Subscribe
Share
Share
Embed
Demystifying the conversations we're already here at RRE and with our portfolio companies. In each episode, your hosts, Will Porteous, Raju Rishi, and Jason Black will dive deeply into topics that are shaping the future, from satellite technology to digital health, to venture investing, and much more.
Raju: You know, using that unified data record, use AI to figure out how to make Will Porteous better, but I need all of the data across multiple devices.
Will: That’s going to take a lot of money and generations [laugh].
Raju: It’s going to take a lot of money and gen—well, no. You’re too healthy already, you know? You go the other way if you want [laugh].
Will: I’m Will Porteous.
Raju: And I’m Raju Rishi. Welcome to RRE POV, the show in which we record the conversations we’re already having among ourselves, our entrepreneurs, and industry leaders for you to listen in on.
Raju: Hello listeners. This is Raju Rishi, and I’m joined by my partner, Will Porteous. Today’s episode of RRE POV is focused on those companies that were started in the pre-AI era but are now living in a post-AI era. What do you do? “AI or die,” or, “AI in time,” or, “AI doesn’t fly,” or, “Deny, deny, deny,” or, “If the glove doesn’t fit, you must acquit.”
Will: [laugh].
Raju: Okay, that one doesn’t really belong, but I like the rhyme. It was kind of just going with the rhyming flow. You know, just get into a little thing, William. You know how it is. I’m a kooky guy.
Okay, so we all have companies that have been around for five-plus years, built sizable businesses, and now we walk into a boardroom and a board meeting, and the AI topic comes up, and so, how do we address it, right? So just, Will, I’m going to decompose this into a few different dimensions because AI can actually have numerous impacts on a business. And, you know, let’s start with the product or solution impact. This is a scenario where, like, you have a product in the marketplace, and now an AI is, like, you got AI competitors that are trying to do the same thing, or you know, you’re nervous because, you know, there could be some displacement. And the question really becomes, can an AI-native competitor come in and build a better solution? Are we, you know, sort of on the tail end of our business?
And then the second question is, you know, sort of, what are our competitive advantages and moats that we can leverage to, you know, basically judo this thing, and you know, come out with AI-native capabilities, but do the things that are hard for the AI people to do, but we’ve already kind of got them locked and loaded. So, I’m going to start there. I’m going to talk about just some areas where, you know, look, we don’t really have a whole lot of portfolio companies doing that is scary, right? Like, you know, like, Google should be worried, right? Because—
Will: [laugh] right.
Raju: —yeah, search, I think design is going to be an area where, you know, you had to pay someone, or you might have, sort of, you know, earlier tools and now it’s going to become commoditized, and there are engines in place that allow you to do things. But let’s not focus on those because I think those are pretty obvious to people. Is writing going to be, you know, kind of change forever? Is, you know, search going to be changed forever? Probably. Like, we know that’s coming.
Let’s talk about, kind of, the safe spaces a little bit. You know, do you have any companies that you’re sitting there saying, like, hey, you know, these companies are kind of safe for the time being. They got to think about it long-term, but they are safe for a period of time. And if you have anything that comes to mind, I’d love to share it. I obviously have a couple in mind that I can share as well.
Will: Well, one that immediately comes to mind that’s perhaps not obvious to some of our listeners is our smart baby monitor company, Nanit. So, we were the founding investors in Nanit, and today they serve over 750,000 families. They’ve grown to really interesting scale. And parents spend three to four hours a day in the Nanit application, and Nanit provides them with all sorts of insights about their baby. But what’s happening in the background is the Nanit camera is a computer vision-enabled device that can look at the activity and recognize the activity of the baby, can tell how much the baby slept last night, can tell when the baby reaches certain developmental milestones, and using computer vision, can classify those activities and compare them against what is actually the largest store of imagery on infant development in the world.
Nanit was an AI company before this AI wave, and the insights derived from that proprietary data set allow us to give new parents a lot of constructive advice, not just on how to teach their baby to fall asleep, but also how to understand the infant’s health. And so, I think hardware companies generally that can produce proprietary data are in a pretty safe spot in the AI era because they can build on that data, they can fuse that data with other data sets to generate AI-powered insights. To me, that’s a really viable strategy for a lot of connected hardware companies in the present era.
