CEO & Co-Founder, SeatGeek
Squawk Pod: Kalshi’s ‘Midterms Hub’ & SeatGeek & Serval's new I.T. dept. - 07/23/26 | Audio Only
Former CFTC Commissioner Brian Quintenz is now on the board of Kalshi Exchange and a senior advisor to the Coalition for Prediction Markets. The morning after Kalshi unveiled a “Midterms Hub” for election contracts, Quintenz discusses prediction markets regulation and impact on U.S. elections. AI-native IT startup Serval is helping customers like SeatGeek bring IT engineers off of IT help desks and into business units, where they can focus less on support tasks and more on big picture problem solving for the organization. Serval CEO and founder Jake Stauch and SeatGeek CEO Jack Groetzinger dis...
About Jack Groetzinger
Jack Groetzinger, co-founder and CEO of SeatGeek, has been a vocal critic of Ticketmaster and its parent company Live Nation, arguing that the company maintains a monopoly in the live event ticketing industry. In a December 2024 appearance on CNBC's "Squawk Box," Groetzinger said that the Department of Justice's lawsuit against Live Nation was warranted, stating that SeatGeek has had "a front row seat to the illegal behavior." He claimed that Ticketmaster controls roughly 80% of the market and that Live Nation uses its control over concert promotion to pressure venues into using Ticketmaster, saying "venues fear losing Live Nation concerts if they don't use Ticketmaster." Groetzinger has called for structural remedies, specifically the dissolution of the common ownership of Ticketmaster and Live Nation, arguing that "behavioral remedies cannot solve the problems inherent in an anti-competitive merger." Groetzinger has also advocated for greater price transparency in the ticketing industry. In December 2023, he participated in a White House announcement where ticketing companies committed to showing all-in pricing upfront, saying "we are major believers in transparency and knowing what you will pay." He has described SeatGeek as a technology company focused on improving the fan experience, noting that the company has signed dozens of major venues as clients by building better technology. Regarding ticket prices, Groetzinger has said that high prices are driven by supply and demand, noting that for events like Taylor Swift concerts, "there's probably 10x or 20x more people who would want to see her perform live versus can actually go." He has also expressed skepticism about using blockchain for ticketing, arguing that live events are inherently centralized because venues ultimately control access.
Transcript (109 segments)
K
Katie Kramer0:04
Hi, I'm CNBC producer Katie Kramer. Today on Squawk Pod, our tech future: first, prediction markets and the midterms. Khi has rolled out a new hub for elections contracts. Former CFTC commissioner and Khi Exchange board member Brian Quinten's on the government pushback on pre. These are new markets, so I don't think it's a surprise that states are grappling with a lot of these issues from what the laws are that have been on their books for a long time. Then the future of work with an AI startup trying to disrupt the IT department. Servil CEO Jake Stout.
J
Jake Stout0:40
Instead of filing a ticket and waiting for somebody to respond to that, we're actually going to automate that resolution with AI. Get you back to the more meaningful parts of your job.
K
Katie Kramer0:48
And CEO of SeatGeek, Jack Groetzinger, a customer of Servil and a believer in a more efficient use of your IT guy.
J
Jack Groetzinger0:55
We see this as a way to move people from less interested work to more interesting work, get them focused on harder problems.
K
Katie Kramer1:01
Those conversations plus the question of course is: is a reframe of this the answer unto itself? I'm of two minds when it comes to AI, too. It is Thursday, July 23rd, 2026. Squawk Pod begins right now. Stand back by in three, two, one, queue up, please.
B
Becky Quick1:20
Good morning everyone. Welcome to Squawkbox right here on CNBC. We are live from the NASDAQ market site in Times Square. I'm Becky Quick along with Andrew Ross Sorkin. And Joe is off today.
A
Andrew Ross Sorkin1:31
Meanwhile, shares of Google parent Alphabet. They are lower in the pre-market this after the company reported second quarter revenue above expectations. 82% growth in Google's cloud business, but overall profit fell short of estimates. Investors are focusing on an increased capex forecast as the company invests in AI. Alphabet now expects to spend between $195 and $25 billion just this year. That's up from a prior forecast of 180 to $190 billion. Let's show you what CEO Sundar Pichai said on the company's earnings call.
S
Sundar Pichai2:07
I do think, you know, it feels like we are in very early innings of what feels like secular shift across multiple areas in our core information businesses. Just the possibilities when I see what all you can do with the absolute frontier capabilities, there's still a lot of work ahead to translate all that into experiences for our consumer users.
