Why Building With AI Got Easy and Maintaining It Got Brutal with Fathom CEO Richard White - The Vault Unlocked

Loading summary

Transcript51 lines

[00:00]
A
Most people are reacting to AI. Our guest for this episode built for it three years before it arrived. Two bets. Transcription costs would fall to zero and AI would get good enough to actually do something with what it heard. Both were contrarian then. Both were right. Fathom is now the top rated AI note taker on G2 and Richard is one of the few people who can tell you what actually changed and, and what didn't. We get into why building software has never been easier and maintaining it has never been harder. Why your years in business are an advantage in this shift, not a liability. And why the real bottleneck right now is not the technology, it's what you can see. If you have been waiting for the right moment to move.
[00:47]
B
This is it.
[00:48]
A
Richard White's background is engineering and product design. This is the vault unlocked. Let's unlock it.
[01:04]
B
Richard, welcome to the show. I'm excited to have you. I just for the, for the guests and the viewers listening, why don't you tell us a little bit of who you are. I'm excited because I use your product. I love your product. It's in our business today. I seen you've changed it quite a bit and I'm excited to have you here. But for the listeners that may not know, tell them who Richard White is.
[01:30]
C
I'd like to think I'm a product designer and kind of technologist. You know, no one's let me write code of production in gosh, maybe 10, 15 years. So I'm not sure I can claim being a technologist as much anymore. But that was my background originally kind of in engineering design. Done a couple of startups. I worked the first batch of Y Combinator if I want to date myself. Their product before this called Uservoice. But as you kind of alluded to for the last five, almost six years we've spent working on Fathom, which is the number one kind of rated on G2, AI notetaker for people on lots of back to back meetings. It's been a really fun ride with a really great team and the most fun part about it is talking to folks like yourself who love and use the product every day.
[02:10]
B
Yeah, I mean I've used a lot of different AI notetakers and I've used Fathom before and then we switch and now I'm back to Fathom and I, I'm, you know, I'm actually, I'm sold. I, it's like it, it's, to me it's, I find it's the most easiest interface. It just usability of It. And I. And I love that. I just feel like it's not overbuilt. It's just built, like, just exactly for what it is. Take us back to where did this start? Like, where did you see that this was needed? Because the one thing I do know about Fathom was way before this huge, the AI craze and everything. So you. You saw something way before. That's what I'm interested in. Like, the vision and the strategy you saw and how you brought it together.
[02:51]
C
Sure, yeah. I mean, it was actually even. Right. Just before COVID honestly was working on a different product, was working on a totally different product and totally different space, and just found myself on a ton of zoom meetings. Like, I think it was like 15 to 20 a day. A lot of them were research sessions, right? Where I've got 20 minutes almost back to back to like, interview someone, demo something, get their feedback, rinse and repeat. And it's kind of one of those things where, like, you know, if you run into a problem once a day, you don't maybe do anything about it. You run into it 20 times a day. You're like, oh, my God, this is really painful. I need to, like, I don't. I need to fix this. Right? And so, you know, I remember just kind of. Kind of thinking how. How kind of crazy it is the way we kind of share knowledge out of, like, meetings and stuff, right? It's like, oh, I meet with someone, I had this great experience. They tell me some really interesting quotes or facts or whatnot. And then I heard we scribbled down notes and then try to, like, clean them after the meeting and remember exactly what they said. It's a very stressful situation, right? It's like being a court stenographer, right? And also being the. The. The. The lawyer kind of interviewing the person on the stand at the same time, right? It's like you're kind of doing both and no one likes it, right? No one likes taking notes. No one likes reading notes. Notes are also like, we're like a really poor artifact. I share with my team, and a lot got lost. You know, I'd have this amazing conversation, I'd share the notes with my team, and they kind of shrugged their shoulders like, okay, right? So I look at this something. Don't we have the technology to fix this at this point? Right? And, you know, if you go back to 2020, there were tools that were doing call recording. Nothing with AI yet, obviously. You know, most of the products are, like, in the sales space. Like companies like Gog and stuff like that. And they're really expensive and they were candidly kind of mediocre, right? It's like it took you 30 minutes an hour to get the recording afterwards. It was mostly just a transcript. No one wants a transcript. What I wanted was just like, I get off the meeting, there's instantly some notes. Great, like, I don't have to do this job sort of thing. And we kind of looked at that space and we kind of had this think thought, like, gosh, where is this space going? We kind of had two core hypotheses that really got us excited. It got me excited about what turned into Fathom. That was Transcription five years ago. Actually pretty. Still pretty expensive, right? It was like three to four dollars an hour to transcribe content, which doesn't sound like a lot, but if you imagine if you build a product in transcription, build product meetings, people are easily going to do 10, 20, 30 hours a month on it. Gosh, your hard costs for that product are already like $50 a month, right? So the fact that Gong and Fuch, that were charging $150 a month makes sense in that context, right? I was like, why is it so expensive? Oh, right, the input costs are expensive. And so we kind of looked at it like we think this kind of commodity, like when I. We tried a bunch of different vendors, like kind of made a little prototype and tried a bunch of different vendors like Amazon and Google, and I think one was called Rev. I was like, these are all pretty good. They're not great, but pretty good. So this hypothesis, like transcription costs will go to zero because they're all good enough and cost is always trending down. We think it'll go to zero. And more important than that, and we think AI is going to get really good. And it's kind of funny now because it's kind of an obvious thing, but go back five years. Very contrarian take because it's hard to remember. But there was a. There was a wave of quote unquote AI companies like 2015-2020 that were terrible, right? That promised you the world and delivered almost nothing, right? And so. But we're. Because I was like, no one wants a transcript. I don't want to get off a meeting and read a transcript. I don't want to read. Nothing to do with the transcript. But the AI will need a transcript to do all the fun stuff I think it could do in the future. Write your notes, write your actions, fill in your CRM. Find, find trends, find themes, alert me when certain Things happen. All that stuff needs a really good, high quality transcript. So we started the company with those two ideas and said, gosh, if those two things are true, transcription cost goes zero and AI gets really good. Could we be the first people to give away this product for free? Right? In a space where people are short you $150 a month, what if we just gave away for free? Because we actually don't think the value is in the meeting itself, it's in building up this database. Then you build a bunch of AI features on top of. And so we always have the thesis of like, we're gonna give away this product for free to individuals with the hope that that gets us into a bunch of companies where we can then sell a different product to the managers of those people. Right? Because the managers have a different problem, which is I, I'm not in the meeting taking notes, I'm out the outside the meeting and I want to know the important things that are happening. I want to know there's a pricing discussion doesn't go well. There's an argument that happened at the, at the engineering standup. There's, you know, a deadline that slipped three times, like, but I don't have time to sit and listen to every meeting, right? And so I got really excited about this business because one, kind of fit the hypothesis of where I thought the world was going. But two, it had this really awesome kind of two sidedness to it. We had one part where you can give away a lot of value for free and feel okay about that because you don't have to like charge people later because there's just nice kind of complementary business built on top of that for their managers. That's kind of how we got started, right? And it's kind of funny you mentioned that we were kind of ahead of the curve. And I think that's probably true because we had a third corollary to those hypotheses, which was if you wait till when transcription cost is zero and AI is really good, you'll be two to three years too late to start this business. Right? It'll be kind of obvious to everyone and everyone will jump in. But like, like any technological revolution, the best companies like build towards that hypothesis couple years out and they do all the other stuff, right? Like we spent two or three years building all the foundational work and the product experience that you talked about, the good user experience, good usability, the reliability, the obviously distribution channels, all that sort of stuff. And so it was a very much a go to where the puck is going kind of thing, not, like, wait for it to get there.
[08:30]
B
Then when. So when you guys are doing the hypothesis, like, this was again back in 2021, you know, Covid days. I don't like to use that word. I. I even.
[08:40]
C
I hesitated to say it because, like, I feel like we just don't use that word anymore.
[08:43]
B
I know I don't.
[08:44]
A
I don't really use.
[08:45]
B
I hate using it. And it's funny because, like, and sometimes my brain, I keep thinking, like, it's only like a couple years ago, but like, no, it's like, that's like almost six and a half years ago now, like, for, you know, so today there's a lot of players in the marketplace, but you. You've had the market share. So are you seeing competitors, like, coming in and taking over? Are you. Are you guys adapting your product now more with AI? Like, how are you staying in the trends and. And how do you see where AI is going? I mean, even just with the note taking, let alone, what is that next vision that you have for where this can go?
[09:19]
C
Yeah, it's kind of funny. I mean, like, I feel like it's been a tale of two businesses, right? There's a. I. I do this whole talk and I show my revenue graph. You could see the point where AI actually really shows up. And for us, that was kind of like GPT4 level of AI. That was the point where the AI could write better notes than the human. Right? And so that's where we went from being a meeting recording business to being a meeting AI business. Right? And it really takes off. And this whole second act of the business is not about hypotheses and stuff like that. It's actually about, like, how do we get really good at building AI functionality? Because it's actually very fundamentally different than building traditional SaaS or just software. And I can talk about that. But so it's been kind of cool. And now we kind of are seeing this, like, you know, the capabilities increase every six to 12 months. And now sometimes it's even like two to three months. Right? And so we're constantly now seeing, like, okay, two years ago, state of the art was we can write a really good summary for a meeting and we can figure out the action items. A year ago, it was, oh, we can look across not just one meeting, but every meeting you've had with Acme or, you know, every meeting you've had in this, with this prospect. And we can surface, like, risks, we can surface trends. And now the state of the arts Move into. Oh no, now we can actually look across every meeting you've had over four or five years right across your org and tell you trends about competitor trends, internal knowledge management type questions like, you know, we asked it the other day, like, hey, we don't like to do a lot of documentation here at Fathom because we just assume that's the point. You just ask the AI, like why generate documentation and maintain it? Just if you need to answer a question, you just ask the AI. Now we're at this point where that actually works and we can be like, hey, we got a new engineer and they're wondering about why we built the system the way we did. Can we, can you give me a history of transcription engines at Fathom? And It'll go over four years of meetings and I'll write a six page memo like yeah, 10 minutes. So I think it's pretty cool we're moving this world where I imagine two fun things are going to happen in meetings. One, I just imagine meetings are going to get really good, right? Like this has been my weird mission. For someone who hates meetings, it's like how do we make meetings actually fun? One, we remove all the work, right? So like you don't have to be a stenographer, but also you don't have to get off a meeting and then have more work than when you started, right? I think that's where this is going. It's like everyone hates meeting. See, even have a great meeting. I still, at the end of the meeting I'm like, oh crap, now I gotta go do all the stuff we talked about. We're not too far away. You get off that meeting, two thirds of it's already done. The email is drafted, the follow up is scheduled. You know, the, the, you know, the presentation we went to build out is already stubbed out. Maybe it's even 80% built, right? So like one is this kind of magic of you speak things into existence on meetings. And then the other thing that I think where we're going and where the space is going is kind of like information finds you. So the other thing people hate about meetings is they're in a billion of them, right?
