Trust & Safety - AI's double edge and the future of platforms - Jeremy Gottschalk (Marketplace Risk)

Speaker A: Hello and welcome to a new episode of our podcast, Beyond Core. The podcast for business leaders who dare to think beyond today's success and build a future. As you know, in every episode we explore entrepreneurial strategies for unlocking new growth beyond your core business. From platforms, digital ecosystems, AI driven ventures, or circular business models, we dive deep with industry pioneers who are shaping what's next. If you are looking to scale innovation, create new markets and future, proof your company you're in the right place. And I'm not doing the podcast by myself. I also have my favorite co host here with me. Hi Nathalie.

Speaker B: Hi Matthias. Lovely to be here to talk about my favorite topic. As you know.

Speaker A: Yes, exactly. Today it's about trust and safety, your favorite topic. And we have a special guest here, it's Jeremy Gottschalk. Um, he released a new book, bulletproof your marketplace, and he's an expert since years about the whole topic about building trust safety within your marketplace, uh, and platform. Hello Jeremy. Hello.

Speaker C: Thank you for having me.

Speaker A: So today you are joining us from Florida. So it's a quite of a distance, but we want to shorten the distance and also let our guests know who you are. And um, why is this your favorite topic of uh, trust risk, buildings, safe environments for, for marketplaces? Maybe you can give a short intro to our uh, listeners and viewers who you are and yeah, why is this your favorite topic?

Speaker C: Absolutely. So, um, I'm a lawyer by training, sadly, um, and went, practiced IP law initially, which was um, super exciting, but then got into technology, started working with a lot of startups. Eventually, um, went in house, uh, into a child care and senior care marketplace right after they raised their series A. Uh, was the first lawyer tasked with, you know, all things liability and exposure and protecting the company. And um, one of my first questions was, you know, what are we doing to protect these kids? Um, I have a specialty in law. When I was in law school, um, uh, called pediatric law, so is representing abused and neglected children. I originally set out to save the world, but with, you know, too many student loans, I had to go into a law firm, uh, just to eat. So long story short, um, going in house was a perfect kind of marriage of my training and my interest, but also presented these huge challenges that um, you know, we find, uh, on marketplaces and the Internet generally. Um, and so it just became a passion, obviously protecting the company, but also figuring out how do you protect vulnerable populations, um, and people on the Internet, uh, in general. Um, and very quickly I realized it's really hard to do um, you know, the Internet's anonymous and there are bad people everywhere and if you're not careful, right. You could be the victim of them. So um, that's really the origin of um, the book Marketplace uh Risk. All of that is really to try and bring um, leaders in the industry, found everyone from the founders and the C level kind of all the way to customer service folks together to talk about um, kind of the lessons they've learned, the tactics and the strategies they've used, ah, in order to create a better, safer environment for essentially users of the Internet. My specific focus is on marketplaces. But um, there's so much transcends, you know, the Internet and E commerce in general.

Speaker A: Um, as you said it uh, brings also a lot of advantages but also attracts a lot of bad people and bad behavior. And um, it's a special topic especially for marketplaces to build this trustful environment. And you also mentioned Marketplace Risk, I noticed as a big conference, uh, or series of conferences where people can, can join and can exchange. Uh, maybe you can also bit of more tell us what, what you are doing. So you are consulting companies or marketplaces and platforms. So what is your, how do you help them at the moment?

Speaker C: Sure. So uh, Marketplace Risk started as a working group of founders and very senior um, leaders at uh, a bunch of different marketplaces, from child care to pet care to the ride Share, can you name it. Um, we started getting together quarterly just to learn from one another. The group grew organically as more and more people were hearing about this. It wasn't some clandestine or secret meeting, uh, but we just wanted to learn from one another. And um, so that evolved into a conference uh ten years ago now, uh, in 2015, um, and has grown and grown and grown, uh, and we realize then there, you know, there are people in um, Europe who are dealing with the same issues and then on the east coast and so fast forward, um, we host five conferences. Um, San Francisco, New York, Austin, um, London. And then this year we're also introducing Sao Paulo, Brazil. All focused on trust and safety, risk management. Really the, the stuff that you never learn in business school or that you know, VCs aren't focused on, um, but are critical to running these platforms. Of course we also have you know, blogs and podcasts and we do a lot of virtual content and, and we've got a lot of other um, kind of tactics and strategies to try and network and, and get the, the ecosystem collaborating. But our main focus are in person conferences where people can learn from one another and share and Collaborate over all of these topics.

