Richard White | How To Embrace AI in your startup

In this episode, we’re joined by Richard White, founder of Fathom, the AI-powered meeting assistant that records, transcribes and summarises your video calls so you can focus on the conversation. Previously the founder of UserVoice, Richard brings deep experience in user-centric product design and has built Fathom into one of the fastest-growing tools in the productivity space.

Stay tuned to hear Richard’s insights on the early product decisions that drove Fathom’s breakout growth, why their freemium model worked from day one, and how they engineered virality into the onboarding experience. We also dig into how to build AI products that people actually want to use, why storytelling matters even in a product-led company and how to stay focused on solving real problems in a rapidly evolving tech landscape.

Richard's Advice:

• When launching a product, make sure it always works - not just some of the time
• Be a target user yourself: what do I value?
• What feels like ‘magic’ and appears to go beyond the possible
• Build things you yourself can and will use
• Word of mouth infinitely more valuable and trustworthy than, for example, an Amazon ‘review’
• When trying out what AI can do for your business be prepared to jettison ideas as you go - sometimes big ones
• Use humans for customer support; this is an area worth investing in

FF&M enables you to own your own PR & produces podcasts.
Recorded, edited & published by Juliet Fallowfield, 2024 MD & Founder of PR & Communications consultancy for startups Fallow, Field & Mason.  Email us at hello@fallowfieldmason.com or DM us on instagram @fallowfieldmason. 

FF&M recommends: 

• LastPass the password-keeping site that syncs between devices.
• Google Workspace is brilliant for small businesses
• Buzzsprout podcast 'how to' & hosting directory
• Canva has proved invaluable for creating all the social media assets and audio bites.

MUSIC CREDIT Funk Game Loop by Kevin MacLeod.  Link &  Licence

Transcript

[00:00:00]

Juliet: Welcome to How to Start Up the podcast that dives into the stories of startups and scale-ups told directly by the founders. I'm Juliet Fallowfield founder of Fallow, Field & Mason, a PR and communications podcasting consultancy where we empower you to take charge of your own storytelling in-house.

In today's episode, we are joined by Richard White, founder of Fathom, the AI powered meeting assistant that records, transcribes, and summarizes your video calls so you can focus on the conversation. Previously the founder of User Voice, Richard brings deep experience in user-centric product design and has built Fathom into one of the fastest growing tools in the productivity space.

Stay tuned to hear Richard's insights on the early product decision that drove fathom's breakout growth, why their 'freemium' model worked from day one, and how they engineered virality into the onboarding experience. We also dig into how to build AI products that people actually want to use, why storytelling matters even in a [00:01:00] product-led company, and how to stay focused on solving real problems in a rapidly evolving tech landscape.

And I'm wondering, listeners, if you can tell which part of this podcast production process is using AI and which is not.

Richard, thank you so much for joining How to Start Up today.

As I've just mentioned, it's our first episode talking about AI. We are 22nd of July, 2025, so I can't believe we're only just getting to it now, but it's super exciting. I would love you to introduce yourself and the business that you started.

Richard: Sure, yeah. Um, Richard White, founder and CEO here at Fathom. Fathom is the number one rated AI note taker on G2, which if you're not familiar with G2, is like Yelp for software, which is very exciting as it sounds. And so Fathom basically joins if you're on Zoom, Google Meet, Microsoft team joining

 records it, transcribes it, but critically writes all notes for you using AI captures action items, fills in your CRM, writes a follow up email, all that sort of fun stuff. And it's free, and so it'll do this for you [00:02:00] as an individual. It'll also do it for, if you're running a company, you can set it fathom for your entire team and then you get these other nice benefits where you can have Fathom tell you when certain things happen.

Like, I've gotta work anytime there's a really positive moment between a salesperson and a customer, or there's a pricing objection that we don't have a good answer for, or something like that. So, it's a pretty cool tool for people that are on meetings that don't want to take notes, which is basically everyone, as well as people running companies and running teams, especially kind of agencies, client facing teams, and they're wanting to keep their finger on the pulse of what's happening in those meetings without watching, having to watch, which are recordings, which no one else wants to do.

Juliet: oh my God, I have so many questions for you. It the fact it's free, the fact that, you know me, for example, I do a lot of Zoom calls with prospective clients and I typically have a team member there who's taking minutes. And actually one of my colleagues is like, look, why aren't we just using note takers?

It's like, you need to learn how to take the minutes like I did, and then we can use AI. But I kind of wanted to instill in them like how hard it is to get their brains really aware of what might [00:03:00] be missing if we, were using, I think Zoom's note taker, which was not good last year, but I haven't tried it since.

However, my question to you is, what problem were you trying to solve when you first built Fathom?

Richard: Like everyone, I was on a bunch of meetings. I think it was actually early 2020, actually right before, uh, the the pandemic. I was actually running a different business, a different company I had started, and I was actually doing a lot of user research for new product ideas and was on, I think I was on 15 Zoom calls a day with prospective customers.

 Yeah, I remember being kind of like excited about the product we were building, but I was even more excited about like, gosh, this taking notes thing is a terrible process all around. Right? It's not fun to talk to someone and try to type up your notes. It's not fun to try to, stress out and clean 'em up five minutes later.

