephanie:

Like doing the creative work, where they can think and solve- [crosstalk]

Alex:

Exactly.

Stephanie:

… problems. And I mean, I think it comes back to, for a while there, everyone’s like, “AI is going to take our jobs.” It’s like, “No, it’s augmenting your jobs. And it’s doing the things that you probably don’t want to do anyways. But now you just get to work on higher level things,” I would think.

Alex:

We do see that. And it’s interesting, we see the wait times… So rather than waiting for 30 minutes, you actually wait for very little time. And then that agent can actually spend the time and energy to have a, just like you said, a much more creative high bandwidth conversation. So we don’t see this, “Yeah, take…” It’s changing jobs, it’s augmenting jobs, it does require some new training, for sure. But at least right now, it’s not this job killer, it actually opens up the world for new jobs. We actually are converting agents to data annotators.

Alex:

So agents in real time can actually go and label conversations and data to improve the AI. So it’s actually advancing their roles in some ways. Again, I’m not going to be naïve, there will come a time where automation and those kinds of things do impact jobs at scale. I think we as responsible business people need to think about what’s the next thing, right? And what’s the next set of opportunities. So I’m hopeful in general, we are pretty hopeful and positive on where we can get to, but I think we have to kind of wade in very open eyed and make sure we do the right thing as we go forward.

Stephanie:

Yeah, so when thinking about doing the right thing, I think it’d be good to get a little lay of the land of the AI field in general, because I feel like it’s had a pretty bumpy couple of years, just I mean, so many headlines were made around unintended consequences of using AI models, labeling things incorrectly. There’s just been a lot out there. So what does it look like now? And especially in the world of commerce, how do we think about, where is AI even being utilized properly, or misused? And where could it be in a couple years? Or where should it be?

Alex:

The biggest challenge with AI is bias. And I’ll explain what that means. So some of the bias is deliberate, some is not deliberate, or intentional and unintentional. So AI is only as good as the data. So what AI is, at the end of the day, it’s a tool that allows you to look at lots and lots of data, examples. And then you build a model, mathematical model, statistical model, that makes certain assumptions based on the examples. And so if you were trying to make an AI service that would recognize oranges, this is when I used to meet people in person at South by Southwest to give a talk, I put this big, gnarly looking orange on the screen. It was not orange, it was mostly white and moldy and green. And I’d asked the audience, “Who knows what this is?” And 95% of people would raise their hand, I’d say, “What is it?”

Alex:

And everyone knew it was an orange. No one saw an orange [inaudible] anything like that. Unless you went on vacation and left an orange on your counter for like three weeks, then you maybe have seen an orange like that. But everyone instantly recognized it, right? So really what’s happening is very similar to AI, you’ve seen thousands of oranges in your life. So there’s a bunch of features of an orange, little dimples, the skin, the [inaudible] possible colors, it’s round. And your brain immediately matched that image to this archetype of an orange, even though you’ve never seen an orange that looks like that. That’s really what AI is doing. So if you have a bad data set, if you showed a young kid, many, many different images of things that weren’t oranges, but you told them it was an orange, right?

Alex:

When that image would pop, they might not recognize it, they might not be able to tell what it was, or they might misrecognize something else and call it an orange. So the data under AI can inherently have bias by where we collect it from. So if we’re trying to collect data to recognize a certain type of person, or a certain type of behavior, or a certain type of language, if we don’t have a representative data set from the right populations, the AI is going to be biased, it’s going to be biased based on that data set. Or if humans label that data, and those humans come from a certain one kind of homogenous background, they might label the data with their own biases. Again, that could be unintentional, but everyone brings their own unconscious biases with them.

Alex:

Or lastly, engineers, when they build AI models have their own biases, too. They think certain things are important and other things are unimportant. Simple kind of innocuous examples, if you’re building a trading system, and you thought the weather really had an importance on stock prices, so that engineer would build weather as being a heavily weighted variable, that’s a bias. So the key actually is recognizing all the ways that biases can creep in and creating some standards, and then really having the tools and process in your company to actually recognize those, and ask the question, and make sure that you’re doing all the things that you need to do to eliminate that. That bias is what we’ve seen in some of these horror stories around AI.