Speaker B: No, actually that raises a small question, a small or big, I don't know question for me. Um, do you see a difference in the European markets and the American market on that topic?

Speaker C: Yeah, yeah, actually. So it's interesting because when you think of, I think the issues are the same, right? The Internet is the same no matter where you are. I think one of the big differences is um, how many platforms are headquartered there or launched there because you get that kind of mind meld and that knowledge sharing. Um, and one of the things that we found in San Francisco, which is kind of still the epicenter of technology, is you've got all of these companies willing to come together and share. Um, but the reality is what we found very quickly is traveling to San Francisco, it's difficult, it's really expensive and if you're working at a startup, the chances of you having resources to fly to San Francisco and stay there, I mean you lose a vacation quite literally really if you end up going to the conference. Um, and so that's why we um, spread out into different geographies. And there are particularly in Europe, right. The EU is very active and has been way more active than any other government on kind of protecting the citizens and really focusing on privacy and um, a lot of just kind of platform related issues. Um, and then of course the US follows slowly but is starting to look at what's happening in the EU and kind of replicate that. So there are a lot of issues that are different. I think the core trust and safety issues are the same but how they're being addressed. Um, one other thing I think that differentiates, um, all the different um, geographies is the size of the market I think will often dictate what solution providers are trying to sell there. Um, and so for example in the US if you look at digital identity providers or if you look at background screening companies, um, just because of the size of the market, um, you'll get you know, a ton. Um, and when you go to other markets, particularly if there's a different language spoken, sometimes they're not addressed at all. Sometimes the solution providers aren't even focused, focused there. Um, and so there's a little bit of a, a lack of uh, variety I guess you could say that we've seen, and we've seen a lot of solution providers, particularly in the fraud prevention identity, uh, space starting to exp, um, you know, into Latin America, which is, you know, Brazil of course, huge market. Um, but still it's um, it's. They're not really being served as much as kind of the US Market and I think you know the, the parts of parts of Europe I think are, are, are much better served now. But um, in the beginning it was you know the, the solution providers I think were very targeted on particularly if they were U S Based right. They were going to English speaking countries and that was pretty much it. I think they're now really gaining traction and so that's not um, it's not the same as it was before. It's, it's improving a lot. But but um, yeah, I think that the, the problems are, are basically the same. It's the solutions I think that, that vary whether it's regulation or solution providers and vendors, things like that.

Speaker A: Um, to make it also a bit more tangible for our, for our listeners. Maybe you can, and you have also a, a chapter in, inside of your book, um, looking back and looking at some failures. So maybe you can describe a bit or you can uh, uh, present some cases where you think this, those are kind of obvious failures of marketplaces. And this underlines why it is so important to have a look at the whole topic of trust and safety. So why is this important?