A
Andrew Ross Sorkin2:36
Another Alphabet earnings highlights: search revenue growing 17%, and the company's other income category which includes stakes in Anthropic and SpaceX coming in just shy of $100 billion. What's so interesting to me about the reaction though in the market right now about the Google piece is just, you know, every time these companies have spent more money recently, you've seen the stocks go up meaning investors have rewarded what's been happening because they've said 'oh this is the future.' This is really the first time where it seems like at least on the margins investors go 'I don't know.'
B
Becky Quick3:11
And it's not just that it's Tesla as well. We may talk about that stock a little bit later. It's the first time they've been cash spent that they've put into bigger projects that they're working on longer term issues. That's the thing you watch for: when the street rewards this long-term payoff saying 'we want you to chase it, we want you to win' versus when the street kind of pulls back and says 'whatever.' That is the headline that's basically been taken away from every single take on this. The AI spending balloons a little different than what we've seen at least in recent years.
A
Andrew Ross Sorkin3:42
We should also tell you Iran's military says that it has hit US targets in Kuwait and Jordan. It also says ships attempting to sail through the Strait of Hormuz are still being attacked. In addition, Iran-aligned Houthis say that they struck two Saudi Arabian oil tankers in the Red Sea. At least one of those tanker attacks has been confirmed by the British military. Attention has turned to the Bab el-Mandeb Strait as the US and Iran fight over the Strait of Hormuz. Yesterday afternoon, US Centcom said that American forces launched more strikes against Iranian military targets. Separately, the Wall Street Journal reporting the United States is moving forces and weaponry to the Middle East in case President Trump wants to scale up American attacks against Iran. Yesterday on social media, the president said that anytime Iran fires on a ship in the Strait of Hormuz, the US will destroy one of the country's bridges or power plants. President Trump was speaking at a rally last night and made additional comments on some of these things, saying that he doesn't think the Iranians are serious about talking right now, but thinks that they will be soon based on the continued attacks.
D
Donald Trump4:54
And I call it a skirmish. This is with this skirmish we have with the Islamic Republic of Iran. And I call it that because let me tell you, they're getting hit so hard and they want to make a deal. But I say they're not ready to make a deal because every time they make a deal, they want to change it and everything. They're not ready. They'll be ready very soon.
A
Andrew Ross Sorkin5:18
But this is why you see oil prices higher this morning.
B
Becky Quick5:20
And there's the question: does this bring them to the table or not? And that's the best negotiating sort of chess piece that's even available at this point.
A
Andrew Ross Sorkin5:32
On this vote, the yays are 232, the nays are 198. The bill is passed without objection. A motion to reconsider is laid on the table. The House approving legislation that would prevent members of Congress from buying individual stocks while they are in office. More than a dozen Democrats joined Republicans in support of that bill. But the so-called Stop Insider Trading Act falls short of what many advocates of a congressional trading ban have called for. House Republicans also attached a voter identification provision to that measure. Democratic leaders urged their caucus to vote against the bill. And by the way, even though the House leadership said that they should vote against it for the Democrats, more than a dozen Democrats did cross over and vote in favor of this. There is the question of 'don't let the perfect be the enemy of the good.' But then again, there are some real questions that have popped up around this in terms of how it's set up.
B
Becky Quick6:23
Yes.
A
Andrew Ross Sorkin6:25
Meta just out this morning with a new video and it's an interesting situation pushing CEO Mark Zuckerberg's take on the future of AI. The company says that he believes AI has the ability to build community and connection and is really sort of pushing forward into something. Let's show you the video and we'll talk about in a second.
M
Mark Zuckerberg6:45
Leave us behind. We couldn't disagree more. Call us optimists. Call us dreamers. Just as we've always done, we're betting on people.
A
Andrew Ross Sorkin6:58
We bet on people when we connected you to the ones you lost touch with. And again when we connected you. This seems to be a new effort by Meta, but I think more broadly by the AI industry. And I don't know if you've been picking this up between some of the comments Sam Altman's made recently, between some of the comments that Dario Amodei has made recently, and now Mark Zuckerberg about trying to reframe AI more about people. I mean this new commercial sort of about humans and the connection with people in large part because I think AI is becoming this sort of political football because people are so scared frankly that it's not about people or that it's going to undo people. Jeff Bezos who was on our air making this argument that he believes that AI is ultimately going to create more jobs rather than less jobs. Obviously there's a whole other side of it. So these are the optimists about what this is going to do. I don't know whether you think this is going to ultimately benefit people or not, but it's so interesting to me that the companies themselves now are sort of, you know, taking that narrative and at least try to push in this other direction whether it'll turn out to be that way.
B
Becky Quick8:11
Because you think they're afraid of the political football and how much pressure is coming down.