Alex:

And it was kind of rushing headlong into it and not really thinking deeply about it. So there’s actually some great organizations that… We actually took the EqualAI pledge as a company. So there’s a nonprofit called EqualAI, that is working with industries to try to eliminate bias in AI. Or I should probably rephrase, not eliminate, but really try to mitigate. I don’t know if we’ll ever be able to fully mitigate-

Stephanie:

Or eliminate.

Alex:

Eliminate, sorry, yep. It’s just super important, and it’s actually the right thing. It’s good for your business. It’s good for your employees, it’s just something we have to do, ethical use of any new powerful tool, no matter what it is, you have to actually consider those things. So I think that’s the key challenge right now. And I think we’re still in the early, early days of really grappling with it.

Stephanie:

Yeah, I mean, it seems like too, right now, a lot of models will have to be thrown away because they were all trained on a set of consumers that are still there, but now there’s all these new consumers who’s come on the market that were never shopping online before, never doing their grocery shopping online. And we had a really good guest from Stitch Fix, she was the VP of data science there. And she was like, the way that the older generation, who’s now trying out Stitch Fix, wants to talk with us is very different than how we were talking with millennials. And so you have to start rethinking about, do I keep my model and adjust it? Do I just throw it out and start over? And it seems like a tricky point now, because you just had this big inflow of new consumers that you never really were talking with before last year.

Alex:

Yeah, I have a 14-year old. And it’s been an interesting journey as most of his communication has really moved online, and the words and terminology and no cap and sus and all these funny terms. They need to be built in to model. So your VP at Stitch Fix is 100% correct. And we need to go and deal with that. And I don’t know if it’s about throwing away, I think it’s more about augmenting and building on top of. The good news is the influx is a big number of people coming, and those large numbers can actually improve the models pretty quickly, if the number that you’re starting with was smaller. But it’s something we actually, because we live and breathe natural language, we actually have to stay on top of really, really regularly. And yeah, this is going to be the perennial challenge. This is where actually that, remember I talked about I’m transforming agent roles?

Stephanie:

Mm-hmm (affirmative).

Alex:

So one of the things that we’ve looked at and we’ve just actually released this, it’s pretty cool, is the agents, as they’re having conversations, can label the intents and they can actually improve intents, and they can actually retag intents and all this kind of stuff. And so we believe, and this is where I think we are changing roles, not eliminating, the agents know your products, they know your language, they know the markets you’re operating in, they talk to your customers every day, and they’re the ones that are going to be best positioned to kind of add and augment those models, so it’s actually really important to have them as part of that process.

Stephanie:

Oh, that’s really interesting, that they can do that now. And I could see it being really helpful too, because I have heard that oftentimes, models can also, like you said, train themselves and turn into a black box where it’s like they keep ingesting the wrong data, wrong data, and then you build up maybe algorithms that, I remember at certain companies I used to work at, you kind of didn’t know what was in there, at a certain point, you’re like, “I don’t know how it’s working. I don’t know why it’s working this way.” How much data do you need? Is there ever a point where you’re like, “That’s enough, let’s stop. Only collect it this many times.” Because right now, it feels like we’re in a world of like, just get as much as you can and ingest as much as you can. Which seems like it could maybe have unintended consequences.

Alex:

Agreed. But explainability is still a big problem with AI. So there’s a startup for folks out there listening, create some technology that offers introspection and explainability to large machine learn models. It’s not a solved problem. And I don’t have a good answer, actually, when do you have enough, when don’t you have enough? I think you need to constantly benchmark, constantly look at your accuracy, and have all the protections in place that you are looking for that bias, you are looking for those negative consequences. And that’s hard work. That’s not like putting some technology gaps in place and a threshold, that’s really having a dialogue internally, asking the questions, turning over rocks, what could be the negative consequences here? It’s kind of active management right now, and it really needs to be baked into your kind of culture, that it’s something that you focus on.

Stephanie:

Yeah, definitely agree there. So what kind of opportunities do you see? Where do you think conversational commerce should be in the next one to three years? Or what do you think is going to start happening?

