The Brave New World of AI Cyber Crime with Evan Reiser from Abnormal AI

0:00Allie:Hey Chicago, class it up with Crocs. You know back to school is coming in fast. So why wait to find your new fave footwear? Step into a local Crocs store and step into your new look. Try it. Style it. Make it yours. Because the right pair doesn't just show up. It shows off. First day fits, handled. Walk out ready for whatever's next. Visit your nearest Crocs store today. This episode is brought to you by Accenture. When your advertising operations fall out of sync, everything else follows. Spotify and Accenture are working together to reinvent the rhythm of ad sales, using automation, analytics, and smarter workflows to simplify campaign delivery and access better data across the business.

0:49Allie:The result? Less time spent on operations, more time connecting brands with the moments and fandoms that matter most. Learn more at Accenture.com slash Spotify. The thing I hear over and over again from folks in cybersecurity is that the threat-led landscape is just expanding rapidly. I think it's undeniably true. I think just the surface areas where how you can get hacked are increasing. Hello, hello. Welcome back to Term Sheet. I'm Allie Garfinkel, senior writer here at Fortune. And every week on the Term Sheet podcast, we're going to go over the latest news and insights in private equity, startups, and venture capital.

1:23Allie:Every week, here's where you can watch and listen to us speak to some of the most exciting figures in the private markets.

1:33Allie:This week, I've been thinking a lot about how the line between the public and the private markets is thinner than ever. Now, this is a natural thing to be thinking about in the aftermath of the news around Intel. In case you missed it, what essentially happened was this. The U.S. government has now become Intel's largest individual shareholder, spending$8.9 billion to buy a 10 % stake in the company. I was initially taken aback by the news, but then I remembered that while this is the first time this is happening outside an active fiscal collapse on a societal level, there is actually quite a bit of precedent, especially around the 2008 financial crisis, when the government actually took on stakes in companies like AIG, Bank of America, Citigroup, and GM.

2:14Allie:GM, it was actually a 60 % stake. So there is some precedent here. That being said, this also does for me really create concerns about Intel. Now, everyone's been concerned about Intel for a while, but I've actually found myself wondering about how this will affect startups. One of the most important things to remember about sort of the relationship between Intel and the startup ecosystem is for a long time, Intel was one of the most active investors, one of the most active CVCs in the system. But I am now wondering if there's no turning back and if the downstream effects on startups are just beginning.

2:46Allie:Now, what's happening with Intel is worrying, but it's important to remember it's also happening in the context of the broader AI boom. And I think it's worth doing an AI vibe check here because we're getting some mixed signals across the system right now. On one hand, there's a sense that Silicon Valley is hunkering in for the long haul. The Wall Street Journal reported that OpenAI's Greg Brockman and Andreessen Horowitz are part of helping set up a super PAC called Leading the Future that has raised$100 million to support AI regulation in the midterm elections. At the same time, we're getting actually a lot of pressure in the public markets.

3:18Allie:You have single day drops for companies like AMD and NVIDIA. You have the NASDAQ under pressure and a sense of uncertainty moving forward. And at the crux of all this is the idea that we're at the center of an AI bubble. Now, I'm just going to raise my hand and tell you what I really think. we are absolutely in a bubble. There is absolutely no question that we are in a bubble. There's a lot of classic bubble behavior here that we've seen since the dot-com boom, at least if not before, where valuations are high, follow-on rounds are fast, and eventually all of this is going to have to come home to roost in some way, shape, or form.

3:49Allie:Now, I do believe the technology is resilient and will change our lives in ways that we don't understand, but over time it's sort of like that cliche from HBO's Silicon Valley that all of these sea turtles are not going to make it to the ocean. And I think that in terms of the bubble, we just don't know the shape of it and who the winners and losers are ultimately going to be. Now, this week's guest has become one of my favorite people to talk to about AI, not only for his science fiction bullishness, but his candor about the things that might be difficult. Evan Reiser is the CEO and founder of Abnormal AI.

4:21Allie:It's one of the fastest growing cybersecurity companies of all time and has been AI focused since the beginning. And Evan and I first met at the very beginning of this year. We were introduced by an investor, and I'll be super honest with you, though we talked about abnormal, we ended up actually talking about how I was using ChatGPT to analyze Macbeth, and he gave me some of the best prompting suggestions that anyone has given me. So Evan is not only a founder building in the space, but he is somebody who is an avid user of all of these tools. Here's Evan. Evan Reiser, welcome to the Termsheet podcast.

