E4: How AI will impact the role of sales with Zack Kass of OpenAI & Jake Saper of Emergence

As Generative AI capabilities continue to improve, there are fundamental questions as to the role of AI in sales.  How should sales teams leverage AI? Can AI ever be indistinguishable from human sales agents? Or will humans fundamentally always still be involved in interfacing with customers? 

In this episode, Zack Kass, Former Head of GTM at OpenAI and Jake Saper, General Partner at Emergence Capital join Regal CEO, Alex Levin to discuss the ways in which AI will impact the role of sales reps. 

Here are the key takeaways from our conversation: 

1. AI is progressing much faster than most people realize. But regulation will likely slow down the development and adoption of AI, in much the way driverless cars have been rolled out slowly to prevent business goals from resulting in negative outcomes.
2. In the short to medium term, AI will be used as a means to augment existing sales professionals’ capabilities rather than to completely displace them. 
3. AI is a technology that will lift all boats. From the lowest performing sales reps to the rockstars, AI will be able to provide valuable coaching opportunities and unique sets due to the plethora of data it sits on top of. 

AI is a massive technology trend that has the potential to fundamentally change all knowledge-based industries. However, the way it makes its impact will largely vary depending on industry, use case, and the role that humans have in the work. 

In this episode you will learn: 

• How quickly will AI be able to do the selling instead of a human?
• What are the limiting factors to AI being adopted across the industry? 
• Is the proliferation of AI ultimately going to be good for customers?

To learn more about everything Outbound Contact Center, read more posts at regal.io/blog or email us at hello@regal.io.

Transcript

0:01

This is Alex Levin, the CEO of Regal io, and I'm here with Zach, formerly from OpenAI. And Jake, the GP of emergence or AGP at Emergence. Uh, and I'm gonna let them introduce themselves in a second, and then after their introductions we'll get into a discussion of AI and specifically how it's gonna impact sales, both in B2B and in the B2C world. So, So, Zach, thank you very much for joining us today from Sunny California. Uh, maybe you wanna give us a little bit about, you know, how did you end up at this behemoth, at this, you know, famous company now, open AI, and, uh, you know, uh, what are you up to now? Yeah, so, uh, I, I was lucky enough to start my career in AI at a company called CrowdFlower. Uh, working for now, now a very prominent ceo, Lucas ald, who at the time was just sort of like, Um, an entrepreneurial data scientist. This is 2009, and data labeling was sort of an emerging, uh, trend, albeit to a very small niche market, which was data science, that of which there just weren't many, and I didn't know it at the time, but I was really lucky to get, uh, to get involved so early in the category. I spent six, uh, hard years at CrowdFlower, to be honest. It became figure eight and then sold the app in. Uh, the market was too small. We were too early and scale AI sort of came in at the, at the tail end of our journey and, and took off, and that rest is history. But I stayed in ai, um, and went to a company called Lilt, which was founded by two researchers building large language models for the purposes of translation that they started that company in 2015. Actually the same year that, um, that Open AI was founded. Which was an important year because it was just as neural nets were, were, were gaining, were, were gaining popularity, um, again, and, um, the, the, the power of the technology was, was exploding. And then when it came time for open AI to hire their first sales leader, I happened to be one of the few people in the world that had. Um, that had done this. Um, and, uh, yeah. And now the future is, uh, I'll, I'll probably, uh, I'm gonna move to Santa Barbara. I've moved to Santa Barbara and I'm gonna spend my time consulting with, uh, companies on how to build, um, a future around ai. And eventually I'll run for Mayor Santa Barbara in 2026 in large part because I think that physical, uh, physical communities, uh, will be the, will be the most important thing we can invest in. In the future, livable, walkable cities, uh, will is how people stay connected to each other. Otherwise, they're just gonna plug into vr, ar headsets and, and tune out. And, um, and I think the world will bifurcate those that get on the Wally bed and those that that don't. Cool. It's a great introduction. So, I mean, if I can ask, like, when you think back to starting in ai, were you prescient? Did you know this was gonna be so big so fast, or was this, uh, you know, lucky, uh, that you applied to 20 companies and that was the one that said this? No, no, no, no, no, no. To be very clear, first of all, it was 2009 and I had, I was a volleyball player, so like I had no fungible skills and, uh, outside of volleyball and there are hundreds of dollars in men's volleyball. So, so I don't recommend it as a career to people. Um, and I, I applied to three companies Square, um, Airbnb and CrowdFlower. Uh, tho those were the three companies hiring at the time. And, and I got a job at CrowdFlower actually as an intern, uh, at first. And then Lucas just kept investing in me. I, I wasn't even pressuring until I didn't feel like I saw the future until 2017, um, when Opening Eyes started launching the large language models. You know, they, they moved from DOTA and robotics to like G P T and it was very clear that we were at the, be very, at the base of a J curve. And it was clear to the smart researchers