Speaker C: Uh, yeah, I mean if I can go one step backwards. The thing about trust and safety, I think um, 10 years ago your everyday user, uh particularly even if they had an iPhone or a smartphone didn't really think about trust and safety. It was more um, coming from the companies themselves. Trust is the biggest barrier to conversion. So you've got this funnel of potential um, users uh, who you want to become customers and in order to convert them they to trust. Um, and then of course once they're, they trust you and they convert to paying customers. Now you've got to keep them safe, right? So there's that component kind of getting them in the door and then keeping them safe. Um, I talk about in the book there's kind of two phases that I, when I look at the Internet it's the first 20 years and kind of uh, since then, right. The, the last 20 or so. In the beginning, um, there was almost no attention paid by governments, regulators, pretty much anybody. Everyone kind of buried their head in the sand when it came to um, the Internet in general and I think just kind of pretended nothing bad was happening. Right. I mean you, you would get uh, you know, you'd talk or, or hear jokes about the AOL email scams from princes. As long as you gave them their bank account information, they'd send you all of this money. But that was kind of the extent of it and the, um, the fraud and kind of all of the bad stuff that has evolved, um, since then has gotten exponentially worse. Um, and, uh, you also kind of simultaneously have governments, like, they missed the boat the first time around. And so now they're really paying attention. And they're, some would argue over regulating, but at least they're paying attention to a lot of these issues. In the beginning, though, because there was really no government regulation, because nobody was really paying attention, the focus was purely on kind of growth at all costs, right? VCs give a lot of money. They, uh, invest in these companies and they want huge returns, period. The problem is when something bad does happen, particularly if it gets media attention, um, you know, Everybody from the VCs to the executives, all the way down to your customers are freaking out thinking, how can this happen? And it frankly stains not only the reputation of that business, but the entire industry, the entire ecosystem, right? If you take dating apps, for example, if there's some, if someone's raped, you know, through somebody they met on a dating app, everybody lumps all dating apps together, regardless of what they're doing, with trust and safety, etc. And so in the beginning there was just simply nothing being done. And when I say nothing at the core, or I should say at the very end, uh, of spectrum of kind of trust and safety, I think about like, digital identity, right? Identifying your users. Because if you don't know who your users are, um, then you can't screen them, you can't do anything, um, to really monitor or prevent them because you don't know who they are. Also, people who aren't using their proper identity are probably there for some nefarious purpose or another, right? Or they're hiding something, something. And so when I say nothing was happening at the beginning, digital identity, when I went in house, wasn't the thing. The best they would, you know, the best solutions might confirm or validate emails or phone numbers, but we very quickly realized that that was insufficient. And the technology to get around that is just way too easy. And then of course you kick somebody off and they just get a new email and then you're kind of in the cycle. Um, and so I think that you had a lot of money being invested in these startups all focused on growth. They, a lot of these, um, apps and these, uh, platforms became kind of ubiquitous. Everyone was using them. And then when things went bad or horribly wrong, the media paid attention. And, um, then we were all left thinking, like, what is happening? What are they Doing like, how are they going to prevent this? The reality is you can't prevent all bad things from happening. You can't prevent it in the real world, let alone on the Internet. Right. And you've added all these, um, anonymous variables on the Internet. Um, and so, you know, I've seen kind of in my tenure, it's, it's not so much that bad things happen, it's more how the platforms react when they do happen. Um, and you know, there were some really, um, uh, I will call them famous or infamous cases, um, you know, around the two early 2000s with some of the big, the rideshare platforms and the home state platforms where something bad happened. One, you know, one, um, Uber had a famous, um, incident where it was an accident, but a driver, um, turning right in San Francisco didn't see this little girl and ran her over and she died as a result of it. Um, and you know, it was horrible. And I think that everybody kind of recognized that it was, it was an accident. Um, there was nothing, no foul play, none of that. But, um, the response was um, kind of less than adequate. And it didn't really address kind of the, I think the nature of, um, or the gravity of, of what had happened. Right. Somebody had lost their life. And um, the public response was extremely defensive and basically jumped to the fact that they don't have any liability because of the Communications Decency act, which is a federal law in the US that protects platforms from liability. But, um, it was, you know, tone deaf at best. Um, and frankly didn't, you know, did just horrible disservice to platforms in general because, um, you know, when, when there's been a loss of life, nobody wants to hear your about your defenses.

Speaker B: Right.

Speaker C: They need a little compassion. Um, and you know, then Airbnb had one where um, this woman, uh, rented her really high end apartment. And I don't remember college kids or who it was, but long story short is they trashed it to the tune of like 50 or $60,000 in damage. And the kind of the same, um, tone deaf response in that we're not responsible. Um, you know, we're, we're, we're simply a, uh, passive medium, et cetera. Um, and this happened over and over. And so I think there was kind of a combination of, um, the how platforms responded, but also the inability to talk about what they were actually doing when it came to trust and safety. Um, because there wasn't a lot happening back in the day. Right. Um, and that has evolved greatly. I, um, think you look at any of the, the certainly the larger platforms and many of them publish trust and safety reports. They have um, you know, trust and safety centers, FAQs, et cetera, et cetera. And they're very clear what they do do and in fact they are doing a heck of a lot more um, ah, and spending a lot of money and a lot of resources trying to keep their um, keep the experience good but also um, keep their users safe. Uh, more or less.

Speaker A: Yeah, I think those are great examples and shows the magnitude uh, it can have uh, destroying your reputation or whatever. Um, maybe also slowing down your growth or network effects and the network creating negative network effects so that people leaving because they see this and then more people leave and so you can also have negative network effects and then it, yeah, the people and the users run away from your platform and maybe join a different ones. Um, so uh, you also said that platforms learned a lot so they implemented a lot of mechanisms starting with identity, uh, checks or oversee communication and stuff like that. Um, now we see maybe the next phase with AI. Uh, we see AI is coming up and with AI you can do a lot of uh, good stuff but also a lot of more bad stuff. So the bad people are also jumping on this train and um, with deep fakes and stuff like that. So how do you see the double edged sword of AI joining this big picture maybe from a very high level point of view and then we can maybe zoom in a little bit more on how can platforms also leverage AI to build and continue to build trust?