[12:06]
B
They're, they're, sorry, they're inability.
[12:08]
C
They're in a billion of meetings, right? Oh yeah, yeah, yeah. We're all in tons of meetings. And it's because it's like the primary way we disseminate information in organizations, right? It's like if you weren't there for the meeting, you're not watching the recording, you you lost it, right? Like, because you don't want to sit through a 30 minute recording or read the transcript, it's just gone. Right. And so if one time you were needed on a meeting, well, now you're going to be on that meeting all the time. Right. I actually think there's a not too distant future here where like, hey, we have really small meetings and if someone not in that meeting needs to know something about that, because we talked about a project they're related to, or we reference a customer that they're, they're in charge of, that information finds that. I actually imagine a world where like you only have two, three meetings a day, but you have an amazing podcast you listen every morning that's basically curated from everything that's been happening on the yesterday and what happened today. And it's f telling you, hey, here's some updates around the org. You might want to go talk to Tim about this update or that.
[12:59]
B
Wow.
[12:59]
C
And so I kind of think that's a world where you've got, you know, AI native teams, there's smaller teams, there's less meetings, but there's actually paradoxically less meetings but more shared context throughout the org. And so I think those two things, the work gets done for you and the information finds you, puts us in this like really exciting world where people can get out of meetings to get back to building stuff again. Right. And doing work.
[13:23]
B
Is that what you're, Is that what you're working on? Is that the, that's, is that like. So is that a different company or is that what fathoms.
[13:29]
C
No, that's, that's, that's our stated mission. Right? Our mission is to like make meetings amazing by kind of continuing down their source of intelligence that finds you and we do the work for you. Or we, a lot of times now we partner with agents that'll do it for you. Right? So we have API, mcp, all that stuff, so it can do some of those actions for you.
[13:49]
B
It's interesting because I, I was just thinking, so basically all these people, I just say all the different, all the different departments, all the different roles are having meetings throughout the day. I just want to understand this because I think it's. Wow. And then at the end of the day, all of that's curated into a 20, 30 minute podcast, maybe. So in the morning, all employees basically, or anybody can like, hey, what's going on in the company? You listen to it, you have full idea what's going on in all the departments. And I love what you said. Information finds you. So if something is happening on the department in a meeting you're not even part of and your name is mentioned or whatever it might be, you would get a notification saying, hey, even though you had nothing to do with it, to either be ahead of it, to understand what's going on, whatever that might be. And you're actually, you guys are working towards that right now.
[14:41]
C
Yeah, I mean we already have a version of this today where you can put in what we call trackers. And it's not like a keyword, it's just like, hey, I want to know anytime a pricing discussion doesn't go well. Or I want to know anytime there was a heated debate in like an engineering stand up or it understands tone, understands semantics and it will compile all those clips together and either daily or weekly, it'd be like, okay, here's every competitor mention, here's every pricing discussion that go well, here's the themes of, of of what these topics were. Right. In cases like that. And so we already have today that it can go find you, but you have to kind of declare what things you care about. Right. We'll opt you into a standard set. But like, but I imagine we're going to keep going beyond that to not only do you just explicitly say here's the types of moments I'm interested in, but the AI eventually just looks at the job title on your badge and says, given you and I know the projects you're working on, I'll go set up a bunch of these myself and I'll listen to all these trackers and then I'll synthesize them and give them to you. So kind of like with meeting notes themselves, I think We've got the V1 today, but I think where it's going is going to be mind blowing.
[15:40]
B
Yeah, I was just. As you're thinking, as you were saying all that, I was thinking the next layer too is it could be an intelligence for the business owner, like for the owner or the, you know, the board of going, what, what's the energy like in the company? What you know, are people happy in the company? Are people dissatisfied? I mean obviously people watch what they say on the meetings, but there is tonality, there's facial expressions, there's things that are happening that, that as a business owner you can just get a report at the end of the week and be like, hey, you might, you know, your engineering team. There's, there's a, there's an issue here, like this thing's about to explode.
[16:18]
C
Yeah, there's Not a lot of folks speaking up. There's, you know, very contentious meetings. There's a lot of stuff. And I, I'm glad you mentioned tonality because, you know, we first got in this business, everyone wanted to just like sentiment analysis on transcripts and I'm like, so much is lost when you don't have tone. Right. Like, especially in business. Right. In business it's all about to. Right. I, my background was engineering, but I ran our sales team for a minute my last startup. And you know, tone is everything in sales. Yeah, yeah. They said they're going to buy. You know, play me that clip of them saying that. Right. Like you'll know from that clip like whether they're going to do it or not. Right. So yeah, it's pretty impressive what they can do now and we're not doing it yet. But I also imagine, yes, facial recognition, like, you know, how engaged are people and stuff like that is something we'll look at in the future as well.
[17:02]
B
I, I haven't done research, like how big is Fathom now? Like the company itself by employees?
[17:08]
C
About a hundred.
[17:09]
B
Okay. Wow. Okay.
[17:10]
C
But we're, we're also kind of, you know, one of my internal goals is I would like us to get to 100 million revenue with less than 150 people. Yeah, I actually have a lot. I think actually like we're now in this era where it used to be that, you know, no one wants to talk about the revenue. No one's like going to be like, oh, here's where our revenue added, here's where our growth. So everyone just uses employees as a proxy. But I feel like that proxy is getting broken. Right. Because so many companies now are like, gosh, I don't need a 300 person sales team now to get $200 million in revenue. Sort of.
[17:37]
B
No, no, you don't. Again, that's the power. Like, I mean as an engineer is someone who's incorporating AI into your product and you've been incorporated and obviously at the next level. Where are you seeing the, like, where, where does the AI stop at some point? Because the one thing I've realized is like as great as it is today, it's still like, I don't care what anybody say, it's still not there. Like if you ever had to, like if you ever had actually asked whether it's Claude or cloud code or GPD to actually do something, it doesn't get it right. Every time like you, you sit there fighting with it. Where do you think it gets to the point where like you don't even, you just kind of like you're just talking and it's literally listening and it's literally building. And where, when does that stop? Like how does that, you know, what's the negative impact of that?
[18:22]
C
I mean I kind of look at it as like kind of going back to like the command line versus like some package software, right? Where it's like, I think we're getting this points where anyone can open the command line that's a Claude or ChatGPT and like get decent outcomes, especially for personal requests, stuff like that. But there's still a lot of room to basically engineer a better answer or a better output by being really intentional about which models you use in which order and whatnot. Right. And so I think like what we're seeing is kind of the, you know, the quads and whatnot are great general purpose solutions. When you're like, I know, I want this specific thing, you can get better speed, better accuracy, whatnot out of purpose still purpose built systems. Maybe we get to point 5, 10 years where it won't matter, right? And it's like, ah, there's like a general brain, it's good at everything, right. But at least for the next handful of years, there's still a lot of value. And I think vendors like ourselves, where we have a whole AI team that is nothing but an R and D lab that's constantly figuring out, you know, everyone thinks, you know, all the time, people are like, hey, give me all my transcripts, I'm going to throw them all in the quad and I'm going to ask it some trending questions. I'm like, you could do that. It will not succeed. Here's your transcripts. Like for us to get to the things I was talking about earlier, like those tracker concepts and be able to like basically give answers across tens of thousands of meetings. There's a lot of engineers, a big pipeline of different AI steps we have to take. Right. It's not like one agent's doing this. Think about like it's a whole team of agents that are taking on different parts of this task.
[19:48]
B
I understand, yeah. What you're saying, it's not as easy as just throwing it up, but let's talk about that so people understand because I know people do that they would throw up all a bunch of their transcripts in say Claude and say give me the, you know, the feedback. But it's, but it breaks and there's a nuance, it misses and, and it hallucinates. I mean, I mean I'M working with it right now and it's like just non stop hallucinating and I'm catching it. But for some people that don't know how to use AI properly, like it's a, it's not as perfect as people think it is today.
[20:18]
C
Yeah, I mean that was one of the biggest challenges, you know, even us kind of productizing things like this was, you know, when you're asking questions like hey, tell me every time there's a pricing discussion that doesn't go well. Well how many of your meetings have that? Maybe 1% hopefully. Hopefully it's not like 20%. Right. But say it's like 0.2%. Well gosh, then you don't need a really high hallucination rate for most of the content you get back to be hallucinated. Right. Like the more you're looking for needles in a haystack, the more, more likely more painful the hallucination problem becomes. And so, and so, you know, it's kind of funny. GPT5 last year was kind of viewed I think commercially as like not a very impactful major release. But it was actually really important to us because they one thing they fixed in that release was hallucinations. They dropped hallucinations by like 85%. And that actually opened up a whole bunch of use cases where it's like all of our use cases, a lot of the interesting ones are needle in the haystack type problems. And that's why the dropping your all your transcripts in cloud doesn't work is because one, the longer the context window gets, the less quality it gets. But two, sometimes 10,000 meanings just not going to fit into that context window. And so you have to employ a multi step process. And if one of those steps involves some agent that might hallucinate a lot, well everything downstream from that part of that process can be terrible. Right, Right.
[21:37]
B
Yeah. Yeah. So it's funny you're talking about the new models I just noticed, I don't know if I'm like I just woke up one day and Opus 4.8 is now out. Like it's creates like it was sonic. The speed at which AI is, is being produced and building. I've never seen it before. No, neither in and I think and I feel like people are not like seeing it. I just, I try to explain to people like it. It's scary if you're not understanding it and you're just sitting back and thinking that like we're going to live in a ex like that you think is going to Exist. It's not like the new world. We don't. Yeah, I'm sure you can agree, like even you as being such a visionary and seeing the future, like very hard to see what this world is going to be in the next five years. There's going to be new jobs, new role, new. New. New things that we don't even have an idea or concept of that we're going to be doing. Do you have any suspicions or any. Have you thought of any ideas of things that you can see how it would be different for us in the next five, 10 years?
[22:41]
C