Alex:

So I think the big opportunity right now is actually the topic of our chat is more commerce, like real shopping, real purchasing, real buying, I think conversational commerce, primarily over the last number of years, has been sort of sat in that care, support, follow up space. And now because digital is a necessity, not a convenience, we’re starting to see, like I said before, all the little breakage and the flat experiences. So I think the big opportunities are around how can we really help people discover?

Alex:

So discovery is really hard. With Alexa, for example, you don’t know what Alexa can do and can’t do, and those kinds of [inaudible]. Discovery is that still a big challenge. Huge opportunity there. It’s how do you stitch conversations together with discovery? And to me, that’s all about actually modeling the behaviors that we would have in real life. So we’re going to go back to stores, we’re going to go back to malls, they’re going to be changed up, they’re going to be very different. I think we’re going to see conversations in the digital world follow us in and try to fill in the gaps and start to really help us in a much more kind of blended way. So there’s something called the Fourth Industrial Revolution, if you really want to geek out. It’s this-

Stephanie:

Oh, yeah, we’ve talked about this before-

Alex:

Oh, cool. Yeah.

Stephanie:

… on a lot of our other podcasts.

Alex:

It’s this blending of the virtual and physical. Yeah. So I think the big opportunities are in real commerce, and how do we start to blend the physical and the virtual? So we see, for example, especially during COVID, blending conversations with curbside pickup, I’m ready, I’m here, are you here? I want to add this, can you get me… So really trying to fill in all those gaps in those interactions and exchanges. So I think that’s where a lot of the kind of next stage plays, is we’re going to see conversations start to power a lot more of our transactions and commercial activities, and starting to blend together that physical and virtual, that’s where we’re spending a lot of our time.

Stephanie:

That’s cool. I mean, so what kind of tech advancements are needed, because I’m even thinking about that Fourth Industrial Revolution and blending that, and okay, if you’re walking to a store, I mean, I know there was a while there where store owners were hesitant to even install the beacons so that you would know who’s coming in your store. And there was a lot of hangups when it came to retail that didn’t allow the digital world to interact with them, because you had to have hardware infrastructure changes, there was a lot needed there. So what kind of things are needed for that advancement to take place?

Alex:

Yeah, well, a bunch of things. I mean, these things have basically become supercomputers, right? These are more powerful than even the biggest machines 10, 15 years ago. So they’re going to take on more and more of the processing. I think image recognition, big space. And then I think a lot of that starts to wrap together the privacy concerns, so giving control back to the consumer about what data I share and when based on my needs and what I want to do. So that’s where I think you’re going to see a lot of technology advancement is, yes, beacons. Yes, image recognition. Yes, the kind of blending of conversations and in person and then live and all these kinds of things and trying to, like I said, stitch that experience together.

Alex:

But if I were, again, entrepreneurs out there and technology companies, I would look at those for sure, but I think we also have this kind of renewed interest in privacy and what targeting is. And we can do a whole soapbox, if you want, on the evils of free social media and the hyper targeting, I think there needs to be legislation to almost eliminate some of that, because what we’ve allowed companies to do is extreme content and extreme information can find audiences now, because the audience is basically free, they just [inaudible]. So I think trying to really understand who I share my data with, why I share my data with, and I’m sharing it only for the purpose that I want, is another whole area of technology that we need to focus on. These are the things at least we’re working on, that we feel pretty passionate about.

Alex:

And then in terms of very specific technologies, I think the combination of conversations with location and image recognition, are going to start to be really interesting, right? Because I’m going to be looking at something, I want to verify something, I want to validate something with a conversation and a dialogue. And a lot of it’s going to be dependent about where I am. So we’re trying to figure out how those intersect in the right way.

Stephanie:

Yeah. How do you approach it in a way that garners trust from the consumer? Because I feel like there’s been a lot of times, even me personally, if I know I’m talking to a computer, I’m like, “Nope, I’m good.” Because I’ve had so many bad experiences, or I’ve mentioned it a couple times, you call in to like Verizon, it’s like… Pretending to type, like… And you’re like, “This is so fake, and I don’t like this.” And then they can’t even help me either. So how do you go about it in a way that informs the person that you’re not really talking to a person right now, but still keeps them incentivized to want to try and work that method, if they’ve kind of been burned in the past?

