Why Vertical AI Agents Could Be 10X Bigger Than SaaS
42:12 Share

Every three months, things have just kept getting progressively Better. And now at this point, where we're talking about full on vertical A I agents that are onna replace entire teams and functions and enterprises is that progression is still mind blowing to me a lot.

Foundation models are kind of coming head to head. There used to be only one player in town with open the eye in seeing in the last batch. This has been changing.

thanks. As like competition is the soil for a very photo marketplace ecosystem, a for which consumers will have choice and founders have a shot. And that's the world I want to live in.

Welcome to another episode of the light cone on gary. This is geared harsh and Diana. And collectively, we ve funded hundreds of billions of dollars worth of startups, right when they were just one or two people starting out. And today, gerri is a man on fire and he's gona talk about vertical A I. Yes.

I am. I am fired up about this because I think people, especially started founders, especially Young ones, are not fully appreciating just how big vertical A I agents are going to be. It's not a new ideas.

Some people are timeout vertical AI agents. We funded a bunch of them. But I think the world has not caught on to just having it's going to get.

And so I am going to make the case for why I think they are going to be three hundred billion dollar plus company started just in this one category. nice. I'm going to do IT by analogy with sas. And I think in a in a similar fashion, people don't understand just how big sas is because most started founders, especially Young ones, tend to see they started industry through the lens of the products that they use as a consumer. And as a consumer, you don't tend to use that many says who is a mostly able companies.

And so I think a lot of people have missed the basic point that if you just look at what s icon valley has been funding for the most like for the last two years, like we've mostly been producing service companies, guys like that's hitter ly been like. Most of what has been coming out of silent valley is over forty percent of all venture capital dollars in that time period. Went to sass companies, and we produced over three hundred SaaS unicorns in that twenty year time period, which is way more than every other category.

Software is pretty awesome.

Software is pretty awesome. I was thinking back to the history of this because we we always like to talk about the how the history of technology informs the future. And um the real catalyst for the ash boom was, do you guys remember XML HTTP request? My god, like i'd argue that that was quite literally the catalyst for this ash boom .

like ajax ajax .

yeah in two thousand and four browsers added this job scope function h idp request, which is the missing piece that enable, due to build a rich internet application, inner web brows or so, for the first time, you could make things in websites that looked like dust up applications and then decorated gool maps and gmail instead of this whole like assis. On essentially the the key technology at lock was that software moved from being a thing that you ve got on a city ROM and installed on your desktop, to being something that you use through a website and on your phone.

Paul graham actually a shares in that lineage age, and that he was one of the first people to realize that he could take the HTTP request and then actually hook IT up to a unix prompt. And you didn't actually have to. You have a separate your computer program that would change your website. So via web was a online store kind of like shop a fy. But way back in the day.

yeah he was basically like the first sass APP ever. Like like pg actually invented sass in like one thousand ninety five. It's just that those first sp s kind of socks because they didn't have example H D P request.

And every time you would like click a button, you would have to reload the whole page. And this is just a studii experience so I didn't really catch on until two thousand five. One example request why it's right.

Anyway, I see this L M thing is like actually very similar. It's like it's a new computing paradigm that makes IT possible to just like do something fundamentally different. And in two thousand and five one, cloud and mobile finally took off.

There is a sort like big open question of like, okay, well, this new technology, this is what should you do with IT? Where is the value going to a crew? Where are the good opportunities for startups? I was going through the list of like all the billion outcomes who were created.

And I kind of a had this realization that you could kind of bucket the the different path ths the people talk into like three buckets. There's there's a first bucket that people started with, which was like I would call them obviously good ideas that could be mass consumer products. So that's like docks, photos, email, calendar, chat, all these things that like we used to do on our desktop, but that obviously could be moved to the browser, into mobile.

And the interesting thing is zero startups, one in those categories, one hundred percent of the value flow to in companies right, like google, facebook, amazon, they own all, all those businesses. Folks forget that, like google, dox wasn't the only company to try to bring microsoft office online. There are like thirty companies to try to bring microsoft office online, but they all lost a gool one.

Then there was like a second category, which was like mass consumer ideas that were not obvious that nobody predicted. Um that's like uber, inter card, door dash, coin base, airbnb, those ones those ones came out of left field like the the dot dot dot between XML H D P request and airbnb is like very not obvious yeah and said the income didn't even try competing in the spaces until he was like too late. And so starters are able to win there.