I mean, I think, I think there's a couple shifts. I mean, one my buddy Emmett, who used to run Twitch and now runs this AI company called softmax talks about. I think there's a really good analogy where he describes models as kind of like, you know, certain level of education where it's like, GPT3 was like a eighth grader, right? Yeah. GPT4 was like a high school student. GPT5, you know, like, you know, it kind of says like, you know, again, four years ago we were at eighth graders doing things. Okay. What stuff would we delegate to an eighth grader? Not a ton. Right. Okay. High school student. Okay. Now we're at kind of like, kind of like unlimited grad students kind of thing. Right. It's kind of like the state of the art. Right. And so you, I think if you think about, like truly think about this as what would you hire a grad student intern to do? It really shifts your, your mindset on all these things, Right. You know, they're still going to make mistakes. And that's where I think the one interesting part is like, what does the grad student lack? It lacks business experience, business acumen. Right. And so I do think there's like this kind of world where we kind of think, you know, youth will always inherent in the world sort of thing. But I think for a lot of us that have been in business for a while, there's incredible argument to be made that actually we're in a better position to build a bunch of agents. Because managing agents a lot like managing humans. Yeah. They need context, they need autonomy, they need like, you know, guard rails, but also not micromanagement. It's kind of this interesting balance. It kind of looks a lot like managing people. And so I actually think a lot about like, how are you building kind of how are you treating the AI and how are you like building processes around it such that like, it is a lot like managing a good team sort of thing.
[24:25]
B
I, and that's where again goes to say where you need, you know, $100 million company, maybe needs a hundred engineers now even maybe less. Right. Like there are people saying that there's going to be a billionaire, you know, billion dollar company worth maybe two people, three people working it.
[24:40]
C
I fundamentally think that too.
[24:42]
B
Yeah, yeah. So my, my, my goal, my goal was, okay, that's to be true. And I, I, I do believe it to be true. Well, how many million dollars, how many $10 million companies will have four or five and, and whatnot. So, but I also see a lot of the big companies are still hesitant on really fully adopting AI still in their practice or they're looking for third parties to adopt their AI because they don't want to take the responsibility. Are you seeing that as well?
[25:08]
C
I mean, yeah, I've seen two things. One, we still see a lot of hesitancy in the enterprise to do anything these things because they're really hesitant about their data being elsewhere now that they can see the value of what you can do with that data.
[25:19]
B
Right.
[25:19]
C
With AI. But on the other hand, we've also seen that like it's actually way harder to build internal AI tools than people thought. I mean the thing you were just mentioning about, hey, there's a new model every three, six months. The other side of that coin, which I don't think people realize, is that that also means there's a model getting deprecated every three, six months too. So you go build something on Opus 4.6, gosh, you maybe get six months before you need to go rebuild that on Opus 4.8. Because the finite amount of compute in the world is sloshing back over the 4.8. And even though they have it technically EOL 4.6, it doesn't, you know, when you ask it a question, it doesn't work 2/3 of the time. Right. And so there's this interesting thing that we're doing is like we're moving a lot off this frontier models and onto open source models. Not to save money though. That's nice. But because like the, basically the upgrade life cycle on these things is insane. Right? And they're not forward compatible. A thing you build for 4.6 will work for 4.8. But like you want to start from scratch if you want to get your gr.
[26:18]
B
So you're just constantly rebuilding, constantly rebuilding.
[26:21]
C
And so I think, you know, I still think there's a place for vendors like us because like I said, for any feature we have, whether it's writing a Summary finding the action items, you know, answering questions. There's a purpose built pipeline there that usually has five or six different models in the mix, some from Frontier Labs, some open source, increasingly more open source. But like it is not. The building cycle has gotten way easier, the maintenance cycle's gotten way worse. And so like it's never been easier to stand up a prototype. This is what I want, this works. And yet that thing you'll have to rebuild every six months is almost the new thinking.
[26:58]
B
I just want to make that sound a little bit more for the everyday user because I think it's super important. The ability to build new products and services, SaaS, whatever it might be, has never been easier before, but the ability to now maintain them is actually harder. And that's because of the instability and, or because of how fast AI is growing that the models are changing so fast that right when you even figured out how to build the product and actually stabilize that product, you're now going back to the rebuild. And I do, and I'm seeing that in some of the products I'm building myself is I go, okay, I get why I need an engineer team now. Like I'm at that point where I can get it from like 0 to 5. But like you want to get it to the point where it's efficient, effective, stabilize. You need the, the AI engineer experts.
[27:45]
C
Yeah. The other interesting part is like we spend a lot of time thinking about what just got easy to build because there's a lot of times where you can go build a feature like, oh, I want this thing to exist and you can kind of almost like brute force it. Like we've had a few features where we spent three to four months to find the right incantation of models and you know, third party services to make a feature work. And then we wait six months and a new model comes out. That just makes that like an afternoon project. Right. And so there's this other part about like just efficiency of building where it's like, oh no, not only do we want to, you know, it's never been easier to build, but we want to focus on the things that are just became easy to build thanks to new release extra Y. And so, you know, I think our AI team spends half their time just reading the white papers and keeping up to date on the newest launches. So you can figure out great, what was hard last week, that's now easy because that's the stuff we want to be building.
[28:33]
B
That's, that's the thing that I'm scared. How do you Keep up. Like, you know, if you're a business owner sitting and you're listening, right. Business podcast and you got a, you know, a small, medium sized business and you're, you're just trying to make the business exist, right. And work and you're, you know, and now you're having to deal with all of this AI. It's not just about adopting AI, it's about adopting AI and then it's changing so rapidly and so fast. What would be your advice? Or you know, what could a business owner do it to. To feel like they're not falling behind but still incorporate as much AI into their business without it being something now a full time job.
[29:06]
C
Yeah, I would. I think those two like. Well, comparing us to, to that, that scenario I think is like comparing like a F1 racing team to kind of like, you know, me, me hitting the track on the weekend. Right. So like we do that because we are, we are in a very competitive space. We're trying to beat the best in the world at this. Right. And we go out every Sunday and we do a race and like we throw away the engine after every, after every race sort of thing for the average business user. I actually think the market looks very different in that like you don't really, you, you could take the thing you built on Opus 4.6 and move it 4.8. It won't be as good. No, but it'll be close enough that you won't care. Right. And the amount of gains you'll get today by just getting started today and building something will be insane. Right. I think everyone, if you haven't had a chance to use an agent or like a Claud co worker, Claud code or something and just start building something. You just got to get started. Like there's no do not let the maintenance cost. Yes, it's there be at all an impediment to getting started because you will be blown away by the amount of stuff you could do. I've talked to so many friends who are not technical who are now automating whole parts of their businesses. Like I've got friends that are salespeople that are building their own CRMs. I've got people that are marketing that are like, you know, I barely can email and are yet like hey, I built swap rating system for my marketing team. It's my right.
[30:19]
B
You're talking to someone here. Like I, I'm a sales guy, you know, traditionally a sales guy who turned into a business owner around sales who's now full on AI developer I developed like four, one of them. You know, we're talking to big companies, right, just under some NDA, but like, and it's kind of, it's like if four months ago, if you said you're going to be doing this, I would never believe you. It's, it's, it's insane. This is why I tell you it's actually insane. What happens if you just sit at the desk and you just ask a simple question. How do I get started in AI? That's what I did. I, I was at a, I tell people the story because I think it's very powerful. I was at an event in February and I was talking to AI expert like you, who's just all in, all in. And I'm talking and just being kind of a pest. And he, and he kind of just got fed up and looked at me straight in the eye and just said, hey. He almost was kind of like, shut up. He's like, listen, you're either all in or you're not. You make the decision. And I went home that night and I, and it was one of those things where it just sits and sits and burns and burns. The next day I woke up, I said, I'm all in. So what does all in mean? Well, I gotta go in and ask that, literally ask, what does all in mean in AI? And then next thing you know, I'm built, I'm seeing how it's working. You don't need to be like, you need patience, you know, and not to be afraid to ask the question. So it's, it's, it's, it's interesting to me because I feel like there's going to be a lot of these come, and I could be wrong, a lot of these, like a lot of companies are going to be coming out and it's going to be a race to, to getting customers and a race to who has the best story or marketing. But the products are going to be half ass and then there's going to be good products where the big guys are just going to gobble up. I just, I just think we're going to have so many, I'm seeing it now, just so many note taking companies out there. But. Okay, well how do you decipher which one's the best? They all have a little nuance. But who's the actual best at it? I think the ones like you or the F1, you know, race team that are working on Sundays every day, you know, like you said, throwing out the engine.
[32:20]
C
Right. Well, and then again, because we're kind of building platform stuff that other people could build on. I, I think from the, for the small business like owner, user, type, it's never been a better time to be a domain expert because the cost of building the software has gone down so much. It now means it's viable to build software in places you wouldn't before. All sorts of niches or small verticals or very specific use cases where, hey, look, I don't know everything, but I know exactly how these 20 farmers do their business and what they need to do. Ten years ago you could, you got to go raise a couple million dollars, go build it. Well, that market's not worth more than a couple million dollars now. You can go build that in a weekend. And that's a very profitable business. And so it's now kind of democratized creating software. It's like you actually, you do need folks like myself and my AI team if you're going to go build the F1 car, if you're going to try to be one of these foundational platforms everyone else is used to build on. But if you're just trying to solve a problem that you know like the back of your hand, oh boy, are you, this is going to be a gold rush for you, right? Because if you have expertise, expertise or stuff, you got those connections, you know the problems people have, you don't need to hire a 20 person team and raise $5 million to get off the ground. You can just get it done this weekend. And I think that's going to be amazing.
[33:34]
B
I'm going to leave it here because I believe we're saying the same thing. And I, and I, and I've been saying to people like with AI today, the only limitation is the mind is what you can or cannot see at this point, there's nothing you can do or can't build or can visualize or can't even bring to fruition that I can't do for you. The only thing that's limiting people is what's going on in their mind. I would agree and say yeah, 100%. Well, listen, I know that you do this. You're talking about you don't need to be on these shows. You do this because, you know, you help podcasters like me and you know, helping other business owners understand the power of it. I will just say this for anybody. If you are on meetings, this is not a plug. Never asked me to do this. I just want to make sure you understand if you are using meetings, if you're on Zoom, Google, whatever type of online meeting. You must have Fathom. It's very, very simple. There's no other product out there that is as easy to use, as efficient, as effective, and just awesome. Fathom is what you need for any last notes or any last thoughts.
[34:39]
C
No, and it's mostly free, so no reason. I'll check it out. Yeah.
[34:43]
B
Yeah. You don't. Like I always say, you don't got a $50 problem. No business in the world has a $50 problem. Again, Rich, thanks so much for being here.
[34:53]
C
Appreciate it. Thank you for having me. This was fun.