Speaker C: Yeah, so I, you know I think I'm not the only person who said this but AI is probably one of the most revolutionary technologies since the Internet came out. Right. Um, it has the ability to shorten um, development times, um, you know, customer service responses, you name it. And there's, there's a valid, very productive use of AI And I think we're just scratching the surface. Um, we're, we're just learning the, the um, capacity and it's, I think it's as we learn it's going to continue to kind of outpace us in any event. Um, the other side of that though is you know, online crime, fraud, um, it's organized. And so you think about organized crime as um, you know, mafiosos in a New York pizzeria right at 3:00am um, it's, that's not the case anymore. Organized crime has gone online and they have the same tools, if not better than you and I have access to, to commit these crimes, um, to the point that there Are AI tools being developed specifically for uh, the purpose of committing um, fraud, um, breaking passwords, I mean you name it, right? You've got a tool that can um, make a, any sort of a process, um, exponentially more efficient. Right? And so there's a positive side to that. But also the fraudsters, they can use those tools as well to crack passwords faster, to um, you know, commit uh, whatever, whatever crime they're, they're seeking to commit that much faster. Um, and so I think some of the obvious misuses, um, we'll call them of AI or like the deep fakes, um, like the synthetic voices and things like that where you can quite literally clone someone's voice, uh, or their image, um, to the point that, that you or I, none of us would, would probably be able to tell that it wasn't our grandmother or our sister who was leaving us a voicemail or you know, on zoom with us, um, which is super scary because it, it leads people to be extremely vulnerable. Um, and um, it's almost impossible to detect from I would say your everyday user or consumer. Obviously they're. Now we're also developing tools to detect when images are AI generated and when videos are AI generated and sounds and things like that. But um, I think the, the high level point here is there are a lot of really good uses for AI, but there are a lot of really bad people who are simultaneously leveraging those tools for nefarious purposes. And they are way ahead of, I think the good guys and they are way ahead of the governments and they are completely outpacing law enforcement. Um, and so we're at a point where we really have to learn quickly, get up to speed quickly to understand because I don't think we're ever going to stop them. I think the best we'll be able to do is detect it, um, and um, you know, prevent people from becoming victims. But I don't think the tools, um, or their sophistication is going anywhere.

Speaker B: So do you think you, you mentioned, um, I mean, sorry, I'm going back a little bit on the previous topic, but it's linked, um, you mentioned responsibility and, and of the platforms and that platforms were initially pushing back, pushing away responsibility and I for one was always annoyed by that and that really threw a really bad image on platforms and marketplaces. Fast forward, um, today. Do you think platforms are doing enough on um, this topic of responsibility, taking ownership of risk of safety, et cetera?

Speaker C: Yeah. Um, so if AI is a double edged sword, this is like a triple or quadruple edged sword. Um, so a couple things. One is, I think just from a um, customer retention perspective.

Speaker A: Don't.