at Lil, which is how I got exposed to it. And then it became so obvious to me and I couldn't understand why the market wasn't updating. Um, and then as soon as Open AI opened this job, I, I, you know, I was like, this is, I, I do this for free and give them an arm in the process. Um, but, but yeah, it, it, you know, it all happened so fast, and Andrew Ing is honest. He talked about it recently at a Sequoia event. No one predict, like, not, not even the people close to it, predicted it really, except Ilia. I mean, Ilia wrote this, the chief scientist at Open AI wrote this incredible paper in 2012 that sort of predicted all this. But very few people in the world cared about it or believed him. Yeah. So. Great. And so Jake, uh, uh, full disclosure, you know, you're an investor in Regal, uh, you know, you're, you're now a GP in emergence. Um, I mean, maybe tell us a little bit about your journey, uh, to being a VC and now into being an expert on ai. Um, it's really fun to do this. Uh, I care both, I care about you both a lot and it's cool that we're all together. Um, I, uh, I grew up in Austin, uh, before Austin was what it's become, um, sort of in the initial tech boom of Austin in the eighties and nineties. Uh, my parents, um, are tech. Entrepreneurs, they've started a bunch of comp companies together, um, as co-founders. Um, and I was sort of sitting shotgun for them, uh, growing up. Uh, I didn't realize that that was a job opportunity, uh, going forward. But, um, in many ways the role I play, uh, Alex with you and Rebecca at Regal, um, is not wholly dissimilar to the role I played with my parents and their companies growing up. Um, I have like more legal rights, uh, in Regal than I did with my parents' businesses, but, um, I've done a lot of coaching founders and problem solving and trying to help however I can since I was a kid. So in some ways I was raised to do this. Um, I also find it really meaningful and joyful. Um, I started my career in consulting. I'm a super structured thinker. I love frameworks. Consulting was a very good fit for me. As a result, I see the world in two by twos. Um, I then did a startup myself. Uh, I helped start a company that was developing solar power plants in India and Africa. Uh, very different type of startup. It was very hard, it was very enterprise salesy in the sense that we were selling, uh, very, very large seven figure contracts to governments. Um, that lasted 15 years. Um, after about four years of that, I came to California. For grad school, uh, and then transitioned into venture and have been doing that for, uh, almost a decade in terms of my path into AI and AI investing in particular. Um, I started to spend a bunch of time on this in 2016. Um, my partner, Gordon Rader and I, uh, started to develop a investment thesis that we call coaching networks. And the core idea behind that is, Um, a lot of the AI energy, um, and, and kind of automation energy back, uh, in the kind of uh, 2015 to 2020 time zone was around how can we use technology to help automate tasks away? And that was kind of the rise of rpa, uh, companies like Automation Anywhere and Blue Prism, et cetera. Um, and we sort of saw that as, uh, commoditizing over time. Cuz once the task is gone, it's gone. We were thinking more about how can we use these technology breakthroughs to help coach workers, uh, in real time on how to do their jobs better. Uh, and the core idea was that the highest value tasks are human to human tasks, and those are gonna be the ones that are hardest to automate away, uh, over time. Um, and so how can we use the same technology that's automating some tasks away to help augment, uh, folks doing it? So, um, our first investment with this thesis is a company called Chorus. In the sales space, it's sold to ZoomInfo a couple years back, but, uh, many of you're probably familiar with, with Chorus, uh, plays in a similar space as Gong that helps coach folks, uh, sales folks on how to, um, have better conversations. We then invest in a company called Textio, uh, which does augmented writing. So they coach specifically writing in the HR domain folks on how to write, uh, more effectively in that domain. Um, and since then we've invested in all sorts of, uh, companies using AI to help workers do their jobs better. Okay. And you know, if you think back to your decision to go to vc, to your point, like you had a lot of fun in consulting. You were in a place that sort of emphasized the values that you care about. You know, why did you decide to switch to VC and, and, you know, what advice would you give to people as they're thinking about, you know, is VC right for me? Um, I have a two by two to answer that question. Um, of course, I don't know if, like, is it a video? Should I draw it or should I just describe it? Cause I can, I can also draw it real quick. Um, so, so basically like there's two dimensions along which I thought about my career and which I think you can plot a lot of careers. Um, You've got on one dimension, like the X-axis is breadth and depth of content. So there's some jobs where you go super deep and you become a world's expert in a very specific thing. And there's some jobs that are very broad in terms of the content, uh, that, you know, you're doing all sorts of things. So in consulting, for example, I worked for Novartis. I worked for Shell, I worked for Campbell Soup, I worked for all sorts of companies trying to help them solve all sorts of problems. That was kind of the definition of breadth. Whereas when you're, you know, starting a company, when you're leading a company, when you're an executive in a company, you're really thinking about how to solve a very specific, you're a world's expert in specific things. So