And then there's a third category, which is all the beauty backed companies, and that's like three hundred of them. And so like look like by number of logos, way more billion dollars annually created in that third category than the first two. I think one reason why that happened is like there is no like microsoft of seas, like there is no company that somehow does like aster, like every vertical and every product. Like for structural reasons, IT seems to be the case like all different companies, and that's why there's so many of them.

They think salesforce is probably like the first true test company. And I I remember mark bending of coming speak out, Y C, and he tells the story is just very early on. People just didn't believe you could build sophisticated enterprise applications like over the cloud or via, says he was just so um there is like a perception. He was like like you don't you buy like your box software and that's .

like the real software .

that he was quite concerning as the early web APP sucked. They like via web, where you had to be a visionary like pg, and understand that the browser was going to keep getting Better and that .

eventually would be good, which feels like quite remediation of today. Yeah yeah, the same thing. Like, oh no, like you won't be able to build like sophisticated enterprise applications that use these L M A I tools because they pollutant ate or they're perfect or they they can like just toys. Yeah, I like that elisa story exactly the same.

And so when I think about the parallels with L M S, I could easily imagine the same thing happening, which is there is a bunch of categories that are like mass consumer applications that are obviously huge opportunities, but probably the income ants will win all of this. So that something like a, like a general purpose, A I voice assistant that, you know, you can ask you to do anything like to do that thing that doesn't, obviously the trade exists, but like all the big players are going to be competing .

to be that thing right? Apples a little slow on that one I series so stupid still what said that makes no sense.

I've in eight hundred account that is like the very obvious this thing is search and maybe google will still win um on search, but perplexity is definitely them.

The classic innovate. I mean, you argue back to what said uber 和 airbnb。 These are actually really risky things from a regulatory standpoint. So if you're google and you have basically guaranteed your giant a part of goal that you know sort of comes to you every single month, like why would you endanger that part of goal to sort of pursue these things that might be scary? Or my room in the party gold.

I think I think that's like probably the primary reason why the incomes didn't end up building those products and didn't even clone them even after they got big. And IT was obvious that that they were going to work gool never launched in uber on. They never launched airbnb clone.

I was listening to this a talk by travis, and one of the things that he said that really stuck with me is that in the in the first shoes of uber, he was very scared of that. He was going to put good at prison like a long time, like he was actually personally risking going to prison in order to build a company. And so yeah, no highly paid google executive to do that.

What do you think about why the comments didn't go into B2B as is a p ar t of the rea son is tha t a l ot of the you th cas es are ver y is a v er y wid e dis tribution.

I is a great question. I love to hear. What you guys think my take is that it's just too hard to do that many things as a company like each B2B ask rea lly req uires lik e the peo ple who are run ning the pro duct in the bus iness to be ext remely dee p in one dom ain and hea r ver y dee ply abo ut a l ot of rea lly obs cure iss ues.

You like take like gusta for example, like why didn't google build a gusta competitor? Well, there's no one to google who really understands paperl and has the patients still like deal with all the nuances of all these like stupidly role regulations, and it's just like like it's just not worth IT for them. It's easier for them to just focus on like a few really huge category .

in the B2B bus iness wor ld. It's it's sure about the unbundling bundling of software argument that comes up a lot as well. I think and why and why did all these vertical B2B Saa S pro ducts of all ver sus jus t lik e ora cle or sap or twi ce yea h to be jus t own ing lik e eve rything and I t hi nk IT mig ht be als o is a n ot hing's tri bute or to the shi ft to lik e say ing the int ernet is in the old way s of set ting sof tware.

Again, like you have this box software that was really like expensive to install and you have like a whole ea around that. And any time you wanted something custom, like the integrators would just say, oh no, like we can like to build you a custom like perl feature or something like that and then sales walls comes along with like a sa solution. And IT just seems like IT could never be as powerful or sophisticated as like the expensive enterprise installation you just paid for. But they prove that I told was the case. And I think that just like open the gates for all of these like vertical sah solutions to emerge doing exact where .

you're saying is the problem is that with other enterprise software, if you are user of oracle in the net. Because they have to cover so much ground, the user experiences actually pretty bad. They are trying to be jacked able trades with matter none. So being a bit of kitchen sink type of experience. And this is where if you go and build up to be sas vertical company, you could do literally a ten Better experience and more delightful because this is stark, their friends between consumer products and enterprise user expense.