Raju: I love that. I love that company, also. I mean, I got to be honest, Will, I’ve kind of… our CEOs have babies. And I tell them, like, hey, hold off. Like we need to get this thing, you know—no, no, I don’t. I actually encourage people to have kids while they’re startup entrepreneurs. But we always ship them a Nanit, and… they love this thing. Like, they were like—I mean, I get more compliments on that product than any other solution. I’m sort of like, hey, you know, free RRE jacket. Well, people do love my sweaters. Like, that is a thing. That’s a known thing. But aside from the, you know, sort of sweater thing, that Nanit it is, like, loved, and I agree.
I want to build on what you said. I think that there are four or five things that companies have that they can leverage and they can use as competitive moats or advantages for their business. And you really, really want to double down on those advantages while you play either catch up with the AI technology or, you know, you kind of create additional capabilities that AI can’t, you know, execute easily. And so, you know, one of those advantages is definitely an existing customer base. If you’ve got deep-seated relationships, contracts, purchasing agreements, you know you’re on, you know, the government’s list of, you know—you know what they call it. I don’t have too many government companies. What was that list called? DSI?
Will: Yeah. Yeah, the general schedule.
Raju: Yeah. There’s, like, GSI, which is government, and then there’s state, SLI, which is, like, state level. Like those things are hard to get, you know? Like, FDA approval, all those kinds of certifications, and so you have an advantage when you got existing customers, for sure. The second, I think, deep advantage some companies have, if they’ve been a law around a while, is data.
And, you know, Nanit certainly has it. It’s got image libraries up the yin-yang, right? We have another portfolio company called Imgix. Imgix has, you know, thousands of customers, they’ve got billions of images that they use to essentially accelerate websites. You know, like, a customer of theirs could be, you know, I don’t know, let’s just say Nordstrom or something like that, a retailer or it could be, you know, like, a home buying selling platform, which, I know there’s dozens of them, but there’s huge ones—these are all customers of Imgix—they keep that image library.
They have billions and billions and billions of images. Now, you have an existing customer, you have their images which they’ve given you, and now you want to do manipulation of those images. You want to basically stretch the image, you want to give a different background, you want to basically, maybe even turn that image into a video clip, which is really, really interesting. Now, generative AI allows you to do that, and there are platforms being created that allow you to do that, but that doesn’t matter. That capability, Imgix could build that capability very, very quickly—and they have, frankly—and they are the trusted provider, they own the images, they’re integrated into the workflow, and so it is an absolute natural fit that they would just provide that service. And so, like, Nanit is sitting on a lot of data, Imgix is sitting on a lot of data. Both companies are sitting on a lot of customer relationships. It’s really, really hard to displace that.
Will: I think that’s such a powerful example and it kind of points to the fact that in the AI era, if you have the data that provides enough of a stub, so to speak, for you to build on or act on, from a generative AI standpoint, you can do all kinds of things around that knowledge and Imgix is already operating at the scale of billions and billions of images for their customers, so the fact that they can step forward and provide rich AI generative experiences that they already know are on target for customers that they already serve is, I think, a huge lock. And you’ve been a good shepherd of that company for a long period of time, and I know you’re seeing a pretty incredible inflection in this environment.
Raju: I am. And I love the CEO, I love the management team, and you know, they definitely are not afraid of reinventing themselves. You know, I mean, listen, the bulk of the revenue is going to come from existing capabilities and existing functionality that those customers love and need, but if those customers are ever thinking about AI imagery, they’re not going to go anywhere else, you know? And it’s lovely to see that the management team and the business itself and Chris Zacharias himself is building some of those solutions.
So, I’ll just play it back. Like, if you have an existing customer base, if you’ve got proprietary data, if you’re integrated into workflow, if you’ve got multi-legged, competitive, complex tasks, and the last thing I’ll say is, if your company deals with PII because PII, you know, it’s going to take a little while before that data is going to be allowed to be manipulated in at least external, you know, sort of AI engines and LLMs. Like, you can create a proprietary version, but you’re only going to get limited leverage on that. So, I think you have modes. And so, you have to be thinking, as a company, how do I take those and give myself enough time to, sort of, build the right level of capability, and what is it that the customer really wants to buy? And you know, I wouldn’t be so terrified of the AI-native platforms.