And then I generally find out, I think I actually took pretty good notes, 

 but even when I take good notes, I'd look at them two weeks later and be like, I don't really remember which conversation that was. And even worse, whole point in me taking notes was to educate my team on what I'm learning.

And when I found out I'd have a great conversation with someone and then I'd go to my team and [00:04:00] show them the notes and they kinda like shrug their shoulders and they got, so much gets lost in translation. And so I think it was around that time where I was like, gosh, don't we have technology?

Can we do this better?.It's 2020, why am I still taking notes? And that's kind of where we first came up with the idea for Fathom.

Juliet: Amazing. Well, A, thank you 'cause it's great and it's so useful to have. But I think, 'cause we work in communications, we are typically talking to founders about how to own their PR or producing podcasts. And I don't wanna be taking notes, typing, or writing. Not looking someone in the eye when I'm having a conversation, but I want to make sure that things are captured.

And of course, in a startup, headcount is expensive and if I have a member of the team there taking notes, it's great exposure for them and great work experience and also useful for us both to be on that conversation. It's something about the live conversation that's different to the watching it back as well.

But for us to have efficiencies in headcount, let's be honest, it's a huge cost saving. It's a game changer, but also for me to have a tool that I can rely on that isn't gonna have a sick day or decide it's late [00:05:00] or it can't join 'cause it's WiFi's died is brilliant. It is free, how, have you managed to do this and make it free?

Richard: That was probably one of the earliest we got, right. If I go back to 2020, one of the things we recognized was that this is a very expensive product to run originally. Because, if you go back five years, transcription cost a alone, it's about $3 an hour. And if you imagine the average user is at least 10 hours a month, that's $30 a month at hard costs alone.

 But we looked at that and we kind of played with a bunch of different transcription technologies and we said, gosh, first of all, we didn't think people wanted a transcript, but we did think you'd want that AI was gonna get really good and AI would need the transcript to do all the fun stuff we wanted to do it.

Juliet: Actually, kinda a fun aside, in 2020, we put AI in our product name and all my investors freaked out on me and they're like, no one likes AI. Which, kind of funny now, right? Like it easy to forget, there was a first generation of AI tools that everyone derided. And this is what, just three years ago? 'Cause this has changed so fast.

Richard: So fast, right? And so we knew like transcription's really important part of [00:06:00] this process in order to give to an AI and that's 90% of the cost. And we honestly looked at all the transcription providers and said they're all pretty good, I think this is a commodity. I think the cost on this actually gonna go to zero.

So I actually bet the whole business that the cost of transcription was going to go to zero over the next couple years. And if it does go to zero, well gosh, then we can afford to give this away for free. And we'll lose a bunch of money in the first year or two while that cost is going down. But we don't have that many users right then, so that's okay.

So we're kind of playing this game of chicken, if you will right? Our user accounts going up while transcription cost is going down and we just need those two lines across at the right point. But yeah, so we had this theory of like, if we can get the transcription cost down, we can be the first people to give a product like this away for free.

And if we can do that, I saw from my last business how powerful 'free' is when it comes to virality. If you have a product that you think people can spread virally or just kinda word of mouth is probably the better way to put it, right? Because I think when I say virality, people think about, something blowing up on, TikTok or something.

I think about virality of just one user tends to spread it to other users, right? In this case, [00:07:00] you're bringing this thing is your meeting, it's taking notes for you, you can send the notes out to everyone on the meeting. It naturally has this like inbuilt word of mouth. 

A lot of products you use, you can't see that someone else is using them. This is not the case with Fathom, and so for that reason, we saw if we can make it free, not only will people see it, but people recommend it. Because if I charge you money for something, are you gonna recommend to your friends only if it's like really, really good even then, not really. If it's free, that's a gift you're giving now, right? So now you're like, 'oh, you should try this thing out, it's free'. And so we bet a lot of the business on this idea that transcription costs go down. We can give this away for free. If we give it away for free, then we can spread like wildfire.

Juliet: So getting those user accounts up and getting the word of mouth out there, the power of word of mouth. I mean, you're preaching to the converters with someone who works in PR, but how do you revenue generate then?

Richard: When we first started this business, the thing that got me really excited about this model was, youin my intro I talked about there's kind of two people we help, right? We help people on the meeting that don't wanna take notes. We also help the people not on the meeting, [00:08:00] the managers, the executives, the colleagues.

And we saw like, hey, we actually think, yeah you could probably charge,$10-$15 a month to individuals at some point, and they would pay you to not take notes. But, there's a way more lucrative opportunity to charge the managers at some point.

Hey, you want visibility into what your team is doing, you wanna help coach them. You want, knowledgesharing across your org, they'll probably pay you $30 a person.So, it's a much bigger opportunity, and then you have this nice separation of our people using the product we're not gonna charge 'em anything. They just get this awesome free product because we know one day their usage will, get an intro into their manager, their founder or whatnot, and rolling that out across the team.

It's much easier to go on, be a B2B product than to be a consumer product from a pricing perspective.

Juliet: And that's also the case is when people use it and they're absolutely dependent on it as well. I've seen this happen in the podcasting space I mean, even three or four years ago we were paying for transcriptions and now it's all built- in software and you become so reliant on it that you're like, you know what?