there was a, there was a time, there was a moment in time where I was probably the top 10 in the top 10 world's experts in building utility scale solar, uh, photovoltaic power PR plants in Western India. Hmm. That's a very specific thing, and I'm definitely not that anymore, but there was a few years where like I was, you know, top 10 in the world, expert in that very, very narrow specific topic. Um, so that's like, you know, kind of the, the polls of the breadth versus depth. Um, and then the other axis I think about when you plot careers is high versus low agency, uh, and high versus low risk. Which are kind of correlated. So there are jobs where you're taking a lot of risk and you generally have a lot of agency, um, and you also have a lot of upside. And then similarly there's jobs where you have little agency, little uh, little risk, um, and little upside And consulting is that as well? At least in the way I practice it. So consulting was kind of lower left though. Like if I've got. This two by two. I've crudely drawn, I dunno if you can see it, um, like consulting is, uh, is here cuz it's very broad and very low agency. Low risk. Uh, I didn't have really any impact in terms of, I couldn't control the outcomes of what I was doing. Mm-hmm. Starting a companies up here, high agency, high risk, very focused, you know, depth, uh, of content. Joining a company, um, is kind of depending upon the stage you joined is kind of somewhere here where like you've got, uh, you're, you're more specifically focused on one thing and you have some agency, some upside, but depending upon how early you are. You may not have a ton. It depends. And in this quadrant, I found to be more venture investing. Um, I started, uh, my venture career interning at Kleiner Perkins while I was in grad school. And, uh, this is, this resonated with me. So meaning you're doing a fair, fair, a fairly number different, uh, a broad number of content focus areas, and you have a little more agency, a little more upside, way less than you do as a founder. Um, so I, I realized that for me, this was like most authentic, like that quadrant felt best for me. There is no best quadrant. It's really a self-awareness question. And then if you wanna get super nerdy within that quadrant, you can actually draw a nest to two by two and plot all the different types of venture investing, cuz there's lots of different types of venture investing you can do. And for me, what I found is that being in the top right quadrant of the venture quadrant was most authentic. Meaning doing it with a fair bit of depth and doing with a relatively high degree of agency and upside. So that means, you know, in emergence all we do is enterprise software investing. That's all we've ever done. So we are very narrowly focused just on B2B software. We don't do consumer, we don't do hardware, we don't do space tech, we don't do anything else. We just do this. Um, and we're very focused in terms of how we invest. Each partner makes one investment per year. Um, as Alex knows, uh, when we backed, uh, when we backed Regal, I, and we more broadly, our full team went super deep. We worked really hard to hire the first VP of sales. We worked really hard to hire the first VP of customer success, and that's about like having some agency and really caring cuz we have meaningful stake in, in the company's outcome. Yeah, I think it's a great way of thinking about, to your point, which of the boxes do you want to be in? And knowing yourself well enough to know which box like you're gonna be happy in. And then following that, so, uh, as a transition, you know, uh, you obviously invested in us. And so, you know, now Regal is a, uh, you know, when you invested in us, we were a smaller company. Now Regal's a little bit of a larger company, taking advantage of building, you know, software for B2C sales specifically. So, uh, a much more transactional, much higher volume sales model than traditional B2B sales. And, you know, with that volume of data comes all kinds of fascinating things we can do from, you know, machine learning to AI to, you know, sort of lots of great stuff. So I think the, the meat of the conversation today, Uh, what's it be around AI and sales and you know, what's gonna happen? So, you know, I know Zach and I were talking before, you know, Zach, uh, is, is sort of, uh, a proponent of like all the things AI will be able to do. So maybe we start there like, how quickly, you know, Zach, do you think that. Um, you know, al comfort for roles in sales, let's put it in the most extreme. How quickly will AI be selling instead of a human? And, you know, uh, will it, will it sort of happen in certain parts of sales? Uh, will it be good for the customer or not? Like, what do you think the second order effects are gonna be as that happens? Yeah, so I'll start by saying, I think that there's, uh, a, we should distinguish between what it will be capable of and what it will actually do. And what it will actually do is a function of, um, the policy that gets passed around it. And, um, you know, the, in the Overton principles sort of applies here, which is like, how much do people, where, where do, how much does the public care about? Ai, and that's, that's accelerating really quickly. And there will be a lot of regulatory capture here. So a lot of industries will just not adopt AI because we don't want, you know, in, at least in the United States, we may not want doctors, uh, being automated, things like that. Uh, although I, I, again, I would take the other side of that, but sales is one of these categories where, There's actually a lot of, um, there's a lot of reason for the consumer and therefore the, the average constituent of a politician to