If you don’t have a lot of that, there’s a little bit of a different story. And the last thing I’ll say is, like, if you’re doing physical tasks today—like, we have a company called Getlabs, and they basically are mobile phlebotomists. So, if you need a phlebotomist in your home to, you know, draw blood or do certain tasks, you know, they’ll dispatch. And I think the difficulty of having a robot do that, [laugh] you know, is quite some time away. So, I’m not so nervous about that. The thing AI can help them leverage route optimization, for sure, but like, not necessarily drawing the blood. So anyway.
Will: [laugh]. Maybe not drawing the blood, but as you and I both know, analyzing the blood, there’s a role for computer vision and AI in that department.
Raju: Absolutely.
Will: That perhaps is a topic for another day.
Raju: That’s a topic for another day. It does require a bunch of FDA approval, and there’s going to be some [derivation 00:12:32] before we get there. Okay, so let’s talk about operational impact. We talked about product and solution impact, which is like, hey, this is what I’m selling. This is what I’m making, and I’m selling. Let’s talk about operational impact.
And you know, I’m going to—Will, I’m going to take the liberty to call out during this section, some of our portfolio companies that are actually tools, AI-native tools that the world, our listeners and other companies should be, you know, sort of leveraging because they will help you do things. I think this is a really, really important area. And no company—zero companies, are immune to this piece, which is, how do I leverage AI tools, AI capabilities, to you know, improve my operational mechanics? And I think everybody should be thinking about it.
So, let’s start with sales and marketing. You know, this is a function that historically has been done with, you know, sort of like, tools. You have a marketing engine, you have a sales engine. Marketing is a lot of stuff. You do webinars, you do, you know, outbound marketing, you do social marketing, you’ll do a bunch of optimization of your website, SEO, SEM, I mean, there’s, like, a big gamut, and then sales, you know, and sales has a lot of tools, like, you know, gong lets you listen to calls and record things, and you’ve got a bunch of ops. I think this area is super, super ripe for AI assistance and AI.
Will: Yes.
Raju: If you’ve got a few things to say about this topic, jump in, otherwise I’m going to describe a couple of our portfolio companies and how they’re working.
Will: No, you go ahead because there’s some specific companies I know you want to call out.
Raju: Yeah, we’ve got two in our portfolio.
Will: I mean, I think we see some things pretty broadly, in terms of the implementation of AI, in terms of generating responses to RFPs, in terms of generating better—kind of, just being smarter and faster in terms of turning things around for the customers. But you’re thinking about some specific use cases, I know.
Raju: I am, and the only reason is, like, sales and marketing—at least for B2B companies—you know, has been, you know, it has gone through an inflection over the last, sort of like, ten years, right? You know, even, maybe even before that, we moved to, like, you know, spreadsheet-based analysis of when deals were going to close, we you know, moved to SAP and Siebel and PeopleSoft, and then obviously Salesforce and HubSpot. Marketing has gone through that same transition. And so, you know, we use digital marketing, we kind of log. Where,we’re trying to do things like attribution and figure out, sort of like, what’s working, what’s not working.
We actually have two portfolio companies, one called Octane11. Octane11 looks at all of your ad and your marketing capabilities, everything that you’re executing, and munges that with external data and gives you a sense of what is performing well and what’s not performing well. And it is using AI to really help decipher where to double down and where to stop spend and execution from a marketing perspective. Avina is a very complimentary company is looking at your entire sales and marketing operation and telling you things like, this customer, you know, would benefit from getting this type of collateral, or this particular, you know, salesperson should be sending out these materials to try to close these types of deals. And they automatically analyze all, ingest all of the data and give you specific, candid, concrete feedback on next steps that will move your sales funnel forward.
I think these companies—well, we wouldn’t have invested if we didn’t think they were game changers. They are absolutely game changers. I think the reality is, if people are not incorporating these kinds of tools, and they’re just kind of relying on, you know, sort of like intuition, or, you know, like, human-manipulated data, it’s a problem.