Well, there's only a what three Price points in software. It's a five dollars per seed. Five hundred dollars proceed or five thousand dollars proceed. And that maps directly to consumer S M, B or enterprise sales.

And then I think time in memorial has taught us that in the past and this is lesson less true, uh, with new software, thankfully. But enterprise is terrible software because it's not the user buying IT. You know some muck muck inside a fortune one thousand is the person whose getting wine and dine for the omega seven figure contract.

And you know they're going to choose something that maybe isn't that good actually for the end user, the person who has actually use the software day to day. And um i'm sort of curious to see how this changes with alams actually. I mean, two date. One of the more sAiling things that we've seen for both S M B and enterprise software companies is that our all software companies, all startups, period is like there is a sense that as revenue scales, the number of people you have to hire scales with IT. And so when you look at unicorns, even in today's see portfolio, uh, it's quite routine to see a company that reached a hundred or two hundred million dollars here in revenue.

But they have like five hundred, a thousand and two thousand employees already and i'm just gna be very curious like um even the advice that i'm starting to give companies that are a month two out of the badge, uh, it's feeling a little bit different than the kind of advice I would give last year or two years ago. In the past, you might say, you know, let me find the absolute smart st person in all of these other parts of the org like customer success or sales or different things like that. And I want to find someone who i've worked with, who is I know is great, and then i'm going to go sit on on their door doorstep until they quit their jobs and can work for me.

And I want them to be someone who can, you know, build a team for me, hire a lot of people. That might still be true, but i'm starting a sense that a the matter shifting a little bit like you actually might want to a hire more really good software engineers who understand large language models, uh, who can actually automate the specific things that you need that are the bottle next to your growth. And so IT might result in, you know, a very subtlety significant change in the way startups grow their businesses sort of post product market fit IT means that i'm gonna build A M systems that bring down my cost that caused me not to have to hire a thousand people. I think we're right at the beginning of that revolution right now.

I know we talked about this in a previous episode. We talked about there will be a future unicorn company that only run if we do limit with only ten employees. That's completely plausible y're.

right? The e the proms.

And I think .

everything is like a trend that was already underway pre alams. Like I remember when I was running triple bite, for example, we need to like build marketing constant, like user acquisition basically um and especially although isse a series b the like traditional way you was supposed to do that is to like higher marketing executive and build out like a marketing team and just like basic spin up this machine to do like cells and marketing.

But i'd actually met like A Y C found the mike he was. His company was basically building like a smart frying pan town's a bizarre a, he was A M I T engineer. Yeah, you member this, his MIT engineer.

And to sell the smart frying pan, he had to get really, really good understanding, like paid advertising and google ads and just the whole one to stuff. So he, he taken this engineers mindset, approached to IT. And I remember just talking to him about yet, and realizing this would be so much Better to have an M.

I. T. Engineer working on like our marketing efforts than any of the marketing candidate I ve spoken to. And he was able to like scale up to like and he was being like at one point, like a million dollars months and just marketing and various like .

camby and tribal by the great marketing. Like I remember like the culture in station take over that you did all like out of homestake that you did. He was like, really high quality stuff at stuck with that. You could tell that he was not being done some like people.

And I was all mike, and like the comment I would often get when people asked me around that time, like how bigger trouble by IT. And we were .

like fifty people.

And I so so this is like a hundreds of people, as I know. It's all because if you put a really smart engineer on some of these like task, they just find ways to make. They find leverage. And I like elements can go even way beyond like the leverage, how I would just peel software.

okay. So here's my pitch for three hundred vertical AI agent unicorns. Literally every company that is assess unicorn, you could imagine there's a vertical AI unique equivalent in like some new e universe is like moves of these sas unicorns.

Before him, there were some like box software company that was making the same thing that got disrupted by a company. And you could easily imagine the same thing happening again. We're now basically every, every sas company built some software that some group of people use. The vertical AI equivalent is just going to be the software plus the people in one product.

One thing might be just enterprises in general right now a little unsure about what exactly they like, what agents they need.

And one approach i've seen from exactly more experience founders like a bret Taylor city of facebook, such as company era, I don't know all the details, but as far as I can tell, this essentially more like broadly about letting enterprises like deploy these AI agents and spring them up like custom for the enterprise verses like, oh, y, we have like this specific agent to do. This is something i've seen for all my companies, is called a Victor shift. That was fun about a year ago.