I will if they up the [00:09:00] pricing. In fact, I really hope they're not listening to this. But I would pay more because I've drunk the Kool-Aid and I need it and I like it and I've tested it, and now I can't live without it. And also $20, $30 a month is far cheaper than a person that you are employing.

Richard: Yeah, I think free was also important to us because we were asking people to make behavior change, even though no one likes to take notes, they all know how to take notes. And so if I tell you, 'Hey, I need you to pay me $30 $50 a month to give you a product to take notes, you're gonna be skeptical, especially if you haven't seen something like this before.

 Again, it's all new AI stuff, so no one had seen anything like this before. And so I think free is also valuable, not just for word of mouth, but for when trying to sell someone something they've never bought before. There's already enough friction there because they've never used something like this.

You don't wanna add more friction by saying, oh yeah, before you even get to that aha moment and get the value, I need you to consider that you might need to pay me money for it. And so I really like that separation for that reason too.

Juliet: So you didn't even do a two week free trial and then you start paying your subscription? 'Cause I always clock, I remember I, who, it was it, there was a podcast about [00:10:00] Netflix and how, subscription income is game changing. And anytime I signed up to a subscription, I cancel it immediately just in case I forget.

And then if it. Cancellation happens and I still want it. Then I'm allowed to pay for it again. But I go through my Apple subscriptions, cancel 'em all. I am the polar opposite to my brother, who probably has about 30 and has forgotten all about them, but scrupulous with finances, I think is a business 101.

For you, you're not even doing a cancel after a month, it's just free.

Richard: It's just free. And now we have some upgrades you can get but the core product that takes notes for you is forever free. And I think we looked at that and said, when freemium stuff first came out, I think a lot of folks weren't savvy or rather consumers weren't savvy to the model.

I think now almost all of us are skeptical of two things almost inherently. One is can I actually use that free product? Or is it like gonna gimme just enough and I know you're gonna be a good enough business to make me trip over something I have to pay for it.

And two, I'm savvy enough to be like, if it's $50 for a 30 day trial. I'm still evaluating as i'm gonna have to pay $50 for this. And so we not only made it free [00:11:00] forever, we had no pricing page in the beginning. For the first three years, we had no premium tiers.

We kept it really, really clean. And in fact, the biggest challenge we had is we had to explain that this wasn't a scam, 'cause people were like, this is too good to be true. And so we actually AB tested this on the homepage it's like, no, this really is free, here's how.

And you click the, here's how, and it would actually pop up a little,dialogue in the corner. It's actually still on our website today, that explained, here's how it's free, because I think, rightful after 10 years of truly free products like Facebook, you are the product, right?

That old saying. You know people are like, oh, are you selling our data and stuff like that, which always funny to me, 'cause can you imagine wanting to read someone else's meetings? I don't even wanna read my own meetings. But had to put in this disclaimer like, 'Hey, we're not selling your data, here's our model.'

One day we're gonna charge your manager, some manager can opt into a premium tier and that's how we're gonna do it. But we're not doing anything shady here, we just found a wayoperate this product really cheaply and so we want to pass that on to you.

Juliet: And is that working now? 'Cause your, what, your four years in?

Richard: Yeah, we're, four years in,We are now an eight figure revenue business, pretty quickly, I thinkin about two years [00:12:00] of monetization frankly. But, still majority of our users are using the free product and we love that. And again, I think, having a forever free makes a lot of sense for us because of that word of mouth component.

Because every time someone uses Fathom, they're in a meeting, basically there's a chance that they will introduce us to a new user. And so, forever free makes a lot of sense for us given that model. 

Juliet: The acquisition because whoever's using it when their colleagues or clients or whoever else joins that meeting sees it and it's just branded right there, it's there. You can't not notice it. And that word of mouth, it's interesting 'cause we as a business teach founders how to own their PR in house and we produce podcasts and, it's that word of mouth. If someone, or we call it in the beauty world, 'mascara effect,' if you see a friend wearing a great mascara and you ask them what it is, you are far more likely to believe them because they tell you and they're your friend than some random person on an ad or a piece of social content.

So, if someone face to face that you are having a live conversation, it's like, oh no, it's great and it really works and it's embedded and it's just sorted. That virality, I'm sure just went crazy. Did it [00:13:00] kickoff straight away, when you first launched, could you see it go nuts straight off, or did it have a bit of a slow burn to start with?

Richard: It had slow burn. I think we did something that I advise most founders to do, which is we spent a long time in a closed beta, like really iterating the product to make it really, really good. I think the cardinal sin I see a lot of folks do is something that kind of technically works, but one's gonna get excited to tell their friends about.

And so, 

you know, we spent. 

 a year basically building this with a group of 50 people, And then it got to the point, okay, this is good enough to launch. And even then, I think it still had its warts and kinks to work out. So I'd say it probably took us 18 months, until we really started to see everything start working really well, and people just saying, 'I absolutely love this,' and NPS scores through the roof and stuff like that.

Juliet:  you mind elaborating on NPS? 'cause it, it's recent in the last three to six months, founders have been talking about this, but I'm not sure everyone knows what it is yet.