Summary6 min read

The Vault Unlocked | Why Building With AI Got Easy and Maintaining It Got Brutal

Host: Kayvon Kay
Guest: Richard White, CEO of Fathom
Release Date: July 29, 2026

Overview

This episode dives into the rapidly shifting landscape of building and maintaining AI-powered products, featuring Fathom CEO Richard White. Fathom, the leading AI notetaker on G2, was built years before AI's current boom, based on two bold predictions: transcription costs would plummet, and AI would become practical for real-world workflows. Kayvon and Richard unpack why starting with AI is easier than ever, but scaling and maintaining AI products is becoming increasingly hard and why experienced operators actually benefit from today's changes. They also explore the new bottlenecks in AI—now less about technology and more about product vision and execution.

Key Discussion Points and Insights

1. Origins of Fathom’s AI Vision ([02:10]–[08:30])

Problem Discovery
- Richard recounted constant Zoom meetings as a founder, juggling note-taking and active participation:
  
  "It's like being a court stenographer and also being the lawyer... You're kind of doing both and no one likes it." – Richard White [03:20]
- Traditional meeting documentation was stressful, inefficient, and a poor artifact for team knowledge-sharing.
Early Contrarian Bets
- Two core hypotheses in 2020:
  1. Transcription costs would fall to zero—enabling mass accessibility.
  2. AI would get good enough to generate actionable meeting intelligence, not just raw transcripts.
- Building before these became mainstream gave Fathom a multi-year lead.
Business Model Insight
- Offered the main product for free, betting real value lay in organizational intelligence for managers.
  
  "We don't think the value is in the meeting itself, it's in building up this database." – Richard White [06:00]
- Saw the need to start before the tech was mature:
  
  "If you wait until transcription cost is zero and AI is really good, you'll be two to three years too late to start this business." – Richard White [07:40]

2. AI’s Rapid Advancement and Major Shifts ([09:19]–[13:22])

Revenue & Product Evolution
- The arrival of GPT-4 marked the moment AI summaries surpassed human note-taking, transforming Fathom from "meeting recording" to "meeting AI."
- Each leap in AI (now every 2–3 months) unlocks new product capabilities—meeting summaries, cross-meeting insights, and multi-year knowledge mining.
Vision for AI-Powered Meetings

"How do we make meetings actually fun? One, we remove all the work... and then the other thing... is kind of like information finds you." – Richard White [11:11]
- Envisions AI that curates and surfaces only the critical pieces of information to each stakeholder, perhaps as a daily podcast or push notification, so people can focus on real work rather than endless meetings.
AI Curation in Practice
- Current Fathom versions already allow for "trackers" that flag key themes (e.g., unhappy pricing discussions) and proactively push information to relevant stakeholders.
- Future: AI will infer user needs and auto-curate updates based on role, project, and context.

3. The New Era of Company Intelligence ([14:40]–[17:01])

Organizational Impact
- Beyond workflow automation, AI can surface organizational health, morale, and engagement through tonality and sentiment—even potentially using facial analysis in the future.
  
  "Tone is everything in sales... You’ll know from that clip like whether they're going to do it or not." – Richard White [16:38]
Team Structures and Scale
- Fathom operates with around 100 employees, targeting $100M ARR with minimal headcount—possible because of AI’s leverage.
  
  "I actually have a lot... that proxy is getting broken, right? ...I don't need a 300-employee sales team now." – Richard White [17:13]

4. Building vs. Maintaining AI Products ([18:21]–[27:45])

AI’s Current Limitations
- While large language models (LLMs) are impressive, they remain imperfect, especially for nuanced, needle-in-the-haystack queries.
  