There too really smart, like harvard computer scientists. And is that what they found is that they wanna build a platform to make IT easy for enterprises to build their own, like use like no code or sd case to build their own lake internal L M. Power agents, but they advise often donor exactly what they want to use these things for.

So bring you back. I wonder if like in like the box software world, you started off with just like a few vendors who just basically we're trying to convince people to use software at all. And I was just like, IT does everything and then he gets more sophisticated, higher resolution, and you get lots like vertical sax players.

We go through that same view with LLM s, where the early winners might just be this like general purpose, say, like we'd like, make IT easy for you you to do LLM stuff and then know the vertical agents will come in over time? Or do you think there's reason is different now in the vertical agents will take off on day one? Yeah.

that's interesting. Because if you think about the history of SaaS, the consumer things worked first, like two thousand and five to two thousand and ten was mostly consumer applications like email and chat and maps. And people got people as individuals got used to using these tools themselves. And I think I made IT easier to sell service tools to companies because of the same people are both employees and consumers.

And I think the also might just be like this is this is all just continuation of software. And this there's no reason that has to reset back like election, have to reset back to a few general purpose, like enterprise ellam platforms doing everything, because enterprises have already been trained on like the value of point solutions and vertical solutions, and like the user experience that can be that different, these things will just be a lot more powerful. And so if enterprises already built the muscle of believing that like startups of article solutions can be Better than my legacy broad platforms, they are probably going to be willing to take a bet on a star art promising a very good vertical A I agent solution today. And I feel like we're all seeing that in the batch now with some of our companies are getting fast attraction in enterprises for these vertical I agents then they we've seen before.

I think we're just early in the game, right, like all software sort of starts quite vertical. And then as the industries actually get much more developed, um then I mean, I just answer my earlier question. It's like where why does the company end up having a employees? It's actually that um you know early early in the game, everyone's making these specific point solutions. And then at some point you've got to go horizontal like you're already doing this crazy spent on sales and marketing. And then the only way you can actually continue to grow once you sort of get one hundred percent or you know some large majority of the market is you actually have to do like not just a point solution, but things that sort of work together.

Me, the other point of why the bullet case for vertical A I agents could be even bigger than SaaS is that says you still needed an Operation team or settle people to Operate the software in order to get all the workflows to be done. I don't know. Approval workflows are you have to input the data. The argument here is that you will get not only replacing all a set of soft, so that will be the one to one mapping, but is also going to eat all of a lot of the payroll. You look a lot of the spent for companies, big chunk is still .

a payroll and software tiny .

exactly and more on employees to do random data entry or approvals or click the software. I agree.

I think it's very possible. The vertical equivalent will be ten times as large as the sas company that they are disrupting.

I mean, the two case, I could be that the vertical point solution could be just big enough and you don't need to do that breathe thing, right? That could be nice.

Should we give some examples? I feel like we've working with so many vertical AI agent companies. We've got like news from the front. I was actually going well.

your former had a product, iron cannon is working on A Y, C. Company called outset that I worked with. and. Basically they're taking l EMS uh to the surveys and qual trix space. So qual trix is almost certainly not really going to build the best of breed a large language model with reasoning.

And then the funny thing about surveys is you know who is actually for is for people who run products for marketing teams, is for people who are trying to makes sense of like wear our customers actually want. And what are survey is like guess what? That's a language. So um and then I feel like these types of businesses um actually have to thread this needle um because enterprise and S M B software ten is sold based on a particular person who is the key decision maker. And um you have to go high enough in the organization so that the people you're selling to are not afraid that their whole their job and or their whole teams job is going to go away.

tally. That's kind of the move that I seen that allow companies are cell need to do because if you going to go and sell to the team that's gonna replace .

a and IT just does not work.

So I think this is an interesting where there a lot of these are top down and you have to go through a company. You can get the C E, O to sign off .

on the company working with mechanic. That's essentially A I agent, but for least where the start is like a testing there are getting really great traction right now. And it's interesting because you remember, I would decade ago, why economies that we work with, the rainforest Q A, like rain forest, was A Q A as a service company and that they had this exact tension of where they couldn't actually replace your Q A team.

So they needed to build software that made the Q A T more like efficient. But really that obvious ly meant trying to replaces many of them as possible, but they can replace the whole team. And so there will always on this sort, like tight rope, between trying to sell the software to, like the head of engineering as that.