Speaker C: I'm not even referring to the law or legal liability, but just from a customer retention acquisition and retention perspective. Um, I think now your users are expecting more from the platforms. And so I think that accepting responsibility, um, not washing your hands from every incident that happens, not necessarily taking ownership of it. Because the reality is, is these platforms don't have control over what users do between and among themselves. Um, there's certainly stuff that they can do, but at the end of the day if somebody's going to try and defraud somebody else, they're, they're um. The platform likely won't have a lot of control over that. That said, um, you know, users want to transact on in safe environments and the platforms do have visibility into communications activity, et cetera. And so I um, think that the, M. The, the. The platforms that you and I want to be on are taking more responsibility, they're taking ownership, they're communicating with you more both about the risks and how to prevent them. M. They're not um, pretending that bad things aren't happening. They're. They're uh, being very forward about the fact that you know, you've got to look out for certain behaviors and protect yourself, et cetera. In some cases you'll see platforms that will, you know, guarantee transactions up to a certain level to give you some, some comfort. It's not necessarily, you know, legally that's not their obligation, but it's a really good way to instill that trust and, and retain those customers. Um, you know, the question about are they doing enough? Is so relative. It's so hard to answer because when I think of enough there's. There are a lot of things that platforms can do that doesn't require even a lot of technology or resources. And I think that um, you know, monitoring communications, monitoring activity, you know, especially these platforms that around for um, any amount of uh, time they have models and they know what normal user behavior is. Right. They can map it out and there's some outliers but um, they can also then detect anomalies or abnormal behavior. And there's a certain, you know, each platform is going to have normal use cases for whatever the features and the functionality are. And when you see um, or you detect um, abnormal activity, then that's an opportunity to look into that and, and to try and intercept whatever's happening. And, and because they have into communications and all of that, um, there are some really low Tech ways to um, uh, protect your users and potentially prevent. Now depending on the scale, right, you're going to have to have a lot of human uh, intervention. And now obviously with AI you've got a lot, you have tools that can also monitor and then raise those issues to a human when it makes sense. Um, so those are kind of the low tech ways, some human intervention, but easy to have a, I would say like a good vantage point over the whole platform and the whole ecosystem. And again it's data driven. So this isn't um, a subjective kind of somebody sitting in a room watching people's online behavior. It's just, it's purely you know, kind of data driven. And then anomalies are, are surfaced for, for um, review. But then you've got a lot of um, what I would call more expensive tools. Right? Um, and so as I mentioned, you know, digital identity is one of those things that um, I think is what I, the cornerstone of trust and safety. If you don't know your user is who they say they are, um, you're already on your back foot. Uh, conversely if you do know who your user is and your user is who they say they are, there's a really good chance that they're probably not going to do anything wrong. People don't use their own identities and commit crimes online generally speaking. Now people will do dumb things, don't get me wrong but they don't set out to commit crimes and they're probably not in any sort of organized criminal um, enterprise using their own identification. So the ability to properly identify your users is key. Um, and I think is, is the kind of the beginning of um, the solution. A lot of those tools, there's a lot, there's a variety of tools or a variety of signals that you can monitor and confirm and validate etc, and it gets, it gets expensive pretty quickly, right, because you're, you're using um, you're, you know this is all outsourced vendors and solution providers and even if you're doing this at scale it's just, it can get expensive. It can also create barriers, um, and you know, slow uh, down onboarding which any you know any technologist is going to tell you is kind of the, the um, the worst possible thing you can do. So there's kind of this balance if you're moving on from there, there's all types of screening solutions, et cetera that they get expensive. Right? So it becomes this balance of how much um, how much uh, want to invest, how much money do we want to Invest how many roadblocks, how, how much friction do we want to add to, you know, our onboarding and then how frequently do we want to, to do that? Um and so when, when the question is asked are they doing enough? Keep in mind if they spent all of their money uh, screening and leveraging all of these tools, very likely they would go bankrupt because they're probably not making enough uh, revenue per user to, to um, to substantiate that. But on the other hand there are a lot of even low tech ways to um, uh, be you know, uh, curating but also um, ensuring that or doing the best you can to make sure that the ecosystem M is healthy and that those bad guys don't get very far. Um, or if they do something that they're gone and you know, you, you keep them out. Um, so that's why I say, you know there's, we owe a lot to um, particularly in the US the CDA which protects platforms. We owe a lot to that federal law that says the platform doesn't have any liability for what happens between its users, more or less. Um, and so we've seen a proliferation of platforms and social media and marketplaces because of that. Um, but I can tell you that in the U.S. most legislators, both sides of the aisle, regulators, they hate it because they think that that big tech has gotten away with too much for too long. Um, and I don't disagree with that. I think that, that there were, there was a lot of irresponsibility that has made that pendulum swing. Um, and so um, there's not only a lot that can be done, there's certainly more that can be done. But it's kind of a balance, um, you know, obviously for the, depending on the, the, the type of platform and, and kind of what their, their features and functionalities are.

Speaker B: It's really interesting because there are some layers, uh, maybe more of safety than trust that uh, indeed you could um, either implement with technology or regulate with um, regulation. Um, and then the trust is much more emotional and harder to grasp. Um, and yet maybe that's also very powerful. I mean uh, you were talking about non or low tech, um, um, elements that can build trust on a platform. It reminds me of an article I wrote quite uh, a while ago in Forbes on the difference and as a user and a professional of platform platforms I had that look, um, when I was uh, buy, I bought a piece of furniture on one E Commerce site and the furniture arrived with damage and the platform said I don't care, it's not my Responsibility versus another one. The same thing happened and the other one said look, it's the responsibility of the vendor. However I'm staying with you and I'm making sure it's going to be sorted it and for me that made all the difference and I stopped going on the first one. And so it's a perception and I guess it doesn't cost them much to have a three way chat with the vendor and me just chasing the vendor, uh, saying make sure you reply. But it's so powerful in terms of trust.