  "It doesn't get it right every time... you're sitting there fighting with it." – Kayvon Kay [17:46]
- Maintenance is hard because hallucinations compound when searching for rare events across vast data.
The Maintenance Bottleneck

"The building cycle has gotten way easier, the maintenance cycle's gotten way worse." – Richard White [26:21]
- AI model updates force frequent re-engineering; a feature that took months to build today might be trivial (or obsolete) with next month’s release.
- "You build something on Opus 4.6... maybe you get six months before you need to go rebuild that on Opus 4.8." – Richard White [25:18]
- Increasing reliance on open-source models, not just for cost, but to manage the relentless upgrade cycle.

5. Advice for Founders and Operators ([28:32]–[33:34])

Don’t Wait, Start Building

"The amount of gains you'll get today by just getting started... will be insane." – Richard White [29:29]
- Maintenance should not deter experimentation; even non-technical people are automating core business processes with current AI tools.
AI as a Domain Expert’s Gold Rush
- It’s never been easier to turn deep domain knowledge into software for a niche audience—even solo operators.
  
  "It's now kind of democratized creating software... If you have expertise... you can just get it done this weekend." – Richard White [32:37]
- For “F1” platform builders, deep AI expertise is required; for business-specific problems, almost anyone can build powerful solutions.

6. The Real Bottleneck: Imagination ([33:34]–[34:43])

Big Takeaway:

"With AI today, the only limitation is the mind—what you can or cannot see at this point, there’s nothing you can't build or visualize or bring to fruition that it can’t do for you." – Kayvon Kay [33:39]

Notable Quotes & Memorable Moments

"If you wait till when transcription cost is zero and AI is really good, you'll be two to three years too late to start this business." – Richard White [07:40]
"Meetings are going to get really good... You get off that meeting, two-thirds of it’s already done." – Richard White [11:29]
"AI isn’t the bottleneck—what you can see is." – Kayvon Kay, Overview [00:45]
"The biggest challenge is hallucinations, especially when you want to look for rare events in a mountain of data." – Richard White [20:18]
"The proxy of revenue to employees is getting broken... $100M in revenue may need only a hundred people." – Richard White [17:13]
"It’s now a gold rush for domain experts." – Richard White [32:40]
"Just start building—the hardest thing about AI is knowing what to even ask or look for." – Kayvon Kay [30:19]

Timestamps for Important Segments

[02:10] Richard White’s background and origin story of Fathom
[06:00] Core bet: zero cost transcription + advanced AI
[08:30] Building ahead of the AI curve
[09:19] AI’s inflection point: GPT-4 and product evolution
[11:11] The future: meetings as creation engines, information finds you
[14:40] AI’s role in organizational intelligence (sentiment, tone, "the podcast for the company")
[17:01] Team size, scaling, and shifting team structures
[18:21] What AI still can’t do reliably and why productizing it is hard
[20:18] The challenge of hallucinations
[25:18] The build–rebuild cycle caused by rapid model updates
[26:57] Why maintaining AI products is now the hard part
[29:06] Advice for business owners: how to get started without falling behind
[32:19] The gold rush for niche, domain-specific AI software
[33:34] The mind as the new bottleneck

Key Takeaways

Building with AI is easier than ever; maintaining, iterating, and staying current is the new frontier (and pain).
Experience in business processes is an asset—AI can automate and amplify domain knowledge more than ever before.
The "maintenance tax" of AI is relentless; the lifecycle of models is accelerating, requiring new team capabilities.
The real differentiator is vision and the ability to see/adapt to newly possible workflows—not access to AI itself.
Anyone—especially domain experts—can now build transformational products with little capital or tech experience.
Imagination, not technology, is now the primary limit.

For founders: It’s no longer about whether you can implement AI. It’s about imagining what AI can do for your business before others do—and building a durable team and product that can keep up when the ground shifts beneath you.