This will mean you need less Q A people. And great, but the new year to have to go sell out to the Q A team who don't want to be replaced. And so I think that was always like a fiction for that business for how I could like scale and grow.

But now like memetic with A, I can actually just replace the Q A people so their pitch is not oh this like major Q A people faster is like this just means you don't need A Q A team at all. So I can just focus to sell to like engineering and energy doesn't need buying from QA at this point. And you can also go in, I mean, to start with, you can go in, sell to companies.

I don't even have B Q, A teams at the moment. They just use something like romantic. And then I would just like .

keep scaling with the killing and we'll just never build accurate you ever. That is a real life case idea of what I end of saying about why these vertical AI agent comings can be ten times as big as the house companies.

I'm seeing this interesting now um like in recruiting too. I had this exact same issue, the triple vision where to build the often um to build software that makes IT easy to like screen and higher software engineers you need buying from most the engineering team that they're joining, but also the recruiting team. And effectively, the software we were building was trying to replace the recruiters, but we can completely replace the recruits. But now within Y C.

so the recruiters were always like opposing yeah opposing IT because I was the threat .

to yes so just always like friction unlike unlike halfa, you can get when the custom know you trying to sell to is worried about being replaced um but yeah I think now it's so early days but now with A I you can build things that do the whole sack like of recruiting.

We have a coming work with last patch o open them a priority which is actually just doing like the full like technical screen, the full initial recruit screen and getting great traction. So I think as those things keep going, like they won't have the same thing, you won't have the friction. Although I need to convince recruiters to use this, you probably just like not build a recruiting team in the same way that you .

used to any other. What example is, even for death tool companies, they have to do a lot of a developing support. And I work with this company called cabillo t eye that basically built one of the best chatbot that responds to a lot of the, a lot of the technical details that are hard to answer.

And I think a lot of the companies that started using them, they actually and that are having deep real teams that are a lot smaller because in just a lot of the developer documentations, even the youtube videos that left tools put up and even a lot of the chat history. So you just keeps getting Better and Better. And is I give really good answers actually one one of the best .

I seen ah I also worked with a customer support that like a AI customer support asian company called partial actually we both did last batch and I learned to a couple interesting things from part help. Um the first is customer like A I agent for customer support was like the category is like famously crowded where there's like supposedly like you know one hundred of them.

If you go and you google like A I customer support agent, you'll get like a hundred results on google. But what I learned through working with parole is like actually kind of bullshit. Like like most all of those companies are doing very simple, like zero shot elm prompting that can actually replace a real customer supporting that is a lot of really complicated work.

Those is just kind of make feel like a nice demo. But to actually replace a customer support team for like an at scale company that has like a hundred customer support reps to do lots of complicated things everyday, you like really complicated. It's offer that does all the stuff that like jack heller was talking about and is there were only like three or four companies they were even attempting to do that and commuting, they cumulatively, they have like a lesson one percent market penetration. And so the market, which is completely open.

I could also see that being another case of um hyper specialization or hyper vertical aliza. You like there's not going na be I mean, maybe eventually there could be a single general purpose customer support, a software company, but we're like that. That would be like a eight, nine inning kind of thing and we're literally in the first inning.

So you know instead of you going to have companies like gigi M L, that you know it's doing IT for eto, doing thirty thousand tickets every single day and replacing a team of a thousand people. And but it's very specific and IT has you know it's not a general purpose democratic of thing like it's ten thousand test cases in a very detailed uh e outset that you know is basically just for zetto and things like epo. But if you are you know any of the other marketplace companies, you probably going to use IT because like that's a very well defined kind of marketplace that see you know instant delivery marketplace.

I think this is the kind of dynamic that LED there to be like three hundred billion dollars as companies rather than like one like ten trillion dollars like meta saying that provides all this often for the world is just like the customer is is require really heavily like tailor solutions, is hard to build one that like works arve.

exactly. I mean, we already give three examples of customer support. There are very different vertical ses like death wall companies, very different kind of that you need in the training set to marketplaces. Very different, right?

Yeah I guess whether you have agents or real human beings working for you, you end up with the same problem, which is every company bumps up against coasts. Theory of the firm, which says that any given firm will grow only so much to the point where IT a becomes inefficient to be larger than that.

And then that's why the networks and ecosystems and you know a full blown economy, you like every firm, will sort of specialized to do what IT is particularly good at. And then the limits the other limits of what those firms can be. It's actually based on um your ability as a manager.