A
Andrew Ross Sorkin8:17
Look, this is my own take: right now there is, every morning we have skeptical conversations about what's going to happen to people's jobs. Is AI going to take your job? Is AI going to take your dignity? Is AI going to help you? And I think you're now seeing, whether it was Jeff Bezos on our air or now Mark Zuckerberg this morning, at least and I believe their views are genuine. I mean I think they come by them genuinely. They really believe that this is going to change humanity for the better. I know that there are skeptics on the other side, but it's just interesting to me that there's sort of an effort now to reframe the conversation around what AI can actually do for people. And so much of it hasn't really been that. So much has been about the technology and tokens and token maxing and business and ROI and all of that. And so the whole sort of effort now I think in this ad specifically to sort of shift the conversation in another way, you know whether it's successful or not I don't know.
B
Becky Quick9:20
Yeah, I was going to say that's the question. What do you mean you've brought up several times about how when AI has been mentioned at commencement speeches it was no, and I think that by the way I think this ad and what I think so much of the conversation in Silicon Valley has actually been shaped in the last 3 months, two months by those commencement speeches and by the booing and by the sort of realization that so much of the work that they're doing scares people. And the question of course is: is a reframe of this the answer unto itself, or is it just a reframe? And is it real? And the truth is we're probably not going to know until we get to whatever that place is.
A
Andrew Ross Sorkin10:00
Sure. And I will admit that I'm of two minds when it comes to AI too. I have incredible hope about what it can do on the healthcare front, what it might be able to do to find new cures, to find new ways to deal with things that we would never be able to figure out otherwise. But you do hear this constant drumbeat of what it means for workers. And we can say we think it's going to create more, but this is happening the same day we just find out that Uber is laying off 10% of its customer service people. And it is blaming AI for this. It's saying we're trying to get better about how we're doing this, but it's also because we're becoming more efficient with the artificial intelligence that we're using and how we're incorporating it. So you're going to see these competing headlines that come out on a daily basis.
B
Becky Quick10:38
So interesting. So Brett Taylor was sitting actually in your seat I think earlier this week. Chairman of OpenAI but also runs Sierra which is his own firm, which by the way is doing a lot of the kind of work that effectively customer service people used to do. The basic premise of his company is that we can do that work more efficiently and at a much lower price. And you hear that, you say to yourself: that's not a terrible thing. Progress has been made by increasing efficiency and making it so that humans don't have to do some of the jobs we used to have to do. It's been a constant since we've been working our way through this industrial revolution.
A
Andrew Ross Sorkin11:16
I think this was scarier because it's happening so fast. And the question is how do you retrain the people who are going to lose those jobs? How do you do it on such mass scale?
B
Becky Quick11:25
And so I think well, so then the question is, and since we won't know the other side of it until we get to the other side of it and who knows what the other side of it is, you know, the timeline. But with someone like Mark Zuckerberg or this ad coming out or other things trying to be more suggestive of the things that this could be helpful with, and I think healthcare is a huge part of where I have the most hope for what AI can do, the most promise for things that would take us millennia to try and accomplish on our own. If we can fast forward that, that's fantastic. But it comes right after the headline we got yesterday on OpenAI with the bots, you know, the agents getting out of control and breaking out of the sandbox and breaking into another company. I mean, those are the type of things.
A
Andrew Ross Sorkin12:09
And that's why this sort of lands in the middle of this whole debate. And I think that's pretty interesting. And what I just don't know ultimately is, you know, what will resonate with the public and how much of it's ultimately going to be a 'show me' story as opposed to a 'tell me' story. I do think that the AI industry hasn't done the best job of articulating exactly how this could work. I mean, it's all seems very vague even to someone like myself. So maybe this is the beginning of something that shifts that conversation. I don't know. I do think trust is going to be important. Well, and if you're someone who's been in this industry for a long time, great. Show us, but show us, like you said, show us. Don't tell us. The jury's still out on this. Show us how you're going to lead us to these areas. Show us and prove it. And I think it is a big lift when you have so much public animosity already and fear that exists at this point. I will say, I don't know if every single executive that I talk to these days says 2028, the election of 2028 is going to be the AI election. It's going to turn into a conversation, turn into a sort of litmus test on what you think of AI with different political candidates taking different positions and maybe even being very, you know, you think Republicans would be on one end a certain way and maybe Democrats on another end.
B
Becky Quick13:39
Meaning that they think like regulating?
A
Andrew Ross Sorkin13:43
I mean, to the Republicans to this point have had a lighter touch with it and maybe kind of trying to bolster more advancement. There is this entire 'okay we're worried about what it does to our society but we also don't want to lose to the Chinese who are working so hard.' I think there's almost two factions inside the Republican party. There's one that is much more in the 'let it' and there's a whole other crew that wants to go in a different direction. And I think in the Democratic side, you could see a whole other push that could push for way more regulation than before. And so I think that's a little bit of the other piece of it that the whole tech industry is sort of watching and trying to figure out how they're going to message or talk about and discuss that.