Loading summary

Transcript51 lines

[00:00]
A
Most people are reacting to AI. Our guest for this episode built for it three years before it arrived. Two bets. Transcription costs would fall to zero and AI would get good enough to actually do something with what it heard. Both were contrarian then. Both were right. Fathom is now the top rated AI note taker on G2 and Richard is one of the few people who can tell you what actually changed and, and what didn't. We get into why building software has never been easier and maintaining it has never been harder. Why your years in business are an advantage in this shift, not a liability. And why the real bottleneck right now is not the technology, it's what you can see. If you have been waiting for the right moment to move.
[00:47]
B
This is it.
[00:48]
A
Richard White's background is engineering and product design. This is the vault unlocked. Let's unlock it.
[01:04]
B
Richard, welcome to the show. I'm excited to have you. I just for the, for the guests and the viewers listening, why don't you tell us a little bit of who you are. I'm excited because I use your product. I love your product. It's in our business today. I seen you've changed it quite a bit and I'm excited to have you here. But for the listeners that may not know, tell them who Richard White is.
[01:30]
C
I'd like to think I'm a product designer and kind of technologist. You know, no one's let me write code of production in gosh, maybe 10, 15 years. So I'm not sure I can claim being a technologist as much anymore. But that was my background originally kind of in engineering design. Done a couple of startups. I worked the first batch of Y Combinator if I want to date myself. Their product before this called Uservoice. But as you kind of alluded to for the last five, almost six years we've spent working on Fathom, which is the number one kind of rated on G2, AI notetaker for people on lots of back to back meetings. It's been a really fun ride with a really great team and the most fun part about it is talking to folks like yourself who love and use the product every day.
[02:10]
B
Yeah, I mean I've used a lot of different AI notetakers and I've used Fathom before and then we switch and now I'm back to Fathom and I, I'm, you know, I'm actually, I'm sold. I, it's like it, it's, to me it's, I find it's the most easiest interface. It just usability of It. And I. And I love that. I just feel like it's not overbuilt. It's just built, like, just exactly for what it is. Take us back to where did this start? Like, where did you see that this was needed? Because the one thing I do know about Fathom was way before this huge, the AI craze and everything. So you. You saw something way before. That's what I'm interested in. Like, the vision and the strategy you saw and how you brought it together.
[02:51]
C
Sure, yeah. I mean, it was actually even. Right. Just before COVID honestly was working on a different product, was working on a totally different product and totally different space, and just found myself on a ton of zoom meetings. Like, I think it was like 15 to 20 a day. A lot of them were research sessions, right? Where I've got 20 minutes almost back to back to like, interview someone, demo something, get their feedback, rinse and repeat. And it's kind of one of those things where, like, you know, if you run into a problem once a day, you don't maybe do anything about it. You run into it 20 times a day. You're like, oh, my God, this is really painful. I need to, like, I don't. I need to fix this. Right? And so, you know, I remember just kind of. Kind of thinking how. How kind of crazy it is the way we kind of share knowledge out of, like, meetings and stuff, right? It's like, oh, I meet with someone, I had this great experience. They tell me some really interesting quotes or facts or whatnot. And then I heard we scribbled down notes and then try to, like, clean them after the meeting and remember exactly what they said. It's a very stressful situation, right? It's like being a court stenographer, right? And also being the. The. The. The lawyer kind of interviewing the person on the stand at the same time, right? It's like you're kind of doing both and no one likes it, right? No one likes taking notes. No one likes reading notes. Notes are also like, we're like a really poor artifact. I share with my team, and a lot got lost. You know, I'd have this amazing conversation, I'd share the notes with my team, and they kind of shrugged their shoulders like, okay, right? So I look at this something. Don't we have the technology to fix this at this point? Right? And, you know, if you go back to 2020, there were tools that were doing call recording. Nothing with AI yet, obviously. You know, most of the products are, like, in the sales space. Like companies like Gog and stuff like that. And they're really expensive and they were candidly kind of mediocre, right? It's like it took you 30 minutes an hour to get the recording afterwards. It was mostly just a transcript. No one wants a transcript. What I wanted was just like, I get off the meeting, there's instantly some notes. Great, like, I don't have to do this job sort of thing. And we kind of looked at that space and we kind of had this think thought, like, gosh, where is this space going? We kind of had two core hypotheses that really got us excited. It got me excited about what turned into Fathom. That was Transcription five years ago. Actually pretty. Still pretty expensive, right? It was like three to four dollars an hour to transcribe content, which doesn't sound like a lot, but if you imagine if you build a product in transcription, build product meetings, people are easily going to do 10, 20, 30 hours a month on it. Gosh, your hard costs for that product are already like $50 a month, right? So the fact that Gong and Fuch, that were charging $150 a month makes sense in that context, right? I was like, why is it so expensive? Oh, right, the input costs are expensive. And so we kind of looked at it like we think this kind of commodity, like when I. We tried a bunch of different vendors, like kind of made a little prototype and tried a bunch of different vendors like Amazon and Google, and I think one was called Rev. I was like, these are all pretty good. They're not great, but pretty good. So this hypothesis, like transcription costs will go to zero because they're all good enough and cost is always trending down. We think it'll go to zero. And more important than that, and we think AI is going to get really good. And it's kind of funny now because it's kind of an obvious thing, but go back five years. Very contrarian take because it's hard to remember. But there was a. There was a wave of quote unquote AI companies like 2015-2020 that were terrible, right? That promised you the world and delivered almost nothing, right? And so. But we're. Because I was like, no one wants a transcript. I don't want to get off a meeting and read a transcript. I don't want to read. Nothing to do with the transcript. But the AI will need a transcript to do all the fun stuff I think it could do in the future. Write your notes, write your actions, fill in your CRM. Find, find trends, find themes, alert me when certain Things happen. All that stuff needs a really good, high quality transcript. So we started the company with those two ideas and said, gosh, if those two things are true, transcription cost goes zero and AI gets really good. Could we be the first people to give away this product for free? Right? In a space where people are short you $150 a month, what if we just gave away for free? Because we actually don't think the value is in the meeting itself, it's in building up this database. Then you build a bunch of AI features on top of. And so we always have the thesis of like, we're gonna give away this product for free to individuals with the hope that that gets us into a bunch of companies where we can then sell a different product to the managers of those people. Right? Because the managers have a different problem, which is I, I'm not in the meeting taking notes, I'm out the outside the meeting and I want to know the important things that are happening. I want to know there's a pricing discussion doesn't go well. There's an argument that happened at the, at the engineering standup. There's, you know, a deadline that slipped three times, like, but I don't have time to sit and listen to every meeting, right? And so I got really excited about this business because one, kind of fit the hypothesis of where I thought the world was going. But two, it had this really awesome kind of two sidedness to it. We had one part where you can give away a lot of value for free and feel okay about that because you don't have to like charge people later because there's just nice kind of complementary business built on top of that for their managers. That's kind of how we got started, right? And it's kind of funny you mentioned that we were kind of ahead of the curve. And I think that's probably true because we had a third corollary to those hypotheses, which was if you wait till when transcription cost is zero and AI is really good, you'll be two to three years too late to start this business. Right? It'll be kind of obvious to everyone and everyone will jump in. But like, like any technological revolution, the best companies like build towards that hypothesis couple years out and they do all the other stuff, right? Like we spent two or three years building all the foundational work and the product experience that you talked about, the good user experience, good usability, the reliability, the obviously distribution channels, all that sort of stuff. And so it was a very much a go to where the puck is going kind of thing, not, like, wait for it to get there.
[08:30]
B
Then when. So when you guys are doing the hypothesis, like, this was again back in 2021, you know, Covid days. I don't like to use that word. I. I even.
[08:40]
C
I hesitated to say it because, like, I feel like we just don't use that word anymore.
[08:43]
B
I know I don't.
[08:44]
A
I don't really use.
[08:45]
B
I hate using it. And it's funny because, like, and sometimes my brain, I keep thinking, like, it's only like a couple years ago, but like, no, it's like, that's like almost six and a half years ago now, like, for, you know, so today there's a lot of players in the marketplace, but you. You've had the market share. So are you seeing competitors, like, coming in and taking over? Are you. Are you guys adapting your product now more with AI? Like, how are you staying in the trends and. And how do you see where AI is going? I mean, even just with the note taking, let alone, what is that next vision that you have for where this can go?
[09:19]
C
Yeah, it's kind of funny. I mean, like, I feel like it's been a tale of two businesses, right? There's a. I. I do this whole talk and I show my revenue graph. You could see the point where AI actually really shows up. And for us, that was kind of like GPT4 level of AI. That was the point where the AI could write better notes than the human. Right? And so that's where we went from being a meeting recording business to being a meeting AI business. Right? And it really takes off. And this whole second act of the business is not about hypotheses and stuff like that. It's actually about, like, how do we get really good at building AI functionality? Because it's actually very fundamentally different than building traditional SaaS or just software. And I can talk about that. But so it's been kind of cool. And now we kind of are seeing this, like, you know, the capabilities increase every six to 12 months. And now sometimes it's even like two to three months. Right? And so we're constantly now seeing, like, okay, two years ago, state of the art was we can write a really good summary for a meeting and we can figure out the action items. A year ago, it was, oh, we can look across not just one meeting, but every meeting you've had with Acme or, you know, every meeting you've had in this, with this prospect. And we can surface, like, risks, we can surface trends. And now the state of the arts Move into. Oh no, now we can actually look across every meeting you've had over four or five years right across your org and tell you trends about competitor trends, internal knowledge management type questions like, you know, we asked it the other day, like, hey, we don't like to do a lot of documentation here at Fathom because we just assume that's the point. You just ask the AI, like why generate documentation and maintain it? Just if you need to answer a question, you just ask the AI. Now we're at this point where that actually works and we can be like, hey, we got a new engineer and they're wondering about why we built the system the way we did. Can we, can you give me a history of transcription engines at Fathom? And It'll go over four years of meetings and I'll write a six page memo like yeah, 10 minutes. So I think it's pretty cool we're moving this world where I imagine two fun things are going to happen in meetings. One, I just imagine meetings are going to get really good, right? Like this has been my weird mission. For someone who hates meetings, it's like how do we make meetings actually fun? One, we remove all the work, right? So like you don't have to be a stenographer, but also you don't have to get off a meeting and then have more work than when you started, right? I think that's where this is going. It's like everyone hates meeting. See, even have a great meeting. I still, at the end of the meeting I'm like, oh crap, now I gotta go do all the stuff we talked about. We're not too far away. You get off that meeting, two thirds of it's already done. The email is drafted, the follow up is scheduled. You know, the, the, you know, the presentation we went to build out is already stubbed out. Maybe it's even 80% built, right? So like one is this kind of magic of you speak things into existence on meetings. And then the other thing that I think where we're going and where the space is going is kind of like information finds you. So the other thing people hate about meetings is they're in a billion of them, right?