Speaker C: Yeah, you see this with Amazon, particularly in the us um, their, their refund policy, you know almost. And I know and I know it gets abused so I, I, I, um, I think that maybe they're a little too generous. But you know I just from myself, um, and I study them because I think that um, they, they kind of get it in the sense that their two thirds of their revenue comes from their marketplace. Um, and regardless of the source of the goods that you purchase, whether it's from an Amazon warehouse or from a third party seller, they treat it the same. Um, and you know, as the customer, it's kind of like the customer's always right. Um, will, you know they generally obviously issue a credit or, but it's, it's immediate and um, there's, I've never once and I've never seen it where Amazon kind of washes their hands of um, the issue. Um, interestingly uh, again in the US you see when you buy something from a third party seller, very often they will include in the package if there's any issue with this, contact us directly and we will make it right immediately. I'm sure on the back end when Amazon has to issue a refund. Right. They're not getting paid as well. And so it's, it's in their best interest as well. But um, they've done a really good job at least here of um, keeping that trust. And as a result, you know I'm part of the millions of people who, when I'm looking for something I just go directly to Amazon because I feel like it's going to be protected no matter, it doesn't matter if it's coming from their warehouse or um, from a, ah, one of the third party sellers

Speaker B: and they set a precedent because then again I remember one day I ordered something and I wanted to return it and the guy said to me look, I'm not Amazon so I never bought anything again, I didn't have that trust.

Speaker C: Right, right. And that's the risk. Right. I Mean that's the risk. Especially when you're doing things, you can't touch it, feel it, right? You can only see it, it's, it's two dimensional. And so there is a much higher trust barrier to buying something. But I will, I don't think twice about buying anything on Amazon because even if I don't like it, and this is probably the excess point, but even if I don't like it, you can send it back and, and you know the, the risk, the answer because you often have, or the response you often have to provide a response to what's wrong is it didn't work for me. And, and so as a result they, I don't know how many thousands of dollars a year I give Amazon. But um, you know, that's the trust that they've built and um, that they continue you.

Speaker A: Yeah, also, also looking a bit into the future and looking ahead. Um, I mean trust, you described it as something intangible, but it, it is built over time, right? Slowly over time. And uh, it can leave via a Ferrari or Lamborghini very fast. Uh, uh, but uh, you build it over time. So a bit of intangible. But, but on the other side, if you are an owner of a platform or marketplace, you still ask yourself it's intangible, but what can I exactly do to build this trust? So um, we also ask ourselves this question and we came back with a kind of a uh, concept which is relying on a uh, concept famously known as the separation of powers. Maybe you know, from, from, from a state. We also ask ourselves is there something similar? And when you compare, when you start thinking how can I build trust in a, in a state, right? If I'm the usa, if I'm Germany, if I'm France or whatever, then when you look at those modern states you have this separation of power and you distinguish between legislative execution, uh, utical powers etc and uh, so we thought, thought this could be also something like a framework or a concept you can transfer to the whole topic of trust. And when you think about it, and you described it uh, with the two, or Natalie described it with the two cases here, with the two different marketplaces, it first starts with your kind of legislative thing and power, what kind of values you have? Um, um, as a marketplace, right? So you compared Amazon, which has a value like, like customer first, right? I don't care about the supplier, I more uh, care about the buyers and I do everything what is in the interest of a buyer and instead of the supplier. And maybe another marketplace is more like balanced or more on the side of the, of the supplier and whatever. So it starts with the top what kind of values you, you have and then it ripples down to the other layers of like what kind of laws I want to implement, what is the rule on my marketplace M, and then also what kind of trust building mechanisms I want to build. So technical things like an identity check or uh, with the chat or whatever, supervising communication or doing kind of checks on the product itself or whatever else is required. So you can also maybe separate. Um, long story, this is a kind of a framework we use when we work with clients. Is there anything, what do you think now with mixing in the AI topic into that? Uh, is there anything what you think is required or a M marketplace owner can do to build up this trust in the future? So what is what becomes more and more important in the future?