Richard: Yeah, Net Promoter Score, so it's just, it's the survey. we've all seen it probably before. It's like scale of one to 10, would you recommend this to one of your friends and colleagues? And, [00:14:00] depending on your industry, it doesn't work super well in B2B, 'cause a lot of people take it literally and they're like, I'm using this accounting software, would I recommend it to my mom?

No. But when you're doing something that's kind of consumerish like we are, it's a good bellwether for 'do people love it?' And the way it's scored is you're looking for people to give you nines and tens. In fact, you only basically get points for nines and tens and then deduct points for everything that's six and below.

And so, all the best brands, Starbucks, Netflix, you name it, all have 70 plus NPS. Software companies are generally in like the 20 to 40 range. We actually saw, we were in the eighties on NPS, and that was like a really good sign to me early on. Okay, we're doing something right here.

Juliet: Going back to your timing of launch, a lot of people have said to me, a lot of founders hold back and they wait for it to be perfect before they get it out. And then it's like, actually just get it out, 'cause then you hustle 'cause it's out there and you have to get going. But for you, I think with something like this, I agree it has to be really working., For you, the game changing moment was the fact that people were excited about it.

Richard: Yeah, so we, for us it's you pointed this out earlier, it's like it needs to work reliably, right? [00:15:00] This is a product where you're gonna come to depend on it, and actually making this product work reliably is actually the hardest part about it. We weren't the first people to build something in the space, but a lot of the first tools that we saw were really buggy and you we'd interview their users and they're like, yeah, it works. Like 80% of the time. Okay, well my note taking process that I've perfected over all my career works a hundred percent of the time, right? If we're gonna replace that. We need something that also works a hundred percent of the time.

So I like to decouple, launch, like the marketing side from you should as soon as possible try to get people not named you using your product. the first user for our product was me, right? And Ithink it's a great hack when you can be your own target user.

And, build something that you absolutely love and use day in day out yourself. And then convince your friends and your family to do it. and I would say if, like you can't convince your friends family to use and love your product day in, day out, yeah you might do, some soul searching about where your project's going. And then expose it to like, what do their friends and family think?

And I think it's so hard in this day and age to make a first impression because there's so many products coming out. I have this series like you kind of wanna see in this beta phase [00:16:00] where the product is working, you're still iterating it, you're putting in front of users, you're just not putting in front of random people on the internet yet. You want to basically wait till you've got some groundswell and then you launch and the time you launch, you're on product hunt, you're somewhere, but people are already raving about it, And it's not a, this might be good, but, oh, I go to the website and I can see lots of people raving about, or I can see that on social media sort of thing.

So yeah, I have a bit of contrarian take on that. Again, I think you should be building and iterating, but I think you wanna wait, from a marketing perspective, to make a really good first impression because it's so hard to make a second.

Juliet: So was that beta phase,tried a tested safe space for you guys? You had a kind of closed cohort where you could throw things at them and they give you feedback and you were comfortable in breaking your own business privately before you went big.

Richard: Right, yeah. for us, what it looked like is, I personally was reaching out to people.Friends of friends or connections, on LinkedIn. All sorts of different, really unscalable, really grindy kinda ways of getting people to sign up for this product 

Juliet: You weren't using AI. 

Richard: Yeah. [00:17:00] We weren't, using AI, we didn't have that yet. I think we put 50 new people into the product every week or two, see how that group went and put another 50 or so. And honestly, it took us a while, I it went through about 800 people before I got to a group of people where I put 50 people into it and roughly those 50 people kept using it day in, day out. We rid of enough of the bugs that were getting repeat usage, but we burned through hundreds of signups before that, right? People were like trying it, oh, we found this bug. Oh crap. All right.

Okay. Never getting that user back, but at least now that bug is fixed for the next user.

Juliet: Yeah. Amazing. And given how quickly you've done this, how did you decide which features to prioritize? Because I have no idea, all of the kind of tiny things that you'd have done to build this amazing, what is now quite a simple thing to use, but behind the scenes it would not be, how have you gone through that prioritization decision making 

process?

Richard: I think one of the hacks here was me being a target user myself, right? And so it's kinda like, what do I value most? But I think my core product principle optimize for magic, especially in [00:18:00] AI. what feels like something I could never do as a human, lean into those things.

'Cause those are things that are going to get people excited and blow their mind. For us it was very straightforward, you know, it was very clear from the early days, people hate taking notes. Give them all the different ways to reference back to points in this meeting, sharing clips, getting summaries, that sort of thing.

We also had a very interesting thing where, you know, when we launched AI actually really wasn't that good yet, right? We launched in 2021. Really, the tipping point for AI getting really good is about two years later is when kinda GPT four came out and the AI could write better notes than the human.

So we had this interesting thing where in the beginning, we actually were just really a recording and transcription engine that allowed you to pull clips outta meetings. And we've always built that, knowing that eventually we'll drop in the AI piece, And so, in the beginning we knew the key killer feature was just, it works reliably.