Will: Well, you know, the thing I love about this is it solves so many classic problems in the scaling of sales and marketing in venture-backed companies. You and I, for 20 years, have watched the growth of sales organizations in our companies. And I used to say to my CEOs as they were scaling, plan to hire twice as many salespeople as you think you’re going to need because your failure rate is going to be at least 50% [laugh]. And the training journey and the learning journey and all of those things used to happen so much in isolation, and tools like this are transformative in terms of producing powerful teams early in the life of the company. So, I love what Avina is doing.
Raju: Yeah. And I do too. And I think part of the trick here in sales is, you know, you run a process if there’s software that can help you run that process faster and better and make average salespeople, above average and above average salespeople excellent and excellent salespeople astronomical, you know, that’s worth every penny, and these tools do that. Okay, the second area where I’m seeing a lot of question marks from my CEOs, and, you know, I’d love to—we don’t have a portfolio company in this space, but I think the reality is, you know, like the big LLMs out there are providing some capability, is development, you know? What’s the impact of vibe coding on my business? And I’m going to let you start this topic. I’m sure it comes up in board meetings where people say, like, you know, like, “What should I be doing here?” Right? We’ve got an R&D organization that’s ten years old, or five years old, and, you know, we don’t have really native AI coders here.
Will: You know, [laugh] I was in a board meeting for one of our companies earlier this week that was founded by a pretty well-known guy in the AI world, and he prides himself on being very much personally into the architecture and the development of the platform product that they’re building, and he said to the board, “We’re capping hiring on the development side right now because we are getting so much leverage through cursor and vibe coding across the company.” And he said, “You think there’s only one of me here, but there are actually three of me here because I have trained agents to code the way I code, and they’re out working while I’m in here with you guys.”
Raju: That’s hilarious.
Will: To me, this is just an unbelievably powerful unlock in the scaling of innovative companies. It’s a, frankly, very powerful point of leverage for a firm like ours, and I’m seeing our portfolio companies do more with less people and less capital as a result of this. So, to me, this is the big acceleration that AI is really bringing to our industry right now. And we hear these stories every day from the companies.
Raju: Yeah. I will tell you, like, there’s been a handful of unlocks that have changed startup dynamics, right? I’m going to give a shout out to our friends at AWS because when AWS came out, pre-AWS, everybody used to have a rack and you had to get a network engineer and you had to plug these boxes in there and it was uptime and everything like that. And that AWS came out, it was like almost overnight, people were like, “I don’t need a rack anymore.” Like, I can put, like, all my code in the cloud and it can render efficiently and, you know, like, they can deal with the akamai effect of not a CDN, but, like, distribution of the code and execution of the code in the cloud.
And that had dramatic ramification on the cost profile of startups and the location, like, and the talent needs. Kind of the same things happening with vibe coding and cursor and all the things here. And the reality is, it is an issue, and I think every single company needs to understand, what does my R&D organization need to look like? And you know, if I’ve got a bunch of, you know, sort of 5, 10-year-old engineers, and like, 30-year-old engineers, and nobody knows how to vibe code, you are going to be at a severe disadvantage from other players that are leveraging this. This is not an option for companies.
You need to learn. You need to bring in somebody who’s—you know, bring in a younger guy or gal that knows how to code using these tools, and you need that proliferation of knowledge and you know, creation needs to go through your organization. They need to learn from the younger talent. And if you have people who aren’t willing to learn, they need to go. And more importantly, you don’t want to lose all of your knowledge, subject-matter expertise because the reality is, like, yes, you know vibe coding and cursor-based coding and all those tools, they’re still going to make some mistakes, so you have to have test points.
But you know, over time, they are going to be less and less. And you got to go through QA anyway, you know? I mean, and so you got to do unit testing, do system tests, you got to do all—that still has to happen. But if you can create code 10x faster, you know, the trick is, how do you change an organization that’s doing things for ten years a certain way and shift gears, and I think a lot of that is hiring. Like, bringing the right people in that can, you know, I don’t know about threaten people, but like, they’re looking over their shoulder and saying that guy just did three times the number of, you know, module creations than I did. Like, I better get on board here, right? Like, and you know, you do need senior people, but, like, I’m seeing this all the time.