want, uh, their experiences to be automated, not least of which is probably that the costs will, will come way down. Um, and the other thing that's actually gonna control whether or not this, this stuff gets adopted is whether or not the companies that are driving sales, uh, you know, actually want or doing sales, actually want to wanna take the first leap. Because it will be a domino effect. Um, industries will sort of topple as, as, as the leaders adopt this, I think that it will be capable. I think that the technology will be capable of running the average sales process. So like, pick a, pick a widget, uh, somewhere, you know, larger than shoebox, smaller than a car. Uh, a, a price, you know, less than a hundred thousand dollars. I think the technology will be capable of running that sales process in 18 months. Um, I don't think that it will, I, I would take the, I would take the over on will it, uh, but I, it certainly will be capable and, um, and there are a lot of reasons for that. The speech technology is getting really good. The listening technology is already there. Thanks to vector databases, the contextual lookup is, is remarkable, and frankly, it's probably better than the average salesperson. Um, the, I think the trouble that companies are going to have is, is, um, one, again, being the first to take that leap, but more importantly, getting their consumers comfortable with it. And, uh, you know, the idea that you're not talking to a human. Is gonna, you know, I think it's gonna fall the same trajectory that mobile did. Everyone was like, no one's gonna buy a car on their phone. You know, that's, you know, that's always gonna be protected. That's always, people are always gonna go to a dealership and you know, you fast forward, mobile has taken over purchases of every size, including homes at this point. Right. And so I think consumers will update quickly and it's just a question of how fast and without terrifying salespeople, I do think it's important that people. Prepare for a future at least where they are very augmented and probably a future where the work that they do is at the edge of, of why is it good consumer? Well, automated salespeople will never go to sleep. They don't have bad days. They don't lie to you or they shouldn't lie to you. Um, as long as the, as long as the technology is aligned, um, and they should lower your cost structure, right? Consumers who buy from salespeople, they're paying a 10% tax just on that person, right? Um, it's gotta come from somewhere. And the deflationary power of AI is the most exciting thing to me about the future. The cost of goods is going to plummet and therefore the price of goods should plummet too. And we will all just need to make less money to live. We should need to make less money to live, uh, the same or better standards of our life. And it just depends on, you know, where in the value stack we, we find the, we find that value. So let's, uh, take your position as whether it's 18 months or five years, or 10 years, there is an end point at which the technology will be able to, as long as we allow it, do the end to end job and not need a salesperson. Let, I'm making it a bit of a straw man, which is not quite what you said, but let's just pretend that for a second. Yep. So I think. Before I turn over to Jake, who I know has like some frameworks, the two by two main truc about like certain jobs where maybe, you know, you'll still need oversight. I think what we see in our world in, in sort of B2C sales at least, is there, is there is value to the human in that it creates a certain amount of trust and a certain amount of empathy that, uh, results in higher conversion. And so the one argument that like I would push back on is, is sort of the cost argument. You know, if, if having a human salesperson gets you even one more sale, it doesn't need to get you many more sales for it to pay for itself to have a human. And even if your cost structure is higher for the company, it's gonna be worthwhile to have that human. So, Yes, for sure. If eventually the, you know, technology's a hundred percent as good as a human, yeah, then there's no difference. But as long as humans are better, especially in higher dollar items, you know, if you're selling a$50,000 item, if you're a salesperson, tells you one more thing a year, uh, it's always gonna, you know, Be a situation, we're gonna want the salesperson rather than nothing. And then maybe, I mean, maybe, uh, you know, it is a better experience for our customer, you know, you know, at some point in the future, if truly you're saying, Alex, assume that it becomes indistinguishable. Sure. I agree with you. But until that point, I actually think there's this long, I would pause it. There's this long middle period where it's gonna coexist where, because the human salesperson is just a little bit better, tiny bit. It's gonna always be needed. They're, you know, they're gonna be needed and the technology will be sort of living side by side with the human. I think that the best salespeople will be a little bit better for a long time. Like if you ask, um, if you ask Andy Sabine, the president of Morgan Stanley, what makes his best wealth advisors. It's, it's an X factor. Like what, what makes the best partners at a law firm? It's an X-factor. What makes the, you know, pick a professional service, it's an X-factor. It's hard to replicate. That's what separates, you know, that's the genius zone. Most salespeople aren't good, and that is like something that we all know, but don't talk about. Very often, the average salesperson doesn't actually add the value that the company thinks that they add, and on the margin. Most people don't actually talk to a terrific salesperson