So like something that we reliably records and transcribe your calls. We had a feature where you could like, while you're on a meeting, click a button to flag that part of the meeting. 'Cause we learned no one to [00:19:00] go back after the meeting and find that interesting tidbit. So we kind of focused on that, we really focused a lot on reliability. That was actually the thing we saw early on, it's like speed and reliability are the two killer features. The speed one's also interesting because I think, note taking or who we're trying to replace is instant. For all the things we hate about trying to take notes, when I get off the call, my notes are. And most of the technologies we were competing with or things that were already in the market for this. They record a call, they transcribe it, maybe they give you a notes, but they take 30 minutes to an hour,You know, I gotta do my work after, I gotta send out the email, I gotta update my task tracker, right?

And so one of the big things we worked on around was just like speed. And I remember my engineering team, like, how fast does it need to be? And I was like, I don't know. Keep making it faster. And I'll tell you when it's close enough, right? We started off, it took us 30 minutes too, and got down to seven minutes and three minutes and then a minute, and then 30 seconds, we hit 30 seconds.

I was like, this is fast enough, 30 seconds versus 10 seconds, I really can't tell the difference, when I get off a meeting. I think that plus reliability were the two killer things we created a lot on. 

Juliet: Which is why I guess a [00:20:00] lot of people are panicking about redundancy. And Idon't look at it as redundancy, I just look at it as an evolution as the next industrial evolution. Like we've gone through things in the past that have been game changing and we are all using them today. Imagine, going 'oh no, we can't use a laptop 'cause we're gonna give up writing with pens.' We send a lot of minutes after our meetings with our clients, and we know the clients aren't gonna read them. It's more for us to have our task list and be like, guys, you promised us you give us this and it's written and we've highlighted it for you and you're still not giving it to us, but we're forwarding the email again.

So if we could just take out a good hour of going through our minutes, going through our notes, highlighting actions, moving things around, updating trackers, all sorts of things. To your point, if it can make better notes than we can, I'm in.

Juliet Fallowfield: We're taking a quick 30 second break from this episode to ask, are you thinking of starting your own podcast? If so, this is great news because we run a course that will train you how to produce your own podcast yourself. This course is for people who do not want to outsource and pay [00:21:00] someone else to do it.

You want to learn how to do it yourself. We teach you everything that you need to get up and running with your own podcast show, and so much more. Just DM us at hello@fallowfieldmason to book your spot.

Juliet: What did your previous experience at User Voice teach you that you brought into Fathom?

So I ran a company before this for about 10 years. We got to about, high kind of seven figures in revenue. Was it your business?

Richard: Yeah, I mean, I was one of the founders,along with a bunch of great investors and colleagues and stuff like that. Honestly, I call that my finishing school.

It was a, business I ran for 10 years, we did every business model under the sun. We started off as like a PLG company before they called it PLG. When I left it was very much an enterprise business, we pivoted a couple different times, got to work with a bunch of great folks.

I think probably the most valuable thing that I learned from that business is I got to, when we'd be in between executives, I would step into those roles. And so, I ran marketing for us for a while, I ran customer success for a while, I ran sales for a while. And those last two are also really kind informative [00:22:00] to fathom, right?

Because not only could I see this problem from the perspective as the person trying to take notes on a meeting, but also from someone who's tried to as a sales manager, consume other people's notes try to figure out what's happening on this meeting. So there was really good perspective, but I think, all in all, it was a good product in a really hard market, my previous company.

And so I think you actually learn a lot when it's a struggle. I've got some friends who are, fantastically successful. And a lot of them, they're very smart folks and they're great business people, but they all say hit the right thing at the right time.

Sometimes you hit the right thing at the right time and it's just like the ball rolls downhill. We didn't do that with my previous company, we definitely had to push the ball up the hill. The upside of that is that we taught us a lot about how to be really good at execution. And so, now we combine all those warnings for how to be good at execution with the ball rolling downhill like we have here at Fathom. It just really multiplies and stacks. I think that's probably the biggest thing I learned. I think the other thing I learned is run a business for 10 years, has, its kind of ebbs and flows and don't get too excited when the graph goes, really up and don't get too down when the graph goes too down. 

Juliet: I think it's, every single [00:23:00] guest, I think a hundred and thirty, a hundred thirty five, be okay with a rollercoaster. I'm like, I'm not okay with it, but I accept it, but I'm not okay with it. I don't like it. I'm like, Chanel trained big luxury brands where everything, you're like a swan, everything is fine. And in a startup it's not.

And if you kind of be at peace to the fact that every day is a learning curve, or someone said, everyday is a school day, I'm like, I'm really tired. And then the other thing that someone said, it was the trampoline effect where the shit hits the fan. The markets all falling apart. The world's falling apart, and you just hit rock bottom and you get laser focused on what really matters and you get your answers. Happened to me last year I was buying a house, moving house, the business was changing and all of our regular retained clients were changing the business decisions. We lost 60% of our revenue in one month and we had a three month notice period with him. But still I was like, can I afford to move my life? This is really impacting every decision. And I just got really focused on what mattered.

And it's like people, headcount, process, what clients, what work do I wanna do? Do I actually [00:24:00] want that work? You have to ask yourself those horribly difficult questions and be totally fine with it. Some days you're gonna have the rug pulled from underneath you and that's okay. So when everyone's like, oh, you, self-employed, you have limitless holiday.

I'm like, are you joking? But do you think founders need previous experience before they do their own thing?