B
Becky Quick14:31
Cheese will be next.
K
Katie Kramer14:32
Next on Squawk Pod, prediction markets are coming to the 2026 midterms. Some states are already warning voters about placing bets on who will win. We will hear from a former CFTC commissioner who now works with betting site Koshi about the bets, the votes, and the challenges.
A
Andrew Ross Sorkin14:50
Yes, things can always go wrong, but it's very hard to police in real time. And we can debate the ethics of it all. Like, is this all legal? We've seen cases of insider trading in prediction markets and in other markets where the news comes out much later than the individual was flagged, the account was frozen, they've been referred for investigation and prosecution.
Welcome back to Squawk Pod from CNBC. Prediction markets betting on the likelihood of any number of events or outcomes came exploding onto the political scene in the last presidential election. Well, another November approaches. Prediction platform Koshi has rolled out a midterms hub to help political junkies and the site's users see speculated results and polling averages for the 2026 races all in one place. The launch came just a day after the elections commission in Wisconsin warned voters there that placing bets on elections they plan to vote in is illegal in that state. How will this turn out? I can't predict. Andrew Ross Sorkin takes things from here.
Joining us right now to discuss this and so much more is Brian Quintes. He is the former CFTC commissioner, Koshi Exchange board member and senior adviser to the coalition for prediction markets. So, of course, you are in favor of what's going on in terms of all this. Let's talk about Wisconsin as a sort of microcosm of a lot of the questions going on in the country about how prediction markets should be used, whether they can be manipulated, and how we should all be thinking about this. Big questions there.
B
Brian Quintes16:31
There's a lot of questions there. But when you saw the news on Wisconsin, for example, what did you think? Well, these are new markets, so I don't think it's a surprise that states are grappling with a lot of these issues from what the laws are that have been on their books for a long time. But in my mind, federal preemption is supreme when it comes to federally regulated derivatives markets. And there are a couple examples there that I think are informative. The first is, Michigan has a law on its books from 1931 that prevents its citizens from trading in grain futures back when it thought doing so represented gambling. Now, if it tried to enforce that on its citizens, I think we would all think this is ridiculous. But just because these markets are new, I think it's just part of our process to debate what the role is of a federal law versus a state law. But in my mind, this is an overstep by Wisconsin.
A
Andrew Ross Sorkin17:27
Okay. Here's a question. I'm going to posit something to you for just a second. There's an argument to be made that when people look at the prediction markets unto themselves, wherever something stands as an election is taking place, that it can influence their decision to vote at all because maybe they think it's a runaway situation and the person that they were going to vote for is otherwise going to win anyway. So they don't even show up at the voting booths or they're not going to vote or they're going to vote for the other person, meaning that the actual sort of where things stand, the horse race itself can influence what people's actions. And I would actually say journalists interestingly, the day of, uh, the day of an election often time hold back in terms of what they report and when they report it in part because they want people to go to the polls to do it honestly and not to influence the outcome. I'm curious what you think about that distinction.
B
Brian Quintes18:29
I don't think it's a lot different than what the arguments are around that with polling, to be honest. Sure, prediction markets provide more real-time information to people, but I think you also have to understand what the benefits of that are so that people can incorporate what a market's view is. And it's different. Polling and markets are different, right? Polling is asking someone what they want to have happen, whereas a market is asking what do you think will happen.
A
Andrew Ross Sorkin18:57
What's interesting though is the last election or last several elections sometimes the prediction markets have been more accurate than the polls.
B
Brian Quintes19:03
Well, that's right. I mean it forces people to take more of a dispassionate view of what they think the outcomes are. Let me take this back. To me, at least, the market mechanism, the data that comes out of this, what it means, is fascinating and, you could argue, beneficial in many ways.
A
Andrew Ross Sorkin19:23
The question I think is a slightly different one, which is ultimately does it and can it impact the outcome of an election because it has an effect on what voters ultimately do, and whether we think philosophically that is a good or bad thing. And whether polls are different, by the way, people don't typically run polls midday, people hold those polls back.
B
Brian Quintes19:49
You don't hold those polls back. I understand why that's beneficial to the marketplace. I'm wondering whether you think it's beneficial to the electoral process.
A
Andrew Ross Sorkin19:57
I mean, I just come from a philosophical perspective, but I think we have to trust people exercising their own self-interest. And if they are going to vote based on that, then I think a poll or a market isn't going to dissuade them from doing it. And if they don't care enough to do it and that's something that would influence it, then I think that's their decision to make. But I just think this is a critical new piece of data and information that is market-derived that can inform and in some cases lower polarization as opposed to having negative impact on the self-interest piece that it's, you know, people should do what is in their own self-interest and go to the polls. One of the other things I think people wonder about and this gets to the manipulation question: if somebody is wealthy enough and self-interested enough to decide they want to put a big bet, especially maybe not on a presidential poll since that is probably gonna have hopefully a lot of volume, but maybe on a smaller campaign, but wants to demonstrate that one person is way ahead and thinks that that can affect people's self-interest on the other side. How should we think about that?