when they're purchasing a good or service. I think that we should, I will take the under on the whatever, you know, whatever limit we're approaching when AI is much better than the average experience today. Yeah. So let's, let's the question, Jake, specifically, so like in this interim period where, BI is not as good as the best salespeople, but you know, to the point, like there's huge variety in the quality of the salesperson. Like what happens, like, where does this get used and where does it fall down? Like the way I view the the next few years is that this should lift all boats in terms of the quality of sales reps, right? If you have bad sales reps today, Uh, first of all, they're not gonna be able to hide as well, because the data we're collecting is so much better. And when you have, you know, recordings like, you know, uh, Regal's releasing a, a call, you know, recording and coaching product that allows managers to coach, uh, their B2C sales reps on how to improve. Like this technology should help us upskill workers. Um, and so the bad sales reps, um, should be coached and the ones that aren't gonna be coached well with this technology in conjunction with humans, will be let go and hopefully the baseline will raise in, in conjunction with this stuff. That's also hopefully true of the best, the best reps. Because like in the ideal world, this technology recognizes like, hey, when this rep used this phrase, um, this type, in this context, the deal closed 15%, uh, more frequently. And it may be the case that rep never really realized what he or she was doing that convert that, that led to those improved conversions. But if we have a system that collects that data in a closed loop manner, That actually gives the insight to the best rep to say like, oh yeah, I guess that is working and I should do more of that. Um, and oh, by the way, that same insight is shared with all the average reps as well, and hopefully they can start to use something as well. So I view that at least in the, the medium term, this technology should raise all boats on terms of sales rep quality. Yeah. The metaphor I sometimes like to think about, and I I, Zach, I was gonna ask you if you think this is an app metaphor, is. Is Uber and that I can't imagine how a delivery person, you know, I, I know I'm sort of talking Uber in general, but how a delivery person or a taxi driver 20 years ago knew what to do. Somebody jumps in the car or you get an order and says, you know, this street name in 15 minutes. Like, how could you possibly do that job? You know, maybe in London where it's a, you know, literally it was like a year long training program. You could figure it out, but it was impossible. And with. The aid of Google Maps and the Uber app or delivery app, all of a sudden to turn every driver into somebody who always got it right, always took the best route, always made sure that you were, you know, going around the traffic that you didn't have to say anymore. Oh, like, don't take Fifth Avenue because I know there's something happening. Like it knew. And so, again, maybe simplistic, but I think about that in the short, in the medium term to Jake's point. Where couldn't just all of a sudden turn every salesperson into somebody that is doing this phenomenal job for the customer. Yes. And I don't think that's simplistic at all. I think it's very analogous. The gap that you're missing is that Uber would love for all of their vehicles to be autonomous. And moreover, we're almost there if not for policy, like on the driving. So let's take the driving example. So this is one that, uh, I think about a lot just from our business. Who's gonna win when it goes autonomous? Is it Uber who has the distribution that they managed to get while they were using the drivers? Or is it the companies that went straight to driverless have never got any distribution in advance? The best explanation I've heard on this is a, is a former Uber exec who will remain unnamed, who told me that they, he, he predicts they, I'll protect his identity that, uh, There will be a bifurcated platform world, much like Android and Apple, where it's probably Tesla and Tesla has a bunch of cars on the road and those cars, you know, run on a close network and then you know, another, you know, maybe Google, who knows? Maybe you know where you can plug your car into an open network. Um, and those are the winners, but I just, just to go back to this, they think so, to be clear, they don't think that the distribution advantage Google has is worth anything once it goes to Dr. Sorry. That distribution that Uber has is worth anything. I don't know. I, I shouldn't speak. I, I'm, I'm, I'm true. Yeah. I, I, I don't think, let me ask your, forget their opinion. What's your opinion? Do you think the distribution advantage helps? I think it, I think distribution will remain king, which is also why I think ai. On the whole serves the incumbents if you can move fast enough. But if you cannot, like if you have too much technical debt and you can't update to the new world, like the, the be the most valuable currency in the new world is, uh, is ability to change your mind. Like, like updating, updating the new information as it, because it's just gonna accelerate is the thing that's gonna separate people and companies. In, in, in the new world. And that's like a combination of courage and vision and confidence and other things. But just to go back to this, your point is so good about Uber and the only reason that we are not driving around in autonomous vehicles is because of a, is is really a policy issue and there's an infrastructure problem sort of, but Cruise is, is taking people around town. If, if the government let us, Uber would quickly