So yeah, that part a little bit breaks my brain, because when we spend time with Parker conrad at rippling a, one of his favorite points is actually, well, be everyone's very obsessed with with that the fact that the rocks can talk, and, you know, maybe they can draw. But the more interesting thing for him running H R I T software that you, he spent a lot of time thinking about H R, like actually the cool est thing about the alums is that the rocks can read from his perspective. Like you, he's, I think, three thousand employees.

He still runs payroll for all three thousand employees through rippling. So I think he spent a lot of time thinking about like how can one person extend their ability as a manager? And I think we're going na see a lot more there. That would be in a reverse argument that at this moment, where tools form managers and ceos are going to get much more powerful .

IT could increase the scale. The firm that you can run, and that's certainly rippling, is trying to do like build like sweet of hr tools where if he wins, he's going to eat a whole bunch of billion dollar companies. And like one.

one giant company is very interesting point. gary. I think what make me think about this is with having all these A S tools gonna give the ability to all these leaders and all these works to basically open the capture of the context window of how much information they can pass, because is a limit of how much as humans we can have. Meaningful relationship is like the whole thing, with a dumarge number for three hundred people, that hundred you can have a meaningful relationship with, but with the eye, because all of these rocks now can read, I think we will be able to extend that zomba limit that we have yeah.

I think a flow. Rivaldo had this interesting post on twitter that went viral around, I think someone had made a voice chat like this weekend project as a CEO but I would call uh all fifteen hundred of their employees and and you know he was in a very short call like kind of sounded like IT was from the CEO just asking kind of personally I mean this sort of reminds me of um that seen in her where IT zoos out and actually you're following the experience of one person using the her O S but actually that her OS is actually speaking to fifty thousand thousands or tens of thousands of people all at all time.

Eight thousand, three hundred sixteen. yeah. I mean, large language models can talk and can have conversations. And then to what extent can, uh, you know, this power actually extend the capability of one or a few people to understand what's going on?

I I had about that, I got got me thinking because as I understand the project, something I get just IT will call up your employees and then employees and just like rainbow about what they've been doing and they will just extract the meaning out of IT and give the CEO like like bullion summary of he is the most important stuff.

And there were a bunch of like as companies that attempt to do this, sort of like weekly s pulses from employees using like traditional sofa. But like that version is is literally a hundred times Better than the prel m version of this idea.

But I wonder with like that particular tool, just like it's not it's going beyond just like reading and summarizing this. This is the argument. Like if writing is thinking, then like there's actually just a huge amount of work that's involved in the effort of figuring out like who's an effective communicator and they got are the most important things to be like, what what are the key things to be focused on as the company. And just wonder if that at some point to the elms, do like they go beyond just like summarizing and reading and doing actual thinking at which point like who's actually running the organza yeah interesting to I .

guess to the other thing that kind of interesting about how Parker congress thinking about IT is um I found out about this recently of an interview with map maginness C O O that are there more than one hundred founders who work at rippling now as sort of specific people who run like an entire set vertical insight .

ripplings it's super cool the way he's built the team heart right knows hot .

about IT as you've done injuries yeah I mean, it's definitely very focused on a recruiting founders and I mean parklike rippling is that essentially the the case against vertical little versy vely trying .

to a horizontal zed and take over all of H R N I T software platform that .

has like lots of value and he wants to recruit founders and teams that built on top of the platform like one is a little bit more to like amazon s where like shared infrastructure yeah I think .

every product that they've released, I mean things like time tracking and what not. I mean, basically they launch a thing and IT hits like multimillions of dollars and err on day one of launching. And that's exactly what we're talking about earlier.

Like once you once you have a vertical, once you have a total ld, what you're saying is, well, I have to spend this money on sales marketing. Anyway, I um basically get higher L T V and hold my cac constant. And that's sort of what if you look at all the top uh, software companies today, it's like that's what oracle is. That's what microsoft is, that what sales forces rippling, not on wood gonna the next. But um it's it's an interesting alternative to going from zero to one totally on your own.

You want to talk about some of the voice companies that we have. I think that's like an interesting like sub category of this of this stuff like really blowing .

up now company that I work with, cosy lit, that basically does A I voice calling to automate a lot. That collection in the order lending space, which traditions call people.

is like he, you were with thousand dollars on your car.

yeah which .

actually actually this kind .

of job is one of those bottle passing job, kind of socks, because a lot of these low weight workers working all these call centers and is like a terrible, boring job. So very high turn and giant had count to run because so many accounts with these banks. I, and this is a perfect task that A I could automate.