B
Brian Quintes21:10
So, I think that's why it's important to recognize that these are markets that are federally regulated, that follow all of the same rules as traditional derivatives markets. In fact, they are derivatives markets. They're licensed derivatives markets. And there are rules on the books and they've been on the books for a long time that require exchanges and require the CFTC to surveil the markets for manipulative conduct. Those cases can be brought. Right. Outsized influence is something that occurs in all markets and not just these, and those can be flagged by size, by timing, etc. So I don't think we can have the perspective of saying you need to guarantee that nothing bad could ever happen in these markets as opposed to saying try to do your best and then make sure that you...
A
Andrew Ross Sorkin21:53
But there's also a timing issue. So I don't know if you saw this news. Another fascinating sort of development also relates to Wisconsin. I don't know if you saw that Elon Musk has now been said that he likely broke the law when he promised a million-dollar payouts to voters who were going to support the president in the last election. That's only happening now. Meaning part of the question is, yes, things can always go wrong, but it's very hard to police in real time. And we could debate the Elon Musk of it all.
B
Brian Quintes22:26
Well, I think the distinction there is when the news is reported versus when the investigation process occurs, right? And we've seen cases of insider trading in prediction markets and in other markets where the news comes out much later than the individual was flagged, the account was frozen, they've been referred for investigation and prosecution. So just because we're hearing about that now doesn't mean that that activity was flagged and is going through a normal judicial investigatory process.
A
Andrew Ross Sorkin22:55
How did you guys catch the teleprompter, the president's teleprompter operator?
B
Brian Quintes23:00
Yeah. Well, so I think there's a fine line between trying to disclose how a market surveils for activity because you don't want to tip people off of how to avoid that. Versus being transparent. But this was a person who was taking positions in real time when he saw the president skip over certain words that were in a speech, patterns of behavior over time. And this is a KYCed account. It means that the exchange customer, the exchange takes the information of who the person is, where they live, employment information, and all of a sudden you can start matching their identity with these traits. I mean, these are new markets and they might tempt people to do bad things, but the important part is that they can be investigated, they can be flagged, and they can be prosecuted.
A
Andrew Ross Sorkin23:49
How much time and energy is spent on trying to make sure that people aren't finding information and getting an edge on the market?
B
Brian Quintes23:56
So I think there's a distinction between incentivizing people discovering information and making a market more efficient than using information to benefit themselves over which they have a duty to someone else. I mean that is insider trading.
A
Andrew Ross Sorkin24:17
Do you have, does the industry have a list for example of what might be described as politically exposed people? Some banks keep lists of what are described as politically exposed people. You would describe by the way potentially the teleprompter operator as politically exposed potentially based on where they work or what they do. But how does that database work in practice?
B
Brian Quintes24:40
Yeah. So Koshi for instance bans certain people, certain government officials, congressional staff, cabinet officials, judges, mayors, governors, anyone in an FEC filing for instance, from participating in the markets at all.
A
Andrew Ross Sorkin24:58
At all means so sports any doesn't matter what?
B
Brian Quintes25:01
Well, you know, then it can get down to specific events and specific contracts. Coaches, referees, people in campaigns on their own campaigns or areas. So there is an actual ability to try to preclude certain people from trading on insider information upfront, which is more than any exchange in the world does. No other exchange precludes someone from trading in any market based off of that kind of sensitivity. So this is a pretty extraordinary step.
A
Andrew Ross Sorkin25:36
I think it's necessary. You guys are making bets that nobody else makes. There's a reason that you probably have a higher standard of things that you have to do.
B
Brian Quintes25:44
Well, I think it's important for market integrity.
A
Andrew Ross Sorkin25:46
Do you think it's across the board? Meaning you're talking about Koshi. There's a big story in the New York Times about the fight taking place between Koshi and Polymarket. It was a story that really suggested it's not just a fight but almost a personal fight between the two leaders of those companies and two different approaches even to the regulatory complex.
B
Brian Quintes26:07
Well, I think the reason that I wanted to join this coalition, the reason that I've been involved with Koshi for the last 5 years, is just because they took a regulatory-first approach. The companies in this coalition have regulatory compliance in their DNA. That's not to say that all companies in this space do, as opposed to taking a position of regulatory evasion and not following the laws on purpose and what bad outcomes that could lead to.
A
Andrew Ross Sorkin26:31
And do you think that's happening?