Will: You’re ringing the bell, I think, loud and clear, and I hope people are hearing you. There has to be a great sense of urgency around this. It can’t be an incremental move within your development organization. You’ve got to be willing to essentially burn the boats to some extent, around old approaches, or at least empower a subset of your team to go off and start rebuilding things in this fashion because the wholesale transformation of this is underway, and it’s the biggest point of leverage in software development, certainly that’s come along, I think in our lifetime.
Raju: Yeah. I agree. Okay. We’re going to move to customer service and I have a company I’m going to mention here is one of ours, and I’m very excited about it, of course. But this is going to change, right? Like, if you are in the business of selling to enterprises, and you’re in the business of selling to consumers or whatever, you have customer service, or, you know, customer support, whatever you want to call it, account management.
And you know, voice agents are going to be real. You know, I would say that they’re like a bit behind the vibe coding framework and the acceleration of that, but they are coming, right? And so, I’ll frame this for our listeners. There’s, like, when you think about having a conversation with an agent, and you know, there’s actually three LLMs involved. There’s an LLM that does natural language processing, which kind of figures out what I’m saying, removes the background noise, what’s the context and more importantly, am I angry? You know, what’s my emotive, you know, sort of demeanor.
Then there’s a second element LLM that says, how do I respond? Like, simple query is easy, but like, you know, if somebody’s really pissed off about a product they bought or it’s not working properly, you know, there’s a lot of nuance there. And then the third LLM is, generate a voice that sounds human, that is believable, and that whole thing that needs to be done in a maximum of 700 milliseconds. And so, it’s a complex task. There’s a couple of ways to do it today. There’s a black box approach where, like, all the LLM sit in one black box. It’s good for simple solutions.
And then there’s, you know, down the road where you kind of like, build from scratch, and I think it’s too complicated for most companies to do on their own. And this is why we invested in, you know, [PrimAI 00:24:56] or VoiceRun—you know, name is in flux right now. But that company effectively creates a solution that gives you the best of all worlds: allows you to integrate with the APIs that you need, allows you to pick the LLMs of choice, and allows you to run this solution either in-house for things that have highly sensitive PII or they don’t want to sit in the outside world, or run in a virtual private cloud or just a public cloud. And so, you know, I just see the handwriting on the wall. And I think that there are certain companies that, you know, have heavy amounts of engagement, you know, on a voice level, that can drop their costs dramatically here.
Will: Well, and I know that, as someone who’s watched a lot of enterprise companies scale, that middle ground strategy was so important in your analysis because it allows the customer to take advantage of assets that they already have, that they can’t let outside of the firewall, that they can’t share with a large language model. So, this is going to be really powerful.
Raju: It’s going to be fun. It’s going to be a fun world. I think I’m going to really enjoy—like, I really do love, you know, our jo—we are lucky to be doing our jobs, and I love working with you, and I love working in this space, but I think, you know, you kind of rethink all of the little tool sets and all of the different, you know, sort of peculiar of how you do this business. A lot of stuff is, it’s in cruise control because you don’t have to worry about it, but I think you got to, sort of like, pay attention all the time now, which is kind of fun. I like that.
Will: Amazing.
Raju: Now, the last, sort of, internal operational piece is just finance. And finance has been done pretty much in a routine way, you know, historically, and it’s going to change. And you know, you have accounts receivable, you got accounts payable, and you got a planning function, and I think all of those are going to effectively get impacted by AI. And I just want to know, like, in your companies, have you sort of, like, went down this pathway in terms of discussing how the financing role is going to change as related to AI, or is it kind of a little bit further away?
Will: I think in my companies, it’s still pretty theoretical, and I think everyone knows that there’s a role for the AI in understanding complex contracts, in being declarative about certain—the treatment of a booking and when a booking becomes revenue, and the implementation of accounting policy and the understanding the cash cycle, and all of those things, AI can not only help us understand well, but bring in insights from a wider world of companies. But I’m not seeing that yet in our portfolio.
Raju: Okay, I am not either. I don’t see a lot of people moving the needle, except we do have a portfolio company which is rocking, and I’m going to talk about it here. And for those who have a lot of accounts payable—or even receivable, but mostly payable here in this situation—we have a company called OpenEnvoy which you need to incorporate into your processes. So, if you have a, sort of, rigorous accounts payable model, like, you have a lot of vendors you’re dealing with, you have a contract with each one of those vendors. So, I’m thinking, like, maybe a manufacturer.