[12:06]
B
They're, they're, sorry, they're inability.
[12:08]
C
They're in a billion of meetings, right? Oh yeah, yeah, yeah. We're all in tons of meetings. And it's because it's like the primary way we disseminate information in organizations, right? It's like if you weren't there for the meeting, you're not watching the recording, you you lost it, right? Like, because you don't want to sit through a 30 minute recording or read the transcript, it's just gone. Right. And so if one time you were needed on a meeting, well, now you're going to be on that meeting all the time. Right. I actually think there's a not too distant future here where like, hey, we have really small meetings and if someone not in that meeting needs to know something about that, because we talked about a project they're related to, or we reference a customer that they're, they're in charge of, that information finds that. I actually imagine a world where like you only have two, three meetings a day, but you have an amazing podcast you listen every morning that's basically curated from everything that's been happening on the yesterday and what happened today. And it's f telling you, hey, here's some updates around the org. You might want to go talk to Tim about this update or that.
[12:59]
B
Wow.
[12:59]
C
And so I kind of think that's a world where you've got, you know, AI native teams, there's smaller teams, there's less meetings, but there's actually paradoxically less meetings but more shared context throughout the org. And so I think those two things, the work gets done for you and the information finds you, puts us in this like really exciting world where people can get out of meetings to get back to building stuff again. Right. And doing work.
[13:23]
B
Is that what you're, Is that what you're working on? Is that the, that's, is that like. So is that a different company or is that what fathoms.
[13:29]
C
No, that's, that's, that's our stated mission. Right? Our mission is to like make meetings amazing by kind of continuing down their source of intelligence that finds you and we do the work for you. Or we, a lot of times now we partner with agents that'll do it for you. Right? So we have API, mcp, all that stuff, so it can do some of those actions for you.
[13:49]
B
It's interesting because I, I was just thinking, so basically all these people, I just say all the different, all the different departments, all the different roles are having meetings throughout the day. I just want to understand this because I think it's. Wow. And then at the end of the day, all of that's curated into a 20, 30 minute podcast, maybe. So in the morning, all employees basically, or anybody can like, hey, what's going on in the company? You listen to it, you have full idea what's going on in all the departments. And I love what you said. Information finds you. So if something is happening on the department in a meeting you're not even part of and your name is mentioned or whatever it might be, you would get a notification saying, hey, even though you had nothing to do with it, to either be ahead of it, to understand what's going on, whatever that might be. And you're actually, you guys are working towards that right now.
[14:41]
C
Yeah, I mean we already have a version of this today where you can put in what we call trackers. And it's not like a keyword, it's just like, hey, I want to know anytime a pricing discussion doesn't go well. Or I want to know anytime there was a heated debate in like an engineering stand up or it understands tone, understands semantics and it will compile all those clips together and either daily or weekly, it'd be like, okay, here's every competitor mention, here's every pricing discussion that go well, here's the themes of, of of what these topics were. Right. In cases like that. And so we already have today that it can go find you, but you have to kind of declare what things you care about. Right. We'll opt you into a standard set. But like, but I imagine we're going to keep going beyond that to not only do you just explicitly say here's the types of moments I'm interested in, but the AI eventually just looks at the job title on your badge and says, given you and I know the projects you're working on, I'll go set up a bunch of these myself and I'll listen to all these trackers and then I'll synthesize them and give them to you. So kind of like with meeting notes themselves, I think We've got the V1 today, but I think where it's going is going to be mind blowing.
[15:40]
B
Yeah, I was just. As you're thinking, as you were saying all that, I was thinking the next layer too is it could be an intelligence for the business owner, like for the owner or the, you know, the board of going, what, what's the energy like in the company? What you know, are people happy in the company? Are people dissatisfied? I mean obviously people watch what they say on the meetings, but there is tonality, there's facial expressions, there's things that are happening that, that as a business owner you can just get a report at the end of the week and be like, hey, you might, you know, your engineering team. There's, there's a, there's an issue here, like this thing's about to explode.
[16:18]
C
Yeah, there's Not a lot of folks speaking up. There's, you know, very contentious meetings. There's a lot of stuff. And I, I'm glad you mentioned tonality because, you know, we first got in this business, everyone wanted to just like sentiment analysis on transcripts and I'm like, so much is lost when you don't have tone. Right. Like, especially in business. Right. In business it's all about to. Right. I, my background was engineering, but I ran our sales team for a minute my last startup. And you know, tone is everything in sales. Yeah, yeah. They said they're going to buy. You know, play me that clip of them saying that. Right. Like you'll know from that clip like whether they're going to do it or not. Right. So yeah, it's pretty impressive what they can do now and we're not doing it yet. But I also imagine, yes, facial recognition, like, you know, how engaged are people and stuff like that is something we'll look at in the future as well.
[17:02]
B
I, I haven't done research, like how big is Fathom now? Like the company itself by employees?
[17:08]
C
About a hundred.
[17:09]
B
Okay. Wow. Okay.
[17:10]
C
But we're, we're also kind of, you know, one of my internal goals is I would like us to get to 100 million revenue with less than 150 people. Yeah, I actually have a lot. I think actually like we're now in this era where it used to be that, you know, no one wants to talk about the revenue. No one's like going to be like, oh, here's where our revenue added, here's where our growth. So everyone just uses employees as a proxy. But I feel like that proxy is getting broken. Right. Because so many companies now are like, gosh, I don't need a 300 person sales team now to get $200 million in revenue. Sort of.
[17:37]
B
No, no, you don't. Again, that's the power. Like, I mean as an engineer is someone who's incorporating AI into your product and you've been incorporated and obviously at the next level. Where are you seeing the, like, where, where does the AI stop at some point? Because the one thing I've realized is like as great as it is today, it's still like, I don't care what anybody say, it's still not there. Like if you ever had to, like if you ever had actually asked whether it's Claude or cloud code or GPD to actually do something, it doesn't get it right. Every time like you, you sit there fighting with it. Where do you think it gets to the point where like you don't even, you just kind of like you're just talking and it's literally listening and it's literally building. And where, when does that stop? Like how does that, you know, what's the negative impact of that?
[18:22]
C
I mean I kind of look at it as like kind of going back to like the command line versus like some package software, right? Where it's like, I think we're getting this points where anyone can open the command line that's a Claude or ChatGPT and like get decent outcomes, especially for personal requests, stuff like that. But there's still a lot of room to basically engineer a better answer or a better output by being really intentional about which models you use in which order and whatnot. Right. And so I think like what we're seeing is kind of the, you know, the quads and whatnot are great general purpose solutions. When you're like, I know, I want this specific thing, you can get better speed, better accuracy, whatnot out of purpose still purpose built systems. Maybe we get to point 5, 10 years where it won't matter, right? And it's like, ah, there's like a general brain, it's good at everything, right. But at least for the next handful of years, there's still a lot of value. And I think vendors like ourselves, where we have a whole AI team that is nothing but an R and D lab that's constantly figuring out, you know, everyone thinks, you know, all the time, people are like, hey, give me all my transcripts, I'm going to throw them all in the quad and I'm going to ask it some trending questions. I'm like, you could do that. It will not succeed. Here's your transcripts. Like for us to get to the things I was talking about earlier, like those tracker concepts and be able to like basically give answers across tens of thousands of meetings. There's a lot of engineers, a big pipeline of different AI steps we have to take. Right. It's not like one agent's doing this. Think about like it's a whole team of agents that are taking on different parts of this task.
[19:48]
B
I understand, yeah. What you're saying, it's not as easy as just throwing it up, but let's talk about that so people understand because I know people do that they would throw up all a bunch of their transcripts in say Claude and say give me the, you know, the feedback. But it's, but it breaks and there's a nuance, it misses and, and it hallucinates. I mean, I mean I'M working with it right now and it's like just non stop hallucinating and I'm catching it. But for some people that don't know how to use AI properly, like it's a, it's not as perfect as people think it is today.
[20:18]
C
Yeah, I mean that was one of the biggest challenges, you know, even us kind of productizing things like this was, you know, when you're asking questions like hey, tell me every time there's a pricing discussion that doesn't go well. Well how many of your meetings have that? Maybe 1% hopefully. Hopefully it's not like 20%. Right. But say it's like 0.2%. Well gosh, then you don't need a really high hallucination rate for most of the content you get back to be hallucinated. Right. Like the more you're looking for needles in a haystack, the more, more likely more painful the hallucination problem becomes. And so, and so, you know, it's kind of funny. GPT5 last year was kind of viewed I think commercially as like not a very impactful major release. But it was actually really important to us because they one thing they fixed in that release was hallucinations. They dropped hallucinations by like 85%. And that actually opened up a whole bunch of use cases where it's like all of our use cases, a lot of the interesting ones are needle in the haystack type problems. And that's why the dropping your all your transcripts in cloud doesn't work is because one, the longer the context window gets, the less quality it gets. But two, sometimes 10,000 meanings just not going to fit into that context window. And so you have to employ a multi step process. And if one of those steps involves some agent that might hallucinate a lot, well everything downstream from that part of that process can be terrible. Right, Right.
[21:37]
B
Yeah. Yeah. So it's funny you're talking about the new models I just noticed, I don't know if I'm like I just woke up one day and Opus 4.8 is now out. Like it's creates like it was sonic. The speed at which AI is, is being produced and building. I've never seen it before. No, neither in and I think and I feel like people are not like seeing it. I just, I try to explain to people like it. It's scary if you're not understanding it and you're just sitting back and thinking that like we're going to live in a ex like that you think is going to Exist. It's not like the new world. We don't. Yeah, I'm sure you can agree, like even you as being such a visionary and seeing the future, like very hard to see what this world is going to be in the next five years. There's going to be new jobs, new role, new. New. New things that we don't even have an idea or concept of that we're going to be doing. Do you have any suspicions or any. Have you thought of any ideas of things that you can see how it would be different for us in the next five, 10 years?
[22:41]
C