Speaker C: Yeah, so, um, I think that where there's a huge misstep with AI, um, is part. And where the customer, I think engages with the platform. Um, from a customer service perspective, um, you know, there's nothing worse than um, these kind of useless bots that go in circles, right? And you can't get your answer solved, you can't get the problem resolved. Um, because that's the, the quickest way to ruin uh, any sort of trust that you've built is to not be accessible for a customer not to have their, their problem resolved. And, and, and the, when I talk about a problem, right, it could be something as simple as I have a question about a product or I got a product or I interacted with this person and it went horribly wrong, right? And in either case, um, if I'm not sure how to use this product or how to assemble it, um, you know, I'm sitting there with the box open and I need an answer and if I can't get a hold of somebody and I've got this bot and going in a circle and you know, they're trying to, trying to make everything easier and more efficient and, and eliminate the number of, of head count. Um, I'm frustrated and so immediately I'm going to think twice about whether or not I shop there. But also, um, that simple question now kind of escalated into a problem because I've got no, I've got no response. And I'm sitting here and waiting on the other end of the spectrum something bad happens right there. The package was broken or there you were shipped something that you weren't, or you were introduced to a caregiver who stole or blah blah, blah, Whatever the issue is whether it's good or services or any type of platform, if you can't access a human being in your time of need and, and whatever that is, it's, it tends to be very critical in the moment, right? Years later we look back and say that wasn't that serious. But in that moment when you, the thing came broken or you had a horrible experience, that's what you want resolved. And that trust evaporates in that moment if you can't access anyone. So the one thing I, that I fear with AI is, is the disappearance, um, of humans, um, and the kind of inability to um, provide really good customer service. Um, having said this, I'm talking about, I think outliers. The reality is I think AI can help answer a lot of questions very quickly and it actually speeds up the response times. And so there's a lot of great, um, aspects to it that I think will probably solve 90 or 95% of the customer service interactions. But it's the, those ones that don't get resolved or um, you know, you're, you're in this like uh, bot doom loop of trying to get, you know, some prefabricated response, uh, and you're not getting what you need. And so you just get frustrated and you know, very quickly it's like, I don't want to deal with that anymore. I don't. You know, there, as you mentioned, Natalie, there's so many um, options and there's so many competitors that if I can talk to or I know that somebody's going to be um, uh, on the other end of that line trying to resolve the problem, that's where I'm going to go. Um, so I think that that's kind of the low end of the AI kind of risk. Um, I just see it being leveraged, um, over leveraged, um, and the customer kind of being um, not uh, the focus anymore. It's more about the efficiencies internally, um, than anything.

Speaker A: So it's also the, you don't need to, or you, you need to avoid to lose uh, sight on the basics. Right? So this is the basic, is uh, response quickness and stuff like that. But um, a lot of people try to implement AI and doing defensive stuff, but then maybe you lose sight of the, of the basics and, but the basics are the basics and they are responsible for 80, 90% of building the trust. So you need to keep them running and working. Um, another opportunity of AI is also personalization. So we see the biggest opportunity for a marketplace in implementing is not just streamlining operations, but it's also in personalizing the experience for the users on the platform. Could be the buyer side, the seller side, but also the, the partner side. Um. Ah, are there any points from your point of view how to build um, trust on the same. In the same way or in the balance in building up personalization or personal experience?

Speaker C: Yeah, so I think that the personalization is uh, it's great right, Because I think what ends up happening is a user doesn't uh, necessarily even realize that they're talking to a machine. Right. Because um, it's that good. I think that um, monitoring it. Right. And, and really paying attention to the results and the, the. The user experience is key because I think that um, a lot of these tools are really, really good, but they're not um, infallible. And um, there's a real danger with implementing some of these tools and then just letting them go. Um, I think we've seen tons of examples of like these AI hallucinations where this, what you think is a personal experience ends up.

Speaker A: Um.

Speaker C: You know there's examples of this tools uh, convincing some trying to convince someone to commit suicide to take these different drugs that were um. Uh it just hallucinated and basically gave them um, a recipe that would have created I don't remember exactly what the. But some, some toxic drug. Right. Um, and so unchecked I think there's you, you very quickly, especially if you become like the, the. About an AI tool gone wrong. Um, I think that there's, there's um, the opportunity for it to go south very very quickly and it's never going to be perfect. Right. So there's always a certain amount of um, kind of monitoring this and testing it and evolving the tool particularly in a specific environment, um, where you know you're collecting data and you're refining it and, and the model and all of that.