B
Brian Quintes26:32
I do. I think it's obvious.
A
Andrew Ross Sorkin26:35
And which firms do you think are doing that?
B
Brian Quintes26:36
Well, I think the story that you described as well as a story in the Wall Street Journal talked about offshore firms marketing to US customers when they're not registered, servicing US customers when they're not registered, not having any conflicts of interest rules. The CFTC has a very robust set of standards and guidelines and regulatory principles for regulated exchanges from customer protection, know your customer, reporting obligations, conflicts of interest obligations, the quality of contracts that can be listed. And I think it's at best an open question whether firms that are based in the United States but claim to be offshore are following those.
A
Andrew Ross Sorkin27:12
One final question has to do with sports because sports is a big piece of the prediction markets. Gary Gensler, former head of the CFTC back in the day, made a very interesting comment that when the original rules were put in place to allow for prediction markets, there was no way that any senators who voted in favor of this contemplated that sports would be part of it given the states that had built their own rules around whether gambling would be allowed or not. I think at the core of this issue is: did Congress in 2010 say 'no, none of the states can regulate this? It's going to this little small agency that I once was proud to run.' And the answer is categorically no. If we had asked Senator Harry Reid, who was then the majority leader who had the Nevada Senator Harry Reid, 'oh, we're going to take this and move this to the federal government,' he would have thrown me out of his office. I'm curious as someone who's sat in the seat you've sat in, what you think of that.
B
Brian Quintes28:21
Well, I don't know why we're trusting Gary Gensler with interpreting the law given what we saw him do at the CFTC and exceeding his authority and then what we saw him do at the SEC in exceeding his authority and for him to come now, having advocated for the law to be written as broadly as possible to give himself as much jurisdictional leeway as he could, and say, 'I'm going to mind-read everyone in Congress at that point in time and tell you what they meant.' You know, there's a great line from an opinion, a majority Supreme Court opinion written by Justice Barrett that said, 'The judicial role is to read text, not minds, and the text is clear.'
A
Andrew Ross Sorkin28:59
Brian, I want to thank you for joining us this morning. It's a fascinating discussion and debate and a real shift in society, and we're all trying to make sense of it. So thank you. We really do appreciate it very much. And we should make clear because this is an important reminder, CNBC and Koshi have a commercial relationship that does include CNBC having a minority stake in the company and a joint venture of sorts.
K
Katie Kramer29:22
Up next on Squawk Pod, the future of work for IT engineers. Stick with me here. Computer problems, benefits questions. Some companies like ticket site SeatGeek are turning to artificial intelligence. CEO Jack Groetzinger is highlighting the tool he is using for efficiency.
J
Jack Groetzinger29:41
You implement something like Serville and we actually now get more IT requests than we got before because they're getting answered faster, and you know it's actually going to be not that big of a pain to deal with.
K
Katie Kramer29:49
And the founder of that tool, CEO of Serville, Jake Stout. He envisions disrupting competitors like ServiceNow and freeing up engineers to do more meaningful work beyond your company's help desk.
J
Jake Stout30:01
The problem with a lot of these legacy cloud providers is the time to value, the time to implementation. So you have this brilliant idea for an automation for supporting employees. By the time you get that rolled out with these legacy providers, it could be months, sometimes years. We're going to have you up and running in days, if not weeks.
B
Becky Quick30:30
You're watching Squawkbox right here on CNBC. I'm Andrew Ross Sorkin along with Becky Quick. Our next guests have teamed up to automate IT support with AI, bringing engineers out of IT departments and into specific business units. Joining us right now is Jack Groetzinger. He is the CEO of SeatGeek. Also Jake Stout who is the CEO and founder of AI startup Serville. Servil's customers for automated IT support include SeatGeek, Perplexity, Together AI, and many more.
A
Andrew Ross Sorkin30:59
Gentlemen, thank you both for being with us today. Jake, I think first we need to dig into what Serville is. You were only launched in April of 2024. Already you've got a valuation of over a billion dollars. You have some big backers and you have even bigger customers, over a hundred customers, including Perplexity and SeatGeek and lots of others along the way. Explain what you do. You are AI native and you help automate IT departments.
J
Jake Stout31:28
That's right. We're an AI platform to get people help at work. It's employee support. So imagine you're an employee, you're having issues with your laptop or you have a question about your benefits. Instead of filing a ticket and waiting for somebody to respond to that, we're actually going to automate that resolution with AI. Get you back to the more meaningful parts of your job.
A
Andrew Ross Sorkin31:45
Which means what? I type into my computer, my computer's not working. I can't...
J
Jake Stout31:50
My computer's broken. And we fix it magically with the power. Like how does it work?