And what sAiling has done is has has been able to actually get very, very accurate. And IT has been going live with a lot of big banks. We are super exciting. And this was a company from last year and demonstrating that, that part of IT that they were able to get in because they sold through top down.

I guess the space feels like it's moving very quickly and that we have incredible companies that are a voice infer companies like and then people can sort of get started right away. And retail also. I mean, these companies have reach pretty fast scale just because it's one of the more exciting like mind blowing things you can get up and running within I mean, literally the course of hours.

Um and then some of the question that you know remains unanswered and we hope they figure IT out, is how do you hold onto them, especially as you are run into things like the new OpenAI voice aps yeah do you go direct like you? It's probably way more work to try to use the underlying aps off the bat, but these are platforms are clearly low bar. And then the question is, can you keep raising the ceiling so that you can hold on to customers forever?

How do you were making an interesting point earlier about like how the apps of people who built on top of alem has changed from, like, early twenty twenty three when I started?

Now, yes, we are talking about is a great example of this. I think even if you went six months back, IT felt lag. The voices were not really stick enough here. The latency was too high, like there was felt like we were probably ways of having A I voice apps that could meaningful ly like replace like humans calling people up and they here we are and yeah I just zoom out thinking back to the first Y C batch where Ellen powered apps first came in was probably winter two thousand twenty three years in almost two years ago now. And the apps were essentially just things that spat .

out some text and .

not even like perfect and copy edmark. I like incremental ah like .

I had a company, I mean the one that six in my head that is a company speedy brand. And what did is make IT very easy for like a small business to just generate a blog and spit out content marketing like a very obvious ous idea. IT wasn't perfect, but I was pretty cool at the time and that we talk about a bunch of the show, but I was like the ChatGPT rapper turned out around the time, hey, like this is what an elo m map looks like. It's just a ChatGPT rapper. IT is a very basic spits out some text like it's going across my opinion and .

exit I I don't know I want did but but the the first that first wave of allam aps mostly did get crushed by the next wave of GPT.

I like we've had this sort of building of the frog effect where from out it's like every three months, things have just kept getting progressively Better. And now at this point, we were talking about make full on vertical AI agents that are going to replace entire teams and functions and enterprises. And just that progression is still mind blowing to me. They were two years in which still relatively early, and the rate of process is unlike anything we've team before. Anything was .

interesting to see is we discussed in the last episode is a lot of the foundation models are kind of coming, had to hit. There are used to be only one player in town with open a eye, but we've been seeing in the last batch, this has been changing. Clad is a huge contender .

as like competition is the soil for a very feral marketplace ecosystem for which consumers will have choice and founders have a shot. And that's the world I want to live in. So people are watching and thinking about starting a startup or maybe have already started, and they're hearing all of this. How do you know what the right vertical is for?

You find some pouring repetitive admin work somewhere instead. Seems serious, like the common thread across all of this stuff is if you can find a boring repensitive admin task um there is likely gonna be a billion dollar A I agent start up if you keep digg deep enough into IT.

But that sounds like you should go after something that you directly have some sort of experience or relationship to.

There is a common that if there there is definitely common thread that i've seen in the companies that are that i'm seeing promise with. And I know I just pops into my head sweet spot. I think I mentioned on this before, like they basic building in A I agent to bid on government contracts. And the way they found that idea as a year ago was they just had a friend whose four time job to sit there on, like a government website, like refreshing the page, like looking for new proposals to bid on, and they they were pivoting. They're like, I like that seems like something could do a company from a recent batch with pivoted into a new idea that's getting great traction, like the waste building in A I agent to do persse, like medical billing for dental clinics. And the way they found the idea was um one of the found his mothers is a dentist and so you just decide to go to work with out for a day and just sit there seeing what he did until like oh like all of that like processing claims seems like really boring, like an L M should totally be able to do that and he decided writing software for break. His mother is dental clinic.

So I guess I mean in robotics, the classic maxim is uh the robots that they are going to be profitable and they're are going to work are going to be um dirty and dangerous jobs. And in this case, for vertical house. Look for boring button passing. Well, with that were out of time for today, will catch you on the light phone next time.