Speaker A: Yeah. So it's supervising the. Not the human but the AI um, could be also that it makes wrong promises about the product as you ask about the product and then it kind of has hallucinating features which are not there. Then it promises you oh there is a feature and then you make a decision and you buy but. And then the product or whatever arrives. Then it's not the feature and there's no, not this feature. And so uh. But you are then responsible. Possible because you are the owner of the platform so.

Speaker C: That's right, that's right, that's right. That's actually just uh. On that point, one of the big questions particularly in the U.S. right. Because we have the CDA section 230 which basically absolves platforms from liability for what happens between their users. Right. So you, let's say you're, you're um, you're operating this platform, you're leveraging this tool. And uh, uh, you know, one user buys from another. Right. And this tool is in the middle. The middle and does exactly that. Right. Sells a product that is uh, not intended for that use or something worse happens. You know, the question then becomes uh, and I think in, in probably most jurisdictions the, the platform has liability. In the US it's still an open question. I mean I would argue that there is because the, the platform has integrated this tool and the tool and they're responsible for the tool. But we have this federal law that says what happens between the users is, is not the responsibility of the platform. So it's a really interesting question that, that we're watching play out to see how these tools will impact the legal liability of these platforms.

Speaker A: Yeah, I think very, very good points. Um, coming maybe to the, to the, to the closure of our uh, podcast and I think there is a lot of more stuff inside of your book. Uh, and um, we can only recommend all the marketplace owners and platform enthusiasts to have a look at, at the, at the. And the book Bulletproof, your Marketplace. Right. Is the title now available on Amazon and, and all the other well known platforms uh, uh, uh, for purchase. And um, your. When is the next uh, Marketplace Risk conference coming up? Is there next week?

Speaker C: Actually, uh, it's uh, May 12th through 14th in San Francisco. Um, and then September in New York, October in Sao Paulo, November in London.

Speaker A: London. Perfect. Perfect. And maybe um, looking back at your, at your career and maybe the last years. So we always ask also our guests, um, to share some kind of major learnings or major advice to other leaders. So is there any kind of bigger learning, bigger, bigger theme or narrative you would like to share with other leaders which kind of guided you through the last years and maybe is also a guidance exciting theme for you for the next years.

Speaker C: So yeah, with me personally it's always been um, kind of who are you doing this for? And why and staying focused. And, and I will tell you, we built this entire platform and these conferences specifically for these platform leaders to uh, to, to network and to learn and to collaborate. That's really our focus. Um, and over the years there's been a lot of temptation because there's a, there are a lot of um, within the ecosystem there are a lot of kind of Outside forces that will offer you a lot of money, um, to do a lot of things. And um, we've, you know, it's, and it's tempting, right, because at the end of the day, like who doesn't want to make money? But the reality is, um, we've stayed very focused on who our target is and why we're doing this. Um, and you know, so I always tell people, I always tell our team, you know, whenever you think about any sort of an addition to a conference, any new location, um, any sort of whatever it has to do with the app, whatever it has to do with the events, you've got to think about our target. How is it going to benefit them? Um, and you know, is it something that, that they're going to benefit from? And if not, we need to evaluate that because, um, that's why we're here and that's what we do. Um, and you know, I guess the overarching theme would be kind of that discipline, but really, really staying disciplined and think. I, I, and I, I say that from my perspective, but even with clients and when I consult, one of my first questions is like, who's your target here? Who, who are you going after? Because what I see a lot of times is trying to go after too many, um, at the same time and it just doesn't work. Um, and so that discipline I think is, is, is really useful, especially in the early days when it's really easy to get distracted, um, uh, and kind of lose your eye off the point all.

Speaker A: So it's know your customer and know your target group and then focus, focus, focus. I think. Very good, very good advice. Um, thank you very much Jeremy for joining and sharing your, your thoughts and your insights on this whole topic. I mean we scratched the surface and there's a lot of more stuff to um, to digest. Um, so we recommend buying the book, um, and uh, reading through it and or uh, joining you on your conferences and going into exchange with other platform leaders. So thank you very much from, from my side, uh, for joining.

Speaker C: Thank you very much. Thank you. It's great to be here.

Speaker B: Thank you very much, Jeremy. It was great.

Speaker A: My favorite topic or linking with Natalie. We will also put links to the book and conferences in the show notes for everybody who wants to uh, uh, uh, yeah. Learn more can click on the links in the show notes.

Speaker C: Excellent. Thank you.

Speaker B: Thank you.