A manufacturer is maybe, you know, a thousand, maybe five-thousand different vendors that they’re dealing with, and, you know, you get screws and parts and yadda, yadda, yadda. And there’s, like, a structured book on how much you know each part costs, and you know what it’s going to cost if you do it with volume, so if you exceed a certain threshold, you get a deeper discount. Well, if you have, like, a thousand vendors that you’re dealing with, you probably have a thousand invoices every month, and that’s being dealt with manually. Well, OpenEnvoy is a brilliant solution. It basically ingests all of your contracts, including all of the discounting structure and when it pertains, what your net payable is—like, is it a net 30 or a net 60 or a net 90, and what your discounting structure is, of course.
And then it ingests all of those invoices every month, and it basically will tell you, hey, you should pay this because it’s, you know, correct, but we’re going to pay it on day 29 hour, 23 minute, 59 because we have a 30-day net payable, so wait until the last minute, or invoice is incorrect. You’re supposed to get a deeper discount because you exceeded a volume threshold. And so, you know, we’ll wait some time, we’ll send an email saying, “Hey, you know, this invoice is incorrect. We’re supposed to get a, you know, 7% discount. We only got a 6% discount.”
And then, you know, when did you get that invoice you get another net 30. Using AI to do that task can save you so much money. And it’s a no brainer. It’s a no brainer, and it goes deeper than that in so many different dimensions. I don’t want to spend the whole time talking about, you know, sort of how this portfolio company works, but listen, some of these companies that we’ve talked about in terms of the operational impact, like, Octane11, Avina, you know, PrimAI slash VoiceRun, and OpenEnvoy, I think they can actually be beneficial to most companies out there. And if you’re a listener and you’re kind of trying to figure out, like, what do I do around AI, at a minimum, you should be looking at some of these, you know, tools.
So, the last thing I’m going to ask you about sort of operating—or, you know, not the last thing. There’s two more impacts I’ll ask you about, and then we’ll, kind of, wrap up. Like, one is the financial and financing impact of AI. Can you get funded, you know, in your next round? Because, like, hey, I’m a Series C company, or a Series D company, or whatever, and I need another round of financing. You know, how do I tell a story, you know, here? What is the implication here? And are you, you know, having that dialog with any of your companies? You don’t have to name them.
And the second thing is—and I will, you know, mention both because I—you know, sort of part of the same bucket is, like, pricing needs to change with AI. And I’ll tell a little story about this in a second, but I’d love to hear your thoughts first, Will.
Will: Well, I think the leverage that people are getting through AI in the scaling of their companies is now the new model for what you should be able to achieve on whatever financing you’ve raised to date. And so largely, I think this is bringing—it’s raising the standards in terms of how mature your business should actually be before you need to raise another round of financing. I’m seeing teams today that are capped at 20 people that are being immensely productive thanks to the use of AI-based tools that would have grown to be teams of 40 or 50 people, and needed to raise a lot more money in the past. And so, I think that if you’re not absorbing and using the tools and the methodologies that we’ve talked about today, your company-building model is out of date and the financing market is going to react to that when you go to seek a new round of financing. So, he heed these lessons at your peril because it’s a big change.
Raju: That’s a super good point. You’re right. That’s right. And you know, I’m encouraging my company to, like, think about, you know, not necessarily raising a blowout round of financing, but being more mature about it because the reality is, it will tell a story to your next financer, whether you have thought through how to make your own company more, you know, sort of like… you’re leveraging the tools to get the efficiencies in your own business that, you know, makes it so that you don’t have to raise an enormous amount of cash, and then, you know, just double down on sales and marketing until you get through the trough. I mean that kind of lifestyle business is over. I think you got to be able to tell a story of, like, why you have resilience versus an AI competitor that’s just out there.
Will: Sure.
Raju: You know, do you have the 10,000 customers? Do you have, you know, sort of like the data assets? Are you leveraging internal AI tools to make yourself more efficient from a vibe coding perspective, and/or, you know, OpenEnvoys and the Avinas and the Octane11s and the VoiceRun capabilities that we talked about? So, we’re leveraging AI to become more efficient and most importantly, like, we are actually going to incorporate AI in our product, where necessary.