I mean, I think, I think there's a couple shifts. I mean, one my buddy Emmett, who used to run Twitch and now runs this AI company called softmax talks about. I think there's a really good analogy where he describes models as kind of like, you know, certain level of education where it's like, GPT3 was like a eighth grader, right? Yeah. GPT4 was like a high school student. GPT5, you know, like, you know, it kind of says like, you know, again, four years ago we were at eighth graders doing things. Okay. What stuff would we delegate to an eighth grader? Not a ton. Right. Okay. High school student. Okay. Now we're at kind of like, kind of like unlimited grad students kind of thing. Right. It's kind of like the state of the art. Right. And so you, I think if you think about, like truly think about this as what would you hire a grad student intern to do? It really shifts your, your mindset on all these things, Right. You know, they're still going to make mistakes. And that's where I think the one interesting part is like, what does the grad student lack? It lacks business experience, business acumen. Right. And so I do think there's like this kind of world where we kind of think, you know, youth will always inherent in the world sort of thing. But I think for a lot of us that have been in business for a while, there's incredible argument to be made that actually we're in a better position to build a bunch of agents. Because managing agents a lot like managing humans. Yeah. They need context, they need autonomy, they need like, you know, guard rails, but also not micromanagement. It's kind of this interesting balance. It kind of looks a lot like managing people. And so I actually think a lot about like, how are you building kind of how are you treating the AI and how are you like building processes around it such that like, it is a lot like managing a good team sort of thing.
[24:25]
B
I, and that's where again goes to say where you need, you know, $100 million company, maybe needs a hundred engineers now even maybe less. Right. Like there are people saying that there's going to be a billionaire, you know, billion dollar company worth maybe two people, three people working it.
[24:40]
C
I fundamentally think that too.
[24:42]
B
Yeah, yeah. So my, my, my goal, my goal was, okay, that's to be true. And I, I, I do believe it to be true. Well, how many million dollars, how many $10 million companies will have four or five and, and whatnot. So, but I also see a lot of the big companies are still hesitant on really fully adopting AI still in their practice or they're looking for third parties to adopt their AI because they don't want to take the responsibility. Are you seeing that as well?
[25:08]
C
I mean, yeah, I've seen two things. One, we still see a lot of hesitancy in the enterprise to do anything these things because they're really hesitant about their data being elsewhere now that they can see the value of what you can do with that data.
[25:19]
B
Right.
[25:19]
C
With AI. But on the other hand, we've also seen that like it's actually way harder to build internal AI tools than people thought. I mean the thing you were just mentioning about, hey, there's a new model every three, six months. The other side of that coin, which I don't think people realize, is that that also means there's a model getting deprecated every three, six months too. So you go build something on Opus 4.6, gosh, you maybe get six months before you need to go rebuild that on Opus 4.8. Because the finite amount of compute in the world is sloshing back over the 4.8. And even though they have it technically EOL 4.6, it doesn't, you know, when you ask it a question, it doesn't work 2/3 of the time. Right. And so there's this interesting thing that we're doing is like we're moving a lot off this frontier models and onto open source models. Not to save money though. That's nice. But because like the, basically the upgrade life cycle on these things is insane. Right? And they're not forward compatible. A thing you build for 4.6 will work for 4.8. But like you want to start from scratch if you want to get your gr.
[26:18]
B
So you're just constantly rebuilding, constantly rebuilding.
[26:21]
C
And so I think, you know, I still think there's a place for vendors like us because like I said, for any feature we have, whether it's writing a Summary finding the action items, you know, answering questions. There's a purpose built pipeline there that usually has five or six different models in the mix, some from Frontier Labs, some open source, increasingly more open source. But like it is not. The building cycle has gotten way easier, the maintenance cycle's gotten way worse. And so like it's never been easier to stand up a prototype. This is what I want, this works. And yet that thing you'll have to rebuild every six months is almost the new thinking.
[26:58]
B
I just want to make that sound a little bit more for the everyday user because I think it's super important. The ability to build new products and services, SaaS, whatever it might be, has never been easier before, but the ability to now maintain them is actually harder. And that's because of the instability and, or because of how fast AI is growing that the models are changing so fast that right when you even figured out how to build the product and actually stabilize that product, you're now going back to the rebuild. And I do, and I'm seeing that in some of the products I'm building myself is I go, okay, I get why I need an engineer team now. Like I'm at that point where I can get it from like 0 to 5. But like you want to get it to the point where it's efficient, effective, stabilize. You need the, the AI engineer experts.
[27:45]
C
Yeah. The other interesting part is like we spend a lot of time thinking about what just got easy to build because there's a lot of times where you can go build a feature like, oh, I want this thing to exist and you can kind of almost like brute force it. Like we've had a few features where we spent three to four months to find the right incantation of models and you know, third party services to make a feature work. And then we wait six months and a new model comes out. That just makes that like an afternoon project. Right. And so there's this other part about like just efficiency of building where it's like, oh no, not only do we want to, you know, it's never been easier to build, but we want to focus on the things that are just became easy to build thanks to new release extra Y. And so, you know, I think our AI team spends half their time just reading the white papers and keeping up to date on the newest launches. So you can figure out great, what was hard last week, that's now easy because that's the stuff we want to be building.
[28:33]
B
That's, that's the thing that I'm scared. How do you Keep up. Like, you know, if you're a business owner sitting and you're listening, right. Business podcast and you got a, you know, a small, medium sized business and you're, you're just trying to make the business exist, right. And work and you're, you know, and now you're having to deal with all of this AI. It's not just about adopting AI, it's about adopting AI and then it's changing so rapidly and so fast. What would be your advice? Or you know, what could a business owner do it to. To feel like they're not falling behind but still incorporate as much AI into their business without it being something now a full time job.
[29:06]
C
Yeah, I would. I think those two like. Well, comparing us to, to that, that scenario I think is like comparing like a F1 racing team to kind of like, you know, me, me hitting the track on the weekend. Right. So like we do that because we are, we are in a very competitive space. We're trying to beat the best in the world at this. Right. And we go out every Sunday and we do a race and like we throw away the engine after every, after every race sort of thing for the average business user. I actually think the market looks very different in that like you don't really, you, you could take the thing you built on Opus 4.6 and move it 4.8. It won't be as good. No, but it'll be close enough that you won't care. Right. And the amount of gains you'll get today by just getting started today and building something will be insane. Right. I think everyone, if you haven't had a chance to use an agent or like a Claud co worker, Claud code or something and just start building something. You just got to get started. Like there's no do not let the maintenance cost. Yes, it's there be at all an impediment to getting started because you will be blown away by the amount of stuff you could do. I've talked to so many friends who are not technical who are now automating whole parts of their businesses. Like I've got friends that are salespeople that are building their own CRMs. I've got people that are marketing that are like, you know, I barely can email and are yet like hey, I built swap rating system for my marketing team. It's my right.
[30:19]
B
You're talking to someone here. Like I, I'm a sales guy, you know, traditionally a sales guy who turned into a business owner around sales who's now full on AI developer I developed like four, one of them. You know, we're talking to big companies, right, just under some NDA, but like, and it's kind of, it's like if four months ago, if you said you're going to be doing this, I would never believe you. It's, it's, it's insane. This is why I tell you it's actually insane. What happens if you just sit at the desk and you just ask a simple question. How do I get started in AI? That's what I did. I, I was at a, I tell people the story because I think it's very powerful. I was at an event in February and I was talking to AI expert like you, who's just all in, all in. And I'm talking and just being kind of a pest. And he, and he kind of just got fed up and looked at me straight in the eye and just said, hey. He almost was kind of like, shut up. He's like, listen, you're either all in or you're not. You make the decision. And I went home that night and I, and it was one of those things where it just sits and sits and burns and burns. The next day I woke up, I said, I'm all in. So what does all in mean? Well, I gotta go in and ask that, literally ask, what does all in mean in AI? And then next thing you know, I'm built, I'm seeing how it's working. You don't need to be like, you need patience, you know, and not to be afraid to ask the question. So it's, it's, it's, it's interesting to me because I feel like there's going to be a lot of these come, and I could be wrong, a lot of these, like a lot of companies are going to be coming out and it's going to be a race to, to getting customers and a race to who has the best story or marketing. But the products are going to be half ass and then there's going to be good products where the big guys are just going to gobble up. I just, I just think we're going to have so many, I'm seeing it now, just so many note taking companies out there. But. Okay, well how do you decipher which one's the best? They all have a little nuance. But who's the actual best at it? I think the ones like you or the F1, you know, race team that are working on Sundays every day, you know, like you said, throwing out the engine.
[32:20]
C
Right. Well, and then again, because we're kind of building platform stuff that other people could build on. I, I think from the, for the small business like owner, user, type, it's never been a better time to be a domain expert because the cost of building the software has gone down so much. It now means it's viable to build software in places you wouldn't before. All sorts of niches or small verticals or very specific use cases where, hey, look, I don't know everything, but I know exactly how these 20 farmers do their business and what they need to do. Ten years ago you could, you got to go raise a couple million dollars, go build it. Well, that market's not worth more than a couple million dollars now. You can go build that in a weekend. And that's a very profitable business. And so it's now kind of democratized creating software. It's like you actually, you do need folks like myself and my AI team if you're going to go build the F1 car, if you're going to try to be one of these foundational platforms everyone else is used to build on. But if you're just trying to solve a problem that you know like the back of your hand, oh boy, are you, this is going to be a gold rush for you, right? Because if you have expertise, expertise or stuff, you got those connections, you know the problems people have, you don't need to hire a 20 person team and raise $5 million to get off the ground. You can just get it done this weekend. And I think that's going to be amazing.
[33:34]
B
I'm going to leave it here because I believe we're saying the same thing. And I, and I, and I've been saying to people like with AI today, the only limitation is the mind is what you can or cannot see at this point, there's nothing you can do or can't build or can visualize or can't even bring to fruition that I can't do for you. The only thing that's limiting people is what's going on in their mind. I would agree and say yeah, 100%. Well, listen, I know that you do this. You're talking about you don't need to be on these shows. You do this because, you know, you help podcasters like me and you know, helping other business owners understand the power of it. I will just say this for anybody. If you are on meetings, this is not a plug. Never asked me to do this. I just want to make sure you understand if you are using meetings, if you're on Zoom, Google, whatever type of online meeting. You must have Fathom. It's very, very simple. There's no other product out there that is as easy to use, as efficient, as effective, and just awesome. Fathom is what you need for any last notes or any last thoughts.
[34:39]
C
No, and it's mostly free, so no reason. I'll check it out. Yeah.
[34:43]
B
Yeah. You don't. Like I always say, you don't got a $50 problem. No business in the world has a $50 problem. Again, Rich, thanks so much for being here.
[34:53]
C
Appreciate it. Thank you for having me. This was fun.