4:54Allie:I'm really excited you're here. Alec, great to see you and super excited to be here as well. I'm really excited to talk to you about the future, but I want to start with the past and start with you. What was your first term sheet? So my first term sheet was for a startup, right? I'm now the CEO and founder of Admiral Security, but that's my fourth company. My first term sheet was probably a little over 10 years ago for a different startup. Is there anything you remember about it? Yeah, it's a little embarrassing, you guys. So I, you know, I've always been a product builder and entrepreneur, but I wasn't really exposed to, you know, kind of Silicon Valley.

5:32So the whole idea of like term sheets and venture capital is a little bit foreign to me until like, you know, my second or third company. I'm now much more educated. But yeah, it kind of threw me off.

5:43Allie:Tell me about day one of Abnormal. You know, it's really hard to tell. We had a very abnormal beginning as a startup. As you know, there's kind of a cliche about how startups start, right? There's two people in a garage, you come up with an amazing idea and a vision for the future, and then you show up at a VC's office and give them the pitch. For us, it was much different. Admiral was my fourth startup. I didn't know exactly what the startup should do. I was really obsessed with this technology vision, the idea of using AI to better understand people and predict things about them. That mostly came from my background in advertising technology.

6:20um and so i went to a you know i went to a venture capital firm that i i knew really well greenlock partners and i said hey i want to go spend the next dozen years maybe you know maybe maybe a couple decades building this thing but i don't know what the product is i don't know what the market is right i just know what the technology is and i probably spent about probably 50 to 100 hours right with these two investors debating different markets and problems and you know we ended up building a kind of 10-year you know plan right together before i ever actually like officially pitched them, let alone got a term sheet.

6:52So it was much different than the kind of normal startup experience.

6:56Allie:When you think back on that experience, and you'd founded companies before, what did you learn in that process that you never would have known otherwise? Well, I mean, I certainly wouldn't have been in cybersecurity. You know, my background's not in cybersecurity. I'm a cyber outsider, despite being, you know, at a big cyber company today. You know, I've always been a very customer-oriented product manager. And so, you know, kind of in partnership with these two investors, we went to go talk to 100, maybe 150 different CIOs and CISOs. And we talked about, you know, our belief about the trends in AI and kind of what was possible in the future.

7:33And then we asked them, you know, what are problems that we could solve, right, as we kind of build this technology? And what will you pay us to go do in the short term? Because we don't really have a product or solution. And we talked to literally over 100 people, we got 100 different ideas, right? And when we asked phrase, you know, concretely great, but like, what would you pay us a quarter million dollars to do right a year from now, as we kind of go towards this technology vision, we got very specific answers on, um, you know, even since people talk about cybersecurity within the cybersecurity piece of feedback was all about fraud, social engineering, um, account takeover data loss prevention.

8:09And so I, I didn't really know much about that world. I didn't really know kind of the extended problems that people had in cybersecurity. And so if it wasn't for some of those like open-ended conversations, I wouldn't have ended up where I am today.

8:20Allie:There are many cyber outsiders out there who are now increasingly aware of cybersecurity, but I want you to imagine the most non-technical person that you know. How would you explain to them what Abnormal does now? Well, I think the simplest to describe it is, we help make sure the communications you get are legit, right? So if we make sure there's bad guys out there that are trying to do bad things to you, maybe get you to steal your password, steal your data, steal your money. We make sure that we're blocking kind of emails, other forms of communications that they might use to try to get in front of you.

8:56So you never have to see those things. It all happens behind the scenes.

8:58Allie:What do you think people most misunderstand about email security in particular? I think the average person, right, the outside industry might have a better intuition about the reality of email security than people in cybersecurity. And cybersecurity is filled with jargon. There's a lot of gatekeeping. There's almost like a lot of intentional complexity. If you ask probably, you know, our families, right, that maybe don't work in, you know, technology or cyber, what are some of the cyber attacks? I might say, oh, there's bad guys trying to, you know, fish me or trick me into stealing my money. Those are actually the number one cyber crimes.