Alex:

I think there’s two topics there. One, both of them, we have pretty strong opinions about. First, an AI should always identify itself as AI and not try to pretend it’s a human. I actually think we’ll see legislation on that. Because I think it’s bad. And I think a couple of things, even their practical issues with it too, you speak differently than you do to human than you do to an AI or a bot. And it’s better if you know for everybody, it’s actually better even for the AI builder to know that you know because the model is going to be different, interestingly. So there’s actually a real practical reason. But I think there’s a lot of ethical reasons, is I should know who I’m speaking to, and what I’m speaking to. That’s one.

Alex:

Two, I think that the gatekeepers, the Facebooks of the world, that want to kind of screen everyone and operate as these Uber, Uber marketplaces that they control the traffic flow, I think we as consumers are going to start to have very negative… I mean, we’ve started in terms of perception. But I think there’s going to be a continued backlash on that. And I think you want to know that if you’re into dealing with Verizon, you want to deal with Verizon, you want to deal with this company, you want to deal directly with that company. So, again, we feel really strongly about that, we don’t sell data, we work on behalf of the brand, there’s no targeting, there’s no selling data, there’s no advertising.

Alex:

So I think we’re going to see a return to kind of truer commercial relationships. Now, we’ve benefited, we’ve gotten all this free stuff by selling our behaviors, we don’t sell our data, by the way, we sell our behaviors, right? We sell behavioral changes to Facebooks and Googles of the world. And I think we’re going to stop doing that. And I think we have to stop, we have to pay for the things that would give us value, and then we know what we’re paying for and we understand how our data is going to be used. And I think that’s really, really important. And I think we’re going to see a shift over the next year or two, for sure. For sure, in terms of what people want.

Stephanie:

Yeah, so a lot of brands are probably hearing this and are nervous because of all the changes that are happening with Facebook and privacy rules. And many of them have been very reliant on search ads and Facebook. So what do you see customer acquisition looking like if you kind of can’t rely or maybe shouldn’t rely on those channels, and now people are maybe opting out of sharing all their data, even though it’s still pretty hard to opt out. It’s like, you either accept it or [crosstalk] look at a website. Like, “Okay, I guess I’ll just accept.” But how do you see it working for brands where it’s like, “Well, most of my traffic was coming from Facebook. And now that’s not really the world we’re going to be living in.” Is that kind of targeting and traffic and customer acquisition?

Alex:

Yeah, I mean, there’s still going to be these aggregate places, I think they’re not going to be eliminated. So our view is moving from a stream and target the stuff at me to enabling people to express their desires, their intents. And then businesses honestly, basically, applying for, “Hey, here’s something that based on what you’ve asked for, we may have.” And so I think the user acquisition, I don’t have the answer to, if I did, I’d probably start that company, or we would be doing it here at LivePerson, but I think that there will be kind of a flipping the model on its head a bit, right?

Alex:

So rather than this idea of, I can have some type of content, because I think the ills that we’ve seen have come from this model, I can have some kind of content, I want to get 50,000 people for a very, very, very low cost, I can go and target those 50,000 people, who don’t know about me and who I think have a proclivity to me, and I can go get them. So that was good for small business in some ways, right?

Stephanie:

Yeah.

Alex:

You can actually build businesses online. It’s bad for lots of other reasons. It’s also as a consumer, that you’re being introduced to something that… Like the whole serendipity introduction is neat when you’re on Etsy, because you know what it is [inaudible], it’s not neat when it could be anything. So I think we’re going to start to ask consumers to express their needs, and like, what do you want? What are you looking for? Can you define kind of your ecosystem of things that you like and appreciate? And then we’re going to ask your permission to actually bring others too. And you’re going to set standards, like, “I really only want to hear from companies that have certain social stances, or I only want to hear from companies that have certain environmental stances.”