Richard: No, I think previous experience, I think it makes you better at some startup ideas. I think it makes you worse at others, frankly. I think there's a lot of kind of B2B problems that you could solve with a startup that you would benefit from having experience running B2B companies or understanding how businesses work. when I was outta college, I didn't even know what marketing did, I couldn't tell you what two thirds of the business functions were sort of thing. And so I think, there is a lot of value and experience in figuring out how to do things faster.

 I like playing video games on the side, and so I think of startups as playing this big video game. And I think of it as like, 

 are playing Minecraft. It's kinda like this open world game where there's no rules. You just kinda start doing stuff and that's kinda like your first startup where like you don't have no idea what the rules are, what you should be doing.

Then you, your second startup is kind of like playing, Minecraft again after [00:25:00] watching 10,000 hours of videos about it. You know exactly all the things to do and things that felt like open-ended questions now I feel like multiple choice questions, so there's a lot of value experience.

The big negative to experience is, you know what we all kinda call the curse of knowledge. You tend to think about things in specific ways. And I don't think it's any surprise then that a lot of the really big companies, especially big consumer companies, are created by founders with almost no experience because they have no cursive knowledge, they have no pre-existing notions of how things should work.

They just have to figure things out from first principles and say like, 'gosh this should just exist in the world. So, I think it's a trade off, right? I think it's helps you with some things, hurt you with others.

Juliet: Yeah, someone said, I think it was two years ago in the podcast, they said, you just have to be so unbelievably deluded that what you're doing is needed. So even when it isn't needed, it still is needed and you'll just keep going. 'Cause you have to be so bullish at keep pushing 'cause if you don't believe it, no one else is going to.

Richard: I mean this is why I love building things that you yourself can use, right? And like before anyone else was a habitual user or Fathom, I was a habitual user of Fathom and I saw how it changed how I worked. It gave [00:26:00] me a lot of conviction 

 because I didn't, I knew like, no, no, I know what this product can do for people's working lives, right? I know the effect it can have, it has it on me. And so I'd always just kind of fall back on that, like, oh, maybe we got this marketing thing wrong, or maybe we didn't nail this feature or whatnot.'

But I know at our core we built something that I think is game changing, and I think that's super helpful.

Juliet: A lot of people say they've created businesses and brands that they wanted to solve themselves. So there's one woman that had, celiac and she couldn't find any food that she could eat, so she created her own food brand. And I've just realized, I started teaching founders how to do their own PR, I started a podcast and now I produce podcasts.

And it's because I needed the podcast that I needed to learn from. So it sort of, I suddenly realized, hang on, I'm doing exactly what other people are doing too it's like you're fixing your own problem. Which I love because, also you are genuinely interested in it and you are much more likely to be passionate about it.

And when you put the hours in, you need to like what you do. Fathom has grown significantly without relying on traditional sales or hype. What's your take on building a product led company in today's [00:27:00] climate?

Richard: I mean, I do think we have a lot of hype from our users, but I know what you mean the traditional, marketing thing, right? Like we-

Juliet: You've let them do the hard work for you. 

Richard: Right, you know, we said early on we can't, I think this is something I see a lot of start up founders do, it's like we try to do all the marketing channels from day one.

It's Most successful startups have one good channel from day one. And for us we knew that was gonna be word of mouth. And so, we like to generate hype, but we generate hype through providing amazing value to our users and then encouraging them to tell their friends and colleagues and LinkedIn connections about us.

And so I think, honestly, I think social media has weaponized word of mouth. There was this era, if you go back to like eighties, nineties, early two thousands where word of mouth actually fell out of favor because when the internet first came out, what you could do with buying ads and what you could do with marketing and whatever was so powerful.

That it kind of dwarfed word of mouth for a period. And then I think about like social media basically put word of mouth back on equal footing. And now I actually think it has the leg up because those traditional channels are so saturated [00:28:00] and so kind of untrustworthy now.

You know, you mentioned earlier about I'll trust a review from a friend over 5,000 reviews on Amazon because we were like in this age of like lack of credibility of things online, right? So I think actually there's never been a better time to build a product that is really focused on word of mouth because it's by the power of social media and it's the last credible source we have 

Juliet: Yeah, that trust is key and people are being very savvy where they spend their time and their money, they're gonna want to trust where they spend it. Obviously AI, everybody's talking about it, it's on the news on a daily basis. For founders thinking of incorporating it into their business, what would you advise 'em to do and not to do?

Richard: How much time do we have? 

Juliet: As much time as you can give it,

Richard: uh,

uh, yeah, I mean, two things, i'll say this. So one is I think I see a lot of folks doing the, sprinkling AI all over their product a little bit as opposed to finding one killer use case for it. I think this is just a classic product mistake, right?

Where it's oh, made a bunch of little features that in isolation [00:29:00] are kind of valuable and we hope that if you add 'em all up, you get this great experience. It tends, not to work that way. So I think, instead of trying to find a bunch of little AI helper things, if you can really focus down, is there one core thing that AI could really revolutionize for us?

And usually it requires you to be willing to sacrifice your best feature. Usually it means look at the thing that drives your business and be willing to potentially, kill the golden goose to get a better one with AI. so I think it's probably the hardest ones, like and that's easy thing for me to say, but when someone's got, millions of dollars working based on this feature, and you want me to kill it with AI? That's tough, that's a classic innovator's dilemma. I'd say the other thing with AI is, it's kind of, you know, I've been doing software for 20 years.