A
Andrew Ross Sorkin31:55
Absolutely. So it could be a chat interface where you're communicating over Slack or Teams or it could be over email. It could even be a phone call. So you call into the support line internally and you say, 'Hey, I've got a question or I'm having issues with my computer. I need access to an application' and we're going to give you that automatically.
Jake, let's talk about how you all have implemented it at SeatGeek and what it's actually meant for your production.
J
Jack Groetzinger32:16
Yeah, we had prior to Serville, IT requests every year, which is often every 7,000 requests every year. So that would be, you know, I need... I feel like I have that many every week around here. Yeah, it feels like that. But things like 'I need permission to this app' or 'I need a new computer' and thanks to Serville, we've cut it down by over half, which ultimately lets the people that were previously doing that kind of manual work focus on more interesting stuff and solve harder problems.
A
Andrew Ross Sorkin32:42
Let's talk about that because part of the fear with AI is that people think it's going to replace jobs. Did you lay off anybody? Did you hire new people? Explain how that works.
J
Jack Groetzinger32:52
No, didn't lay off anyone. In fact, I was looking at the stats yesterday. Our hiring thus far this year is already meaningfully above what it was for all of last year in IT. Just overall across the company and IT, I don't think we've made any changes, so it's flat. But company-wide, we've seen more and more hiring. There's this interesting effect where you implement something like Serville and we actually now get more IT requests than we got before because they're getting answered faster, and you know, it's actually going to be not that big of a pain to deal with. So overall, we see this as a way to move people from less interested work to more interesting work, get them focused on harder problems.
A
Andrew Ross Sorkin33:27
So what do the people in your IT department do now?
J
Jack Groetzinger33:29
Give you a random example. One person spent a number of months kind of embedding with our customer support team and looking at what is their life like, how do they spend a day? How could you automate things? How can you turn eight clicks into two and create custom stuff for SeatGeek, sometimes with the help of AI, that ultimately allows them to get back to customers more quickly.
A
Andrew Ross Sorkin33:50
So you're a big fan.
J
Jack Groetzinger33:51
Big fan.
A
Andrew Ross Sorkin33:51
Yeah. Let's talk a little bit more, Jake, about what exactly happens with this. And you know, you're going after companies like ServiceNow. That's kind of your...
J
Jake Stout34:01
Most people know us as the ServiceNow competitor.
A
Andrew Ross Sorkin34:05
Yeah. And what do you do differently than what ServiceNow does? Bill McDermott was on this week talking about his earnings too and saying that they have AI agents as well. What do you do differently?
J
Jake Stout34:16
I think we get people implemented so much faster. I think the problem with a lot of these legacy cloud providers is the time to value, the time to implementation. So you have this brilliant idea for an automation for supporting employees. By the time you get that rolled out with these legacy providers, it could be months, sometimes years. We're going to have you up and running in days, if not weeks. And you're going to be getting all these requests automated. You're going to get people back to the work they actually enjoy. Get them out of those ticket queues.
A
Andrew Ross Sorkin34:41
Why are you able to do that faster?
J
Jake Stout34:43
So fundamentally the technology is built around these codegen agents. So we use AI to write code that powers all these automations. So we take this amazing capability of these AI models to be able to generate code. You're seeing this being activated in engineering organizations today. We're taking that same approach to IT and other back-office teams. So they can use codegen to automate not just help desk requests but onboarding, offboarding, compliance workflows, all these different things that take up so much time for all these different teams.
A
Andrew Ross Sorkin35:11
Jake, you know what the next question has to be after what we just heard from the OpenAI agents that got out of the sandbox and broke into Hugging Face. What happens? How do you control the agents? What do you do to make sure that they're not taking or doing taking information they shouldn't be taking, doing things they shouldn't be doing?
J
Jake Stout35:29
It's a really important problem and something we're thinking about a lot. One of the core pieces of our architecture is this two-agent model where we actually have agents that IT and admins use to build workflows that are fully deterministic, that don't have the ability to go into these different systems but are very constrained, and then a separate agent that end users talk to that can only run those automations in a deterministic way. So you can avoid a lot of these scary problems when AI agents have direct access to these underlying business systems.
A
Andrew Ross Sorkin35:55
Is there a kill switch for agents that go rogue at any point?
J
Jake Stout35:57
There's more than a kill switch. There's actually almost a whitelist of things that it can do. So it cannot do literally anything other than what you've expressly designated, which is much more...
A
Andrew Ross Sorkin36:07
Difference between those OpenAI agents which were essentially told that they could, the guardrails were taken off and they were kind of seeing what would happen.
J
Jake Stout36:16
That's right.
A
Andrew Ross Sorkin36:16
Okay. So let's talk from your perspective, Jack. Do you have concerns about any of these things or you feel like this is all good?