Alex:

So I think it’s really all about empowering the consumer to kind of define, it’s a little more work, and I think that’s the thing that’s going to be interesting to see. Because I think we as consumers have gotten very lazy, it’s just like, “I want to scroll and you’re going to send me stuff.” I think we’re going to have to be asking consumers for a little bit more work to define those things and tell us more, so that we can give them things that are much more open, honest and transparent.n again, I don’t know what the format it’s going to look like, right now-

Stephanie:

I’m thinking of a whole new browser right now. Just need a whole new browser that operates in that way, because right now it’s like, where do you get all those ads and everything? It’s from your own Chrome, you’re on Safari. But it seems like you need a whole new world for it to operate in that way.

Alex:

So we’re actually experimenting with, call it a messenger, but I wouldn’t kind of categorize it that way, that it is intent driven. So you define, I’m looking for X, Y and Z, almost think of the kind of anti Alexa in some ways where it’s not just this transactional thing, I want to play this music and turn this thing on, it’s much more, I’m looking for this, I need this. And you understand the ecosystem of services and providers that actually can come together, all permission based, all about transparency. Early days, kind of experimenting and thinking it through, and talking to a lot of partners and companies also, because I don’t think we’re alone. I think many, many folks think there needs to be a change here, and we need to figure it out together.

Stephanie:

Yep. So we’ve had a debate on the show a couple of times about this whole trend of shopping on the edge, which to me seems like kind of where you guys are headed up, like being able to have conversations kind of wherever you are. How are you thinking about where people are shopping now? Do you see it moving to being on Instagram, being within Facebook Messenger, being on Tik Tok and being able to have those conversations there from the brand and selling on those platforms, and less about driving directly to one single website, or just on Amazon?

Alex:

So I do think this idea of the destination starts to fade. I do think that brands will be able to speak to you wherever you are, right? Again, I think it needs to be permission based. I think it needs to be based on your intent. But I do think brand… It’s funny. I mean, the idea that you don’t have a website sounds insane, right? If you’re a company, but what was a website 20 years ago? Nobody had a website really. The brand found you where you were, you saw the store, you looked in different magazines, you saw them on different television channels advertise. It was a much more organic process.

Alex:

And these gatekeepers have become very, very dominant. And again, I think if that changes where we’re not willing to give away our behaviors anymore, or sell our behaviors anymore, then I do think you’ll start to see brands engage in ways across all the places you live and breathe. Again, should be permission based, for sure. So I do think this shopping on the edge is kind of funny because isn’t that where we all did years ago?

Stephanie:

Yeah, but now we’re back. In the digital world though.

Alex:

Exactly. But that’s where we’re going to… The Fourth Industrial Revolution kind of back again is like, I think all these things start to blend together and we don’t want, we don’t want these kind of singular locations and gatekeepers, I think we’re going to start to see different properties have different purposes based on what we’re in the mood for, what we need. It’s interesting, I think the biggest thing that I would leave you with, and leave listeners to, is digital was convenient. It is now a necessity. There’s not more meaningful things that can shape change in terms of the format of an experience and the business models. And I don’t think we’re going to go back, I don’t think we’re going to go back to what was before, there’s going to be something new, and that [inaudible] is what’s really going to drive a lot of it. So I think you’re going to have to as a brand be where people are.

Stephanie:

Yeah, which sounds chaotic for me.

Alex:

And this idea that people are… It sounds chaotic, but I actually think it democratizes things, I actually think it means that we can eliminate some of these gatekeepers who make billions and billions of dollars on our behaviors, which I think would be a good thing in the world.

Stephanie:

Yeah, I agree. All right. Well, let’s shift over to the lightning round. The lightning round is brought to you by our friends at Salesforce Commerce Cloud. This is where I ask a question and you have 30 seconds or less to answer. Are you ready, Alex?

Alex:

Sure.

Stephanie:

All right. First one, what one thing will have the biggest impact on ecommerce in the next year?

Alex:

So I’m sorry to repeat, but the fact that digital is now a necessity, I think is going to have one of the biggest impacts, for sure.

Stephanie:

That’s all right. You’re allowed to repeat on this show. You can do whatever you want. Let’s see, what is something that you believe that many people don’t agree with you on?