This is easily the biggest change I've ever seen, and it fundamentally has changed how I think about building software. Where historically you build your roadmap based upon what do you think is gonna provide the highest value for the lowest effort, and prioritize that first.

Now we kinda work more like an R&D lab. Where I have a team of people that are constantly playing with the latest models and playing with them against a, we almost got a board of like 20 different [00:30:00] things we'd love for AI to do that we think you might be able to do in the future. And they're of constantly playing, I call it the Jenga style of development where it's like we're pushing on blocks and like sometimes we push on a block and like, which like a use case and like a AI, it's like, oh, can AI do action engine really well? No, not really. Okay, let's try something else. Can I write summaries really well? Oh, okay. You can do that. And so there's this interesting thing where you almost need to work from, what can the AI do well today?What's in the sweet spot? and the sweet spot is fun. It's yeah, maybe you could do it today, but it's horrendously expensive, which also means it's horrendously slow, right?

So there certain things we could put in Fathom today, but it would cost me $200 every time someone clicked that button. Price is a function for speed in AI world. So it also means click the button and 30 minutes later I would be out $200 and they'd have some cool output.

But like, okay, that's not where the sweet spot is. So one is, I think you've gotta get comfortable with this world where you almost need an R&D style where you're just constantly playing around with things and being willing to throw away 90% of it to find, ah, I found a sweet spot of, this model does a really good job of doing this job for our users.

[00:31:00] Also because if you force it, you'll find that six months later new AI comes out and just could spend six months making it, forcing it to work, and the new model comes out. And I think the last thing I'd say about AI is that it's really easy to get AI to say something.

And I think when you look at all AI products where they fall down, you mentioned Zoom's AI product earlier, where a lot of them fall down is, yeah it'd spit out a summary of your meeting. But like it was really verbose, it was almost verbose transcript or it's so concise that it lost all the important nuance of the discussion.

And so I think there's a real, almost like artisanal artistic lens to look at AI products. You don'tnotice this, but the right AI products have teams of people on the backend that are really meticulously looking at the outputs and tuning their prompts and tuning the LM's they use and whatnot.

Because That's where I think you win. it's really easy to get something to spit out something. It spits out the right thing and the right thing. Is almost impossible to spec out, right? It takes a human look at it and be like, yeah, and that's why, chatGPT asks you all the time, do you like option A or option B of these responses, in the [00:32:00] beginning you don't have the scale to do that. You yourself need to sit there and mind numbingly honestly, look at outputs of AI's for hours on end. And finally, yep, this is the exact right incantation of this LM.

Juliet: Is there anything you don't use AI for in your business?

Richard: kinda had take a barbell approach to this where it's like we either want AI to do it or we wanna really invest in humans doing it. And so I think, classic thing you know, we started using AI for customer support a few months ago, but we also continue to invest in our support team and now we get to put them on the harder support

problems. And we tell them, great, spend more time on the top 10% of problems that are the most gnarly and stuff like that. And we do this across the org, whether it's support, customer success, sales, I still think there's a lot of opportunity I think there's a lot of AI companies, like they do the AI and you can't ever find a human or talk to them.

 but that's kind of weird. It's like, it's taken away the lowest value stuff. So now I can actually afford to give you a really high-end white glove experience. And this was also a theory we had on Fathom early on. Yeah, we're a free product, but we're gonna give you [00:33:00] service as if you're paying us thousands of dollars.

And so from a support experience, from a sales experience, because again, I'm betting on the fact that long term the AI's gonna do a lot of the easy questions. And so I can afford to give you, if you don't spend a lot of money with us, a really white glove treatment, if you've got questions that are different than what the AI can answer, I think that's a really good model for how to build a small but like highly available team that customers ve about.

Juliet: Yeah, and customer service is everything. I remember working for luxury brands and they didn't have any customer service, it's like, 'you can't charge this much money and not have someone sorting out problems'. But yeah, that level of service I use, uh, Buzzsprout for our directory host and I,also, be their biggest fan forevermore because their customer service is phenomenal.

$18 a month, I will email help desk and within five minutes I always have a, someone come back with the answer they've never not been able to answer. Even if it isn't a problem that they should be fixing for me, they'll help. And that customer service is worth its weight in gold. And I don't think I'll ever leave.

Even if I stopped doing the podcast, I probably still [00:34:00] will pay them 'cause I love them so much. But yeah, that human element, they may be using AI in the background to find the answers, but yeah, that is super, super impressive. 

Richard: What I like to do with AI is actually use it in the foreground. So for support, we will tell you, Hey, we've got two paths for you to get support. You can talk to our AI, which is instant, or you can ping our team and at any point you can, know, no, I wanna send it to the team.

I think it's also like respecting the users enough to be like, look, I trust you to make a decision about what level of service you want. And so I think what people hate is when they get forced into an AI or like you're pretending it's a human, but we all know it's AI sort of thing, right?

And I think, I don't know why companies do this. I also don't know why companies don't invest more in customer support and customer care, 'cause it's such a high ROI investment because talk about trying to stand out from the competition, I don't know, my last company, we, I spent 10 years trying to convince companies to invest in support and never could.