start to eliminate drivers. Here's where I think the analogy falls down. Um, I, I agree with that point, Zach. And like in, in my neighborhood in San Francisco, there are for every, you know, three driver cars. There's a fourth non-driver car. Like it is very standard here now. Um, but the difference between delivery and driving and sales is empathy. Ultimately sales and the best salespeople have insane eq. They're like able to incept themselves into the brains of the person they're talking to, understand what they're feeling and thinking, and then understand how to represent what their, the product or service they're selling to actually get them to buy. And that I think will remain at the highest level, like the top of the game, a very human experience for a very long period of time. It's very hard for a bot to say like, oh yeah, I feel you. I've been there, blah, blah, blah. I agree, Jake. And so let's assume that the world's best black car drivers who drive a, like the important people around will remain because those people don't want to get in a robotic vehicle. But most people are like, yeah, I'd love to spend 50% less to get around. Yeah. And get, and, and get an autonomous vehicle in the same way that most people who, the used car, the used car salesman, became such a trope for a good reason. And Yeah. But that's, that's, you're conflating metaphors here. Like it's not, it's not about selling the car, it's the service that you're providing, the transportation. I agree. And all I'm saying is I think that you're giving credit to an industry that doesn't deserve all the credit. This is coming from a sales leader. Sales. I know, I understand. I understand. This is coming from a sales leader. I think that the best salespeople are, are are, you know, operating in a genius zone. And it is a power law rule. If you ask most people who buy software, like the average person who buys software, were they impressed by their salesperson? People are gonna say probably not. Like if they're being honest. Yeah. And I think that is, and, and, and software attracts some of the best salespeople in the world. I, I'm not, I'm not trying to be inflammatory here. I just think that like, I, I, I think we are kidding ourselves if we think that like people are not, wouldn't be interested in, in saving 20% on their good and, and, and talking to a machine. Moreover, I think that machines will, will learn empathy in the same way that it depends on the good it, it depends on the good Zack, like in my mind's, like what an Uber is providing a commodity, which is getting you from point A to point B. If you're selling something that is less commoditized and particularly has more of a services element to it, that's like, hey. I'm selling wealth management and that a lot of that is like the experience you have interacting with me. Like that's not, but that's not, you nailed the point. Wealth management is a com commod, like at the limit, most things commoditize and, and the limit in this case is not that far away. Yeah. I don't know. I think, I think that, um, just I'll add like one thing that surprised me in, in, in sort of the sort of industry moving online, retail came online first, then. Retail like quickly did away with customer service, right? They hid the number, they took the number away. Like there wasn't a number, you couldn't contact any human being because they thought to your sort of general line of thinking that the more that you took away the cost, it, you know, it would save you money and it wouldn't hurt your conversion. And I think in some industries that sort of became true, like selling CDs online, that probably became true and that industry disrupted itself and then went a whole different way with, you know, Uh, Spotify and other things, but when I think about, uh, industries like healthcare, you know, uh, where you're sort of deciding between, you know, I, I hurt my knee not too long ago, and one doctor told me I'd never walk again. One doctor told me, uh, that I needed surgery today to walk again, and another doctor told me I was a wimp and I should walk out of the room. So, you know, you're deciding between these very complicated outcomes. And you know, effectively the doctors work salespeople. They may not call themselves salespeople, but they're onboarding me and trying to convince me to work with'em. Or if I'm talking about life insurance and there's the struggle between, do I leave it to my kids or my wife or my, you know, somebody else, or if I'm talking about an industry like education where it's a very emotional decision. Do I take time away from my evenings while I'm working full-time to go to school because I want to have a different career and try to move out of my current career. These things we tried to bring online, so I was part of trying to bring some of these industries online and I was shocked and I shouldn't have been. I was shocked when removing the human component from an entirely resulted in much lower conversion rates. Now, think as possible. You were just, you don't fund the account, so you didn't, it's possible you were just early. Yeah. So I do, I buy way in the future when truly a AI is the same as a human. Of course I'm not, you know, there's no, there's no act. If I can't tell that I'm talking ai, then this is a moot point. We're all in violent agreement that when the AI is as good as a human, of course it's the same. I think in the short term though, this is the question. When you have these complex decisions and the AI is not as good as the human, what is the human gonna want? What, you know, what is the business gonna want? And I think to Jake's point, this is where I agree with him, was like there is something. Even if the