Alex:

I’m hopeful. I do a lot of these podcasts, I get a lot of scary questions. I don’t go there, you’re not going to get me there. I think-

Stephanie:

[crosstalk]

Alex:

No. No, you didn’t. You didn’t. This has been fun and positive, which is great. I really enjoyed it. I’m hopeful about the future, I actually think AI is going to be a powerful tool of change, positive change. I don’t think it’s going to kill everyone’s jobs. I actually think we’re going to find new ways to make it augment and enhance us in ways we don’t even expect. So I guess in the AI space, in Big Tech space, I spend a lot of time talking, I hear a lot of fear and the sky is falling. And I guess I don’t think that way. I think I’m pretty uplifted and positive about what the future is to come.

Stephanie:

I love that. I’m on the same page. Normally-

Alex:

Good, yeah, I can get that.

Stephanie:

… [crosstalk] all that stuff. Yeah, you can get us on a space. What’s one thing you don’t understand today that you wish you did?

Alex:

I don’t understand, and I think about it all the time and debate it all the time, and I’m not going to go all political on you, I don’t understand the device of this right now, that we can’t find ways to communicate and talk and debate real issues to find solutions. We like to divide. And I’m kind of confused by it, to be honest.

Stephanie:

Do you ever just look back at your news and media days and be like, “That’s the stem of it, a lot of it.” Like the targeting and the way articles are written. Oh, man.

Alex:

Yeah, I don’t know. I think it’s easy to go blame the media. I’m not saying you’re doing that. I don’t know. I don’t know. Are we at peak Western civilization and there’s always a crest in your fall? Maybe, that could be it. And I’m part of that problem, probably. I don’t know. But I’m confused by it. It’s something- [crosstalk]

Stephanie:

I’d like to see that change. That would be nice to see, [inaudible] everyone just come together in love, like the yellow debate, in a friendly manner. That would be nice.

Alex:

Yeah. And I think we’ll get through it. So I am still positive about the future, I think. But I’m confused by the current state of it.

Stephanie:

Yeah. Yeah. Same. If you were to have a podcast, what would it be about? And who would your first guest be?

Alex:

If I were to have a podcast, it would probably be a little bit far from tech. It would be about how we bring magic back into our lives.

Stephanie:

Oh, I like that.

Alex:

Yes.

Stephanie:

You need the Alex Spinelli show.

Alex:

A little journey I’m on. Yeah, the little journey I’m on. I started going to Burning Man a number of years ago, and there’s just an infectiousness of [inaudible] in wonder and magic and art, and dancing in the desert, into your life. And I think more people need to dance in the desert at sunrise.

Stephanie:

I love that. That’s great. All right. And then the last thing, what’s the nicest thing anyone’s ever done for you?

Alex:

The nicest thing anyone’s ever done for me. I have a pretty amazing group of friends and family, so I got a lot of nice things done for me.

Stephanie:

You’re a lucky dude.

Alex:

I am lucky. I really do appreciate it. I think my wife marrying me is probably the nicest thing anyone’s ever done for me. It changed my life and it’s been wonderful.

Stephanie:

Go her.

Alex:

She’s amazing.

Stephanie:

We’ve had a couple of guys say that on the show, which is so sweet. I’m always like, “I hope your wife listens to this then.”

Alex:

I’m lucky. She’s amazing.

Stephanie:

That’s awesome. Well, Alex, it’s been such a pleasure having you on. Yeah, I love the conversation. Where can people find out more about you and LivePerson?

Alex:

Yeah, I think you start on liveperson.com. And there’s plenty on me on LinkedIn and our various social media. So I look forward to it.

Stephanie:

That sounds great. Thanks so much.

Alex:

Yeah. You got it. Thank you.

Subscribe in your favorite podcast app.

Love this? Share it with your friends!

Facebook
Google+
Twitter
LinkedIn

Up Next in Commerce is brought to you by Salesforce Commerce Cloud.

Respond quickly to changing customer needs with flexible ecommerce connected to marketing, sales, and service. Deliver intelligent commerce experiences your customers can trust, across every channel.

Together, we’re ready for what’s next in commerce. Learn more at salesforce.com/commerce

Subscribe to our newsletter.

Get exclusive updates and new episodes straight to your inbox. 

Subscribe Now To Get

– Our daily newsletter designed to increase your wealth, health, and wisdom.

– Access to exclusive giveaways from The Mission full of awesome swag and prizes.

Our Podcasts