J
Jack Groetzinger36:24
Listen, I think anyone should be constrained about giving AI unconstrained access to lots of different things. We've done the opposite. We've tightly sandboxed what an agent can do within SeatGeek to human-approved stuff that we've reviewed, we've published, that's auditable afterwards. And it's part of the, you know, broadly we're becoming an AI company very quickly, both in terms of our product, but almost more interestingly in terms of how we operate. And obviously you have to do that responsibly.
A
Andrew Ross Sorkin36:51
But when you say you've sandboxed it, OpenAI thought it sandboxed it too. So do you actually think I mean that this is the fundamental question. How confident can you be in sort of the quote unquote sandbox?
J
Jack Groetzinger37:06
Yeah, I think listen, there were hacks and mistakes made before AI. I'm sure there will be more in the future. It ultimately comes down to taking great care and building really strong systems that are reviewable by humans that have tight permissioning. And that is where it all starts for us. And I think if you start from a place of 'you can only do XYZ', not unconstrained as in the OpenAI Hugging Face case, then you're in a pretty secure spot.
I think increasingly the product becomes the governance, the ability to control what the AI can and can't do, what it has permissions to perform, who has to approve those actions. And I think the products that will succeed are those that really emphasize the governance layer, not just the underlying capability of the models.
A
Andrew Ross Sorkin37:45
Can I just ask you a broad sort of philosophical question? We were talking about this ad that Mark Zuckerberg put out this morning about humans, about the need that AI is going to make humanity better, and sort of reframing the narrative around all this. On one end, the industry seems to be trying to move in that direction, at least talk about it like that. At this at the other end, as a business, you have to talk about the ROI on what your product can ultimately create for your clients. And those things to me, at least in the short term, feel like they're in conflict.
J
Jake Stout38:22
I don't think that's true. And I think it's because the AI is automating the work, not automating the job. There is actually infinite work that we want our employees to do. There's so many valuable things that workers can do. And so when you automate a lot of these low-hanging tasks, the IT support requests, the password resets, you can take folks like IT, which are the most technical people in your organization, and you can deploy them like SeatGeek has into all these really productive activities, more productive than what they were doing before. So I actually think we're seeing unlimited demand for more and more work from these employees. So we want more of them, not less of them, even as AI is taking over.
A
Andrew Ross Sorkin38:59
Jack, I got to ask you while you're here. The World Cup was the biggest event, the finals that you've ever seen on SeatGeek. You had massive success with the NBA playoffs, with the Knicks tickets. All of this though is happening against the backdrop where Congress is trying to heavily regulate the ticketing industry. How do you kind of balance the best of times, the worst of times?
J
Jack Groetzinger39:20
Yeah, World Cup is amazing. I'm personally still in a hangover for not being able to watch every week. I think everyone agrees on the problem, which is fans want lower prices, they want more access to tickets. Where the evidence disagrees is that capping fees or capping the total resale price actually leads to better solutions. We've seen as a random example they tried this in Ireland and saw a meaningful increase in scams. They're now trying it in Ontario and are having a ton of problems with enforcement. So I think the real answer is transparent pricing. There was actually a big change last year that was really positive which required every marketplace to show prices all-in, meaning Jake's buying Nick's ticket, what he sees is what he pays. Another issue which has not been tackled and I think is really really important is speculative ticketing, which is basically someone selling a ticket that does not exist. We do not allow that on SeatGeek. We think it should be banned industry-wide and it leads to lots of problems which some marketplaces saw with the World Cup where they had meaningful numbers of fans that bought tickets, they might have flown internationally, they show up and they can't get in.
A
Andrew Ross Sorkin40:24
How do you prevent that?
J
Jack Groetzinger40:25
You verify that every buyer actually has the ticket ahead of time. The problem is that it's actually from a short-term economic perspective kind of appealing to sell spec tickets because you can make some money and you can offer inventory that not everyone else has, but it comes to roost because it doesn't exist. Exactly. And the seller is hoping that they can then acquire it and fulfill the order. Sometimes they can, sometimes they can't.
A
Andrew Ross Sorkin40:47
Yeah. Okay. Jack, Jake, thank you both for coming in. We appreciate your time.
J
Jack Groetzinger40:51
Thank you so much.
K
Katie Kramer40:54
And that is Squawk Pod for today. Thanks for listening. Squawkbox is hosted by Joe Kernan, Becky Quick, and Andrew Ross Sorkin. Tune in weekday mornings on CNBC at 6 Eastern. Get the smartest takes and analysis, the interviews you can't miss, all the good stuff from our TV show. Get it right into your ears when you follow Squawk Pod, wherever you get your podcasts. We'll meet you right back here tomorrow.
A
Andrew Ross Sorkin41:17
We are clear. Thanks, guys. Thank you so much.