9:32You go to the FBI data and you say, you know, ask the FBI, where do people actually lose money? number one crime in the world is social engineering via email of businesses. If you look at any breach report, number one cause of breaches is phishing. However, if you go to a cybersecurity conference, you'll hear about a thousand different problems, and they're all there. But I think people misunderstand the magnitude of some of these problems, and they're obviously getting worse today as criminals are adopting more of these AI tools themselves.

10:00Allie:Cybersecurity is, I mean, it's sort of to your point. I think a lot of people still don't understand, But I think people are more aware of it now than they maybe would have been 10 years ago. Is that true in your experience? Or is that just my corner of the universe, my little bubble? I think it's undeniably true, right? If you, I'm kind of making this up, so maybe someone will correct me. But my guess is if you look at like how many articles on Fortune were the last year on cybersecurity versus 10 years ago, it's definitely gone up, right? Cybersecurity is a kind of household conversation, right?

10:31It's a political conversation, much more in the media and the news and the culture. I think a big reason for that is there's more stuff online, right? We're spending more time online. There's more data online. All the stuff online is interconnected, um, for, for better or worse. Um, and it is becoming a bigger problem. If you just look at any chart of cyber crime by any reporter, right? By, you know, on any dimension, it's going up into the right. And, you know, just, and we're here in 2025 and as a civilization, we've never spent more time or energy and focus trying to, you know, combat global cyber crime gets never been worse.

11:03And so it feels like, you know, it's surely more important today and maybe even more important tomorrow.

11:10Allie:Do you think as a civilization we are set up to combat cybercrime right now? To what extent? I would say like yes and no, right? It's kind of like the best of times, the worst of times, right? Like are we – we're probably better prepared today than we were last year or the year before. But there's also more crime than there ever has been. Again, there is more money lost from cybercrime in 2024 than 2023. and so forth. You have more criminals. There's a lot of tools that allow non-technical criminals to become cyber criminals. You no longer have to speak English to scam someone. You no longer have to know how to do technical things.

11:48All the technologies that benefit our civilization are also helping criminals send out more attacks. And then with the rise of these AI power tools, where any large language model, it allows any cyber criminal bozo to have really deep understanding of their victim and scams and really sophisticated ways of tricking their kind of targets. So, you know, we're in a tricky spot today. I think long term, I'm pretty optimistic on our ability to win the war against cybercrime, but I would say the next couple of years looks pretty scary.

12:18Allie:Well, I was going to say sort of to that end, with the rise of AI, you know, the thing I hear over and over again from folks in cybersecurity is that the threat landscape is just expanding rapidly. It sounds like that's been totally true for you. You're nodding. I think it's undeniably true. I think just the surface areas where how you can get hacked are increasing, right? There's more stuff online. The IT architecture is more sophisticated, more complicated. More of our personal lives are more integrated into the digital world. So that makes the problem harder. But also the sophistication, the attacks are increasing, especially the rise of generative AI.

12:55Anyone on the internet can write a very compelling email, right, or to go trick some company to send them money. Anyone can use a free generative AI tool to create a deepfake video or deepfake voice. These are things that were, felt like science fiction five years ago, and they're clear and present problems today.

13:13Allie:You and I have talked about science fiction before in our time together. What is something that you thought would have been impossible five years ago that is just true right now? Oh my God, I feel like almost half the things we do day to day. I'll give a really simple example. you know one thing i do is i record most of my meetings right with a tool a country's transcripts at the end of the day i have a little deep research agent go back through all my transcripts and give me feedback about how i can be more effective right both positive negative things um and so basically have like a little ai ai executive coach um and that is uh it's super effective right it's incredible right so in some ways in some in some areas it's better than my real executive coach um so like that that's that's pretty amazing you can imagine a world in the very not too distant future where every kind of person, every employee has a personal coach and that coach doesn't care about your business or your boss's politics or your OKRs.

14:05That AI coach is trying to help you become a better professional, better person, help you get more joy out of life. And these things cost, you know, cents, right? Not a hundred thousands of dollars. If you told our former selves that everyone in 2026 could be walking around with a personal coach that was making them better, making their life more enjoyable, that that certainly would sound like science fiction.