salesperson is not perfect, where the AI will lift the salesperson before or replaces them. Alex, do you think it's also possible that you are getting premium, premium experiences when you look for a doctor or when you consider a major purchase? As I think about these industries, like, and I'll pick on one that's famous so that I'm not insulting anybody, like. You know, uh, uh, buying a car to your point is not necessarily a premium experience, but convert. I know conversion rates are higher with a human, without a human going and submitting, putting money in a bank account, again, not a premium experience, but convert, the rates of funding accounts are double, if not tripled, when you have a human being involved and without a human being. I, I'm not saying that it, it, uh, I, I'm, or I'm just arguing the technology is gonna get where it needs to go. The last thing, the last point I'll make on the healthcare and, and just broadly. You know, of the, of the eight something, 9 billion people on Earth, something like 70% have never seen a doctor. And I think it's really important when we talk about, uh, the, the power of ai, it, it means accessibility to things that are considered luxury today. Turning luxuries into staples. And a lot of that is going to require fully automating an experience because we're just not training enough doctors right now. Yeah, that's where there are a lot of people who are unbanked still in the United States, and if we can figure out an effective way to bring them into the financial system where they can start saving money, but that has to happen in an automated way. We're going to do it like, and it's so that, that's where I push you as an example. So let's, you know, I, I, I, I buy, like the value of the internet is increased distribution. For sure. And the value of of AI and like the kind of data that we get at Regal is that we can offer millions of customers a one in a million experience. So my wife's friend is, uh, an insurance agent. She used to sit in the physical office. She's fantastic. People would walk in, they loved her. She's incredible. How do we create that same experience or take your medical example, like how do we take the, the experience you have with the best doctor in the world, but offer that to millions of people by leveraging. All the things we know about how all these different, uh, doctors operate, how patients react, how to deliver good news and bad. And so, yeah. Is it possible that there, uh, there's gonna be more technology, of course, but, you know, we serve, for instance, a diabetes, um, care company where, you know, to your point, they don't have many nurses where they use our software to identify the moments in which technology can do stuff and the moments in which a human needs to get on the phone and say, If there's something going on, you need to talk to me right now instead of seeing a doctor for four minutes a year. You know, you know, at one point in that one, you know, wrong point. They're gonna make sure that you have the 32nd conversation at the right moment when they're gonna save your foot from being cut off because they identify something that's going on and, and, and I agree that the, the line is probably asymptotic, that there is always going to be work that is done where a human will be in the loop always for some amount of work. All I'm arguing is that, The curve to, to that line is, is much more accelerated than you, than you and Jake think. So Jake, I guess you've talked about sort of some, uh, ways of thinking about this, particularly in like how critical the function is. And I think your criticism of the technology is that it might hallucinate, not intentionally lied, but unintentionally lied. So like, do you think, Jake, that's that sort of your opinion changes over time? If that goes away, if the technology was not gonna hallucinate. Yeah, there, there's two dimensions. Think about this. So the first is accuracy, which is related to hallucination. So like how bad is, like how often is this thing wrong or lying? Uh, the thing being AI and, and if they're wrong or lying, how severe is that consequence? So like accuracy as well is just like outcomes. Like that's, that's one dimension. And the reality is like, that's going to get better over time. Like, technology will improve. We'll f we'll find better ways to fix that. I think we have a long way to go on that. So like there's a lot of pay to be made, if that's the right analogy, uh, over the medium term until that gets fixed, I. Um, but I think until it gets fixed, like people really, like human in the loop is a very, is a much safer way. Cause you can nudge the human and if the AI is wrong, the human can override it. So I do think that's gonna be a really important thing for the next era of software. Um, I think that's a whole era of human in the loop software that's going to exist because, uh, of hallucinations in a business context, it matters less than consumer context because like if you're chatting with a dead celebrity, there is no such thing as right or wrong. Um, but to me, even in a world where we get to a place where like this thing is fully reliable. I still believe that in really high value human to human transactions, a human being involved matters. Like, you're not gonna close, you know, a, a really important, really emotional transaction without a human being involved. And I don't think a car is a good analogy cause I don't think that's an emotional transaction in the same sense. I, by the way, I think you're mentioning, I agree with this mostly. I don't think we disagree here. I just think that, that, I just think that, that, that market is like, yeah, that moment is fast. So, so let's go back to the point we were