Will: Well, I was just going to add, as you know, I work closely with one of our portfolio companies that has a very well-known consumer brand, a very large installed base, a lot of data about those customers, but they also have a product experience that is vulnerable to generative AI. And I think that they have to run really hard to embrace new generative AI experiences in their product methodology, and that involves, kind of, teaching their end-users how to take advantage of those kind of capabilities. And it’s a bit of a product wake-up call, and they’ve always competed on the strength of their product experience. Obviously, we’re big believers in the company, and it’s a very well-run business that has been a consistent grower and profitable for a long period of time, but this is very quietly, a moment of urgency for them, and it’s a moment of urgency for a lot of established software companies.
Raju: I agree. I agree, but the one thing that I think you know is an [undeterminant 00:34:40] right now is pricing. And I will tell you, you know, like, SaaS—I’ve said this to other people—like, SaaS is dead. And, you know, just to be controversial, you know? It’s not dead, right? There’s lots of companies that offer tools and capabilities that don’t require a lot of processing power to, you know—if you incorporated AI into it, wouldn’t change their, you know, gross margins very much. And then there are others where the gross margin is going to really be impacted if you incorporate AI.
If it’s a text question, like, [acknack 00:35:12], I ask a question, I get some feedback back, and it’s kind of simple, like, you know, the SaaS model lives well, but if all of a sudden you’re making a query, and you got to create an image, or you’ve got to generate voice, or you’ve got to generate video, that’s incredibly expensive processing power, and if your customers use that at a SaaS pricing level that they currently have, but you layered those capabilities in, your margins would go underwater. And so, you know, there will be a pricing shift that goes on in this market, and I think we’re going to move back to token-based pricing. You remember, like, cell phones, Will, like, the cell phone minutes—
Will: I know where you’re going [laugh].
Raju: Like, so we get these buckets, right? And it was like, okay, you know, you get a thousand minutes or a hundred minutes, and then there’s this overage bucket and, you know, but if you pre-buy it, you know, it’s cheaper, but you know, if you go over, it’s very expensive. I think we’re going to move down this route with AI, and it’s going to be like, got to get people to pre-buy tokens. And tokens, it might be one token for, like, a text query, you know, like, a token is, like, four characters, I think technically, right? So, if you ask four characters and get four, it has two tokens.
But, you know, let’s say you’re going to generate, you know, an image, that might be 100 tokens, you know? If you do video, it might be, you know, 10,000 tokens per two seconds, or something like that, you know? And so, I think that model is going to have an impact, and those costs are going to actually have a real impact. I think the reality is, in today’s world, we’re seeing a lot of the LLMs hide the costs.
Will: Oh, gosh, yeah.
Raju: And so, you pay X, and you’re like, just get people to start using it like it’s crack cocaine or something like that, and then, you know, eventually you jack up the prices. But I don’t know if that’s a hundred percent possible in every dimension. So, I think people should keep that in the back of the mind. Okay, so last question, before we get to Gatling gun—and these will be quick ones, hopefully—is just like, I think it’s really important for companies to think about their strategic impact to AI market leaders, and whether they’re important to the future of their industry, you know, if they don’t have AI or they do have AI. And I’m just going to give one real example, right?
So, when you think about, you know—I’ll give a couple examples, actually—like, New York Times has a library of material, right, and you have voice catalogs that have a library of material. I think they’re actually valuable and incredibly useful in the AI world, and are important to the future because you have to learn off of something. So, if you hold a bunch of data assets, if you’re not an AI company, and don’t know how to leverage that, you are still going to be very valuable downstream and you should think about how you can leverage that. We have a portfolio company called Redox, and they are an API for healthcare.
And API is a huge market opportunity for AI. Like, who should get clinical trial, you know, for this particular new drug. Or, you know, I want to find a bunch of patients that are, you know, have this malady so that I can talk to them about this particular clinical trial. Or, I want to create a unified data record, and you know, using that unified data record, use AI to figure out how to make Will Porteous better, but I need all of the data across multiple devices.
Will: That’s going to take a lot of money and generations [laugh].