14:24Allie:I mean, one of the things this makes me wonder about, too, is just our ability to predict change over time. And in a lot of ways, you're sort of in the business, right, of trying to figure out where the puck is going. Where do you think AI is headed on the expectation that maybe we're not that good at predicting change over time? Yeah. I mean, I think human brains think very linearly, right? And I think everyone's got different predictions about how much AI is going to accelerate, how it's going to change the world. I'm probably personally very far like the AI pilled side right like you know maybe it's because I've read a lot of science fiction growing up but one kind of parallel that I think about a lot is um the rise of like the internet I'm probably dating myself a little bit here but um if you go back to like the early days the internet right the mid 90s and you surveyed like a thousand technologists and you kind of rank them from their most bullish predictions about how you know um how much the internet would change the world if you went to the kind of top one of a thousand right the person that was talking about the crazy stuff, how the internet's going to change the world, I think they probably have underestimated, even the most crazy bullish person, they've probably underestimated how much the internet has changed our lives, even in 2025, right?

15:33I don't think that person could have predicted Uber and DoorDash and Google Documents or any of this stuff, right? Amazon. Yeah, we find those things incredibly commonplace and almost mundane and obvious today. So I have to imagine there's some similar thing with AI, right? If you take 100 researchers in AI and you say, how much do you think AI has got really changed the world over 30 years? I bet you most of them are underestimating the true impact. What that is, I think, is very hard to say. But even the last one year, I'm working in machine learning AI for 15 years. I've personally underestimated the impact across my personal life, across my professional life, across our business.

16:13So I don't know exactly what that world is going to look like, but I can guarantee it's going to come kind of faster than we think. And it's going to be more extreme than we can imagine today.

16:21Allie:Bringing it back to abnormal, what is something that has been true for abnormal over the last few years that you could never have predicted? I mean, I don't know about the last few years, but I'll say the last couple of few months, the way we're using generative AI to change how we work, right? If I look at abnormal, even the beginning this year, so seven months ago, I feel like abnormal operated more similarly to how probably a big software, like Oracle operated 30 years ago than how an AI native company should work, you know, in 2027 or 2028, you know, how you listen to customer feedback, how you route that to different engineers or product managers.

16:56In the past, there was, it would take days and weeks. And like the chance that a customer gives some feedback or actually make its way to the code was like close to zero. You know, today we can basically listen to every customer call, every ticket, every email, basically every voice of the customer. And within 24 hours, that's being analyzed, summarized by AI. is going to the engineer product person responsible for that part of the code. We can go to customers and say, we will listen better, more broadly, and apply your feedback more effectively and fast than any other company in the world. That's a crazy statement to make, yet it's gone from not true to very true in the last six months.

17:32And so I think about even the next six months going forward, it's hard to even imagine all the ways that these different AI tools will help us transform how we run our company.

17:40Allie:speaking of which you guys recently changed your name to abnormal ai from abnormal security why the change well kind of um so actually the original name of the company was abnormal ai so if you go back to when we started in 2018 yeah so it was a bit of a uh went back to back to our origins so the reason we changed it from abnormal ai to abnormal security was because in 2018 ai was kind of a dirty word you know people would tease us for they would make like um uh like terminator references and skynet references and be like all this ai sounds like you know some crazy fantasy. So we actually kind of stripped everything related to AI and machine learning off our website and out of our company, from the outside of our company in 2019.

18:19It wasn't effective for us. It was actually just tracking for us to try to sell to security buyers. We did change the name earlier this year back to Abnormal AI. And it wasn't really for external reasons. I don't think anyone outside our company cares what the name of the company is. We never know. But I wanted to do it inside the company to remind ourselves that AI is not just how we implement our product, right? AI is changing the enterprise landscape, right? The needs of customers are changing. It changes what problems we can solve. It changes how we solve our current problems. It changes how we work.

18:51And there's things that we felt were impossible and maybe insurmountable many years ago that are now totally possible. So I wanted to do it more as a reminder about our ambition and vision as a company for our employees more so than, you know, to kind of capitalize on the AI hype wave, which I don't think any customers get any brownie points for having AI in our name. But I think it still is important for us.

19:14Allie:Evan, you also have a reputation for saying no to a lot of investors. Can you tell me about that? Walk me through Abnormal's fundraising history because it's very curated. I think specifically, you know, I just don't spend a lot of time with investors. I think my core belief is that if you're a great company with great people building great technology, great market, like the investing will work itself out. And so, you know, we've raised, I think, over a half billion dollars since inception. I probably met with 15 investors total, right? In the last, you know, eight, you know, seven or eight years, I've never talked to a single investor outside of a fundraising process.