making with Zach before around distribution. So, Jake, I know you've talked about. Uh, uh, this, this sort of idea that, you know, uh, perhaps the companies that have a, a sort of a UI already that, that they're sort of building a software for a job to be done already that have distribution already may have the greatest advantage. Now this new technology, and it may, uh, be, this may be a boon for incumbents, you know, how do you think about like, who's gonna take best advantage of this technology? Yeah, so, um, I'm thinking a lot about this right now cuz to out, to Zach's point, um, in many ways incumbents are best positioned here. This is different than the, the moved from on-prem to cloud when incumbents were into trouble because it was impossible, basically impossible to re rewrite your entire stack to be cloud first. This is as easy as plugging in api. And effectively, like most public companies are not playing with this, um, and doing like cool stuff with it. So the question is for startups, like what is the, what is the opportunity relative to the incumbent? And I think there are, there are potentially a few, there are opportunities where, what was the job to be done that couldn't have been done before with technology had to be a service before. Um, and there's a lot of examples of that. There's a bunch of interesting stuff happening in legal tech, which was primarily a, a service before, but it's now being able to be productized more. Um, there, uh, there's all sorts of stuff happening in the medical space, uh, as well on that front. So like, and, and those are places where incumbents may have, uh, an advantage, but in often, in many places, they were services companies, not software companies. And so software companies as startups may have a, a role. I think secondarily, um, there's a bunch of industries that are gonna be created by this. Certainly all the tooling, infrastructure tooling to support this is new. Um, there's also gonna be a bunch of companies around compliance and, you know, trying to implement this stuff in a safe way that don't exist yet that are gonna have to be built. So that's another opportunity for startups. And then lastly, I've been thinking, and we have been thinking a lot about. Um, where is there a likely to be an innovator's dilemma issue for the incumbent, or where is there likely to be a counter positioning opportunity for the startup? And so examples of that may include, um, you mentioned before user interface, uh, changes. The, the chat-based interface is sort of a new concept in B2B software. Um, and I think it's unlikely that chat based interface on its own is going to be the dominant interface going forward, but I think there are gonna be elements of chat based interfaces that become core to many B2B software experiences. And I think that many incumbents may have, uh, a challenge of cannibalization of their core business if they try to make that move. So, as an example, let's imagine a world where the ideal CRM going forward we realize should be built with a chat-based interface. And, and I don't know if that's true or false, and Alex, you're gonna know more than I do cuz you're building a CRM product. But let's imagine that that's gonna be core to that product going forward. If your Salesforce. It's not that you don't have the resources to rewrite your software to do that, it's that you have millions of daily active users who pay you billions of dollars who are used to a certain interface. And so completely changing that risks fundamentally disrupting your core thing, which is unlikely to happen cuz there's just too much incumbency there. So I think there's some stuff like that. I think there can be some pricing changes that happen in software companies that got used to charging on a proceed basis, um, may be more vulnerable in a world where, you know, a new startup can say, Hey, I'm gonna charge based on outcomes. Alex, one of the things that that Regal did that was most impressive to us as we were doing the, uh, diligence on the investment is how closely you tie the interventions that you guys have to outcomes for your customers. How much increased revenue you guys are able to show. And in many cases, and this is like a super cool thing that I talk about a lot and maybe, I dunno if you'll mind if I share it, Alex publicly, you'll often actually hold back reps. With customers. So you'll say like, okay, the customer, 95% of the customers reps can use our product, but 5% can't. Just to demonstrate the increased, uh, ROI that we're driving, what is the real impact? You can imagine a world where as you add more and more AI and that impact becomes even greater. You start to shift to a model where it's like, Hey, we're gonna charge not on a Percy basis, but we're gonna charge on the incremental dollars we live for you. Which was really hard to do in a world before because it was hard to establish causality and most software vendors didn't have that much, you know, quantifiable impact that they could prove. I think that's starting to shift. I think Regal's sort of at the forefront of that, and I think that if you're an incumbent who's used to charging in a certain way and frankly used to obscuring the value, create behind some sort of proceed or per usage metric, you may be vulnerable. So thank you, Jake. So I think a lot, a lot sort of think about today. Uh, I know Zach, you gotta run. So we're gonna, we're gonna stop there for today, but appreciate you guys opining on how quickly we're gonna get to AI replacing sales and in the meantime, sort of what businesses can do to take advantage of AI in this, uh, in this, uh, interesting industry. So thank you. Thanks, Alex. Good to see you, Jake. See soon. Hi guys. Cheers.