Raju: It’s going to take a lot of money and gen—well, no. You’re too healthy already, you know? You go the other way if you want [laugh]. But my point is, like, interoperability is going to be an incredibly valuable asset in order to get the data so AI can munge it. And so, you have companies that are in pole position, like Redox in the healthcare space, to say, “Hey, you know you want to access to the data? We’re going to make it easy for you to get access to data, but you got to go through our conduit because we have tentacles into all the medical device and medical application companies, and tentacles into all of the EHRs and clinicians around the country.”
So, I think it’s going to be an interesting opportunity for companies that, you know, can be sort of conduits or service providers into AI, for either the data or for just, you know, connectivity, or whatever it may be, I think their future is super bright. But anyway. Anything else you want to mention on this topic, will? You know, I know we covered a lot of ground on how people should be thinking about their own products, how they should be thinking about their internal operations, how they should be thinking about their strategic impact in the sector. Is there anything else you wanted to touch on?
Will: Just I think it’s a really important moment for CEOs to be paranoid about the things that they don’t know, not just about their competitors, but about the way people are actually building companies right now. It’s important to get out of the office, get in front of—talk to other companies of similar size or similar category. Talk to other companies. If you’re an RRE company, talk to other portfolio companies. We’ll facilitate that. And learn from what other people are doing because the velocity of change that AI is bringing to the company-building and scaling process is enormous.
Raju: Yeah. That’s a beautiful thought. Okay, so I’m going to ask only one Gatling gun question—
Will: [laugh].
Raju: —and, you know, answer however you feel. So, are you afraid of being replaced by AI either professionally or personally?
Will: [laugh]. I think I am—am I worried about it? I think I’m really excited for AI to give me more leverage. And my access to the large language models has already given me leverage. There are things that I—I get answers to things that I want to know and the friction is so low to get those answers that I think it’s changing the way we acquire and consume knowledge in really radical ways. That’s a topic for another podcast, but you know, if it’s important in our business to understand data and markets and trends at all times, you know, the LLMs are giving us a gateway to that information at all times, which is making us better at our job, I think.
Raju: Yeah, I a hundred percent agree with you. I am not afraid about being, you know, replaced professionally. And I’d be remiss not to mention our portfolio company, Originalis, that helps us. Like. You said it right on the nose. It was perfect.
I think AI will make VC industry much more powerful. It will give us leverage. It will make us 10x. It’s like vibe coding to a software developer. You still need the software developer, but a software developer juiced on vibe coding is 10x more capable. And I think the same thing is true about the VC industry. If you’re juiced on Originalis, you can be a 10x more productive VC.
And you know, we have an inside track on this, and it probably is another podcast that we should do, just on Originalis because I think it’s quite powerful. I think personally for me, I… if my wife hasn’t left me yet, she’s never going to leave me, you know [laugh]? I mean, I try to do the right thing, but obviously I overcommit, underdeliver on many aspects, and she’s awesome. She’s, like, the most awesome human being on the planet. But, like, I’m not worried about being replaced by AI, mostly because, you know, I look, I’d love for AI pieces of me to do the things that I’m not doing properly for her at home, in terms of, I don’t know, I’m thinking—
Will: [laugh]. You want—you need the AI handyman? Like—
Raju: Yeah, I need the AI handyman, I need the AI, you know, just follow up, guy or gal. I will tell you, AI will never replace my massages. I’m very good at that.
Will: I was going to say, you’re legendary for your hand massages, and I think, you know, that’s not going away anytime soon.
Raju: Yeah, I think that’s why, like, she won’t leave me. I’m just, you know—I think that’s actually the answer [laugh]. So anyway, this was awesome. Thank you. I’m going to let you wrap up with our audience, Will.
Will: Well, RRE POV listeners, we’re so grateful to have your attention. These are truly the kind of conversations we’re having inside of RRE every day as we grapple with the way AI is changing our business, changing things for our companies, and changing our world. We know that you are too, and we want to encourage you to reach out to us, share your thoughts about this episode or any other. I’m just will@rre.com and Raju is raju@rre.com. Thank you for being RRE POV listeners, and we look forward to being back with you soon. Take care.
Thank you for listening to RRE POV. You can keep up with the latest on the podcast at @RRE on X or rre.com, and on Apple Podcasts, Spotify, Google Podcasts, or wherever fine podcasts are distributed. We’ll see you next time.