Juliet: So now I'm just like, well, I'm just gonna use that as an advantage since no one wants to invest in it, we will invest in it. And if you go look at our tier views, it's the number one thing we get ranked for. I, years [00:35:00] and years ago, I heard of this story that a guy in the States, it was the days where you'd phone up, say a telephone company or something like that, and you'd go, hit one for this, press two for that, hit the, and you have a menu of 10 numbers. And he wrote out the shortcuts to every menu for all of the big companies in the us and he published it saying, you just press 1, 7, 10, 9 and then you'll get where you want to get immediately.

I was just thinking for AI now he'd just be like, done. Easy. He was just like, I'm just gonna save everyone else a world of pain. Running a startup is a rollercoaster as we know, and anyone that's listened to this podcast will have heard in every episode that, but how do you manage the highs and the lows of it?

Richard: I still live my metrics a bit, right? We have a good month, rich is very happy, right? We have a bad month ehh I'm a little less happy, right? So I've just tried to over time to dampen that effect a little bit.I'm a big fan of, there's a book called,Score Takes Care of itself by Bill Walsh, who's the football coach from the San Francisco, 49ers [00:36:00] from the eighties, won all the Super Bowls. And I love the philosophy, which is kind of more of like if you take care of all the little things.

Score takes care of itself. And I think a lot of founders hang on revenue numbers or usage numbers and stuff like that. And yeah, I still do too a little bit, but I try not to because that's like the most lagging indicator, right? I tend to say focus on, hey, do I still love this product? 

I still see high NPS scores? Do I still see anecdotally in like in all the feedback channels we have, do I still see all this stuff? And I look at all the little stuff and realize yeah, sometimes metric, ebb and flow. Sometimes it's just seasonality, right?

Summers are slow, like spring is fast. if you just live, the only metric you look at are like top line stuff, revenue and top line usage. Yeah. You're gonna see a ebb and flow. You want to get down one or two levels down to the stuff that's a little bit, more closer to the ground and also more anecdotal and just be like, is this true? And again, do I still love this product and still I have high conviction that it's working for me. Great. Then let's put some blinders on.

Juliet: Yeah. Oh, I love that. are you happy in your own job that you've made up for yourself is always a good one. It's do I [00:37:00] wanna do this today? Yeah, I do. Cool. That works. And finally, what's the one thing you wish someone had told you before starting Fathom,

Richard: I mean, I wish someone told me how long COVID was gonna last. 

Juliet: in, when you get it? Or- 

Richard: No, no, no. Like how long that pandemic was gonna go on for, I think, 'cause we started middle of 2020. And, we always aspire to be a fully remote company anyways, but,someone told me about the market drop of 2022 that would've been helpful.

yeah, I, you know, honestly, I think I wish someone would've told me how quickly AI was gonna come and how fast. I think we at the company and AI and even we were surprised by how quickly it arrived and how good it was. And so we went from, in 2021, I'm telling investors, Hey, don't you know we put AI in our name?

Don't worry. we believe in this to, literally 18 months later, people are like, oh my gosh, this, it's here, thing. Probably one of the most amazing revolution Yeah. In, in technology.

Juliet: Goodness, yeah. I was thinking it's quite a similar analogy to dating in terms of right place, right time, right person. You need a few things to line up and you've got all of those sorted and then it's [00:38:00] happening. very exciting. And it's so exciting to see the business fly. I love it when you see companies just completely, the trajectory is there, something that we do is a question from the previous guest has a question for you.

And our guest this morning actually was Daisy Bird. She runs a beautiful travel PR agency here in the UK. She's 15 years in, two months ago, and we were talking a lot about people and culture and how she really nurtures her team. And we both had this thing. We're both in our forties of managing the next generation and how we adapt our management style or a leadership style to.to really cultivate that culture with that Gen Z employee. And her question was, how do you get the best out of your Gen Z employees? Do you adapt your leadership style to create the leaders for the future? It'd be great to get your take on that.

Richard: It's a tough question for me. 

We're a fully remote company. And I think maybe that's part of the reason why I don't have a good answer to this. I don't have any idea how old anyone of my team is. I don't even know who on my team. I don't even know if we have Gen Z people.

If we do, I don't know how many, we're about 90 folks now, and we're spread from Philippines to Portugal. [00:39:00] to meet everyone on the team, but I don't know, maybe it's 'cause of filters or I just, I couldn't tell you even if we had Gen Z folks on our team, 

Juliet: That's so interesting. No, I remember getting CVs of people being born in 2000 and something. I was like, what? You are far too young to have a job. No, you are not. I'm old. And I think it's that attitude of say the younger generation coming outta Uni and wanting to work and I dunno if this is the same in the states, but I've had some sort of difference in expectations that I would work until midnight if my boss would expect that of me. Probably I was an idiot. Worked far too hard for other people before I worked for myself, but there seems to be an attitude shift in that younger generation was like, no, I don't do that. I don't want to do that. And it's like, how can you get through that? Basically, I think was her question. And she said it in a very diplomatic way.

Richard: Okay. I don't think that person would make it through the screen process to get to me, maybe.I will say one of our, we tend to hire folks later in their career at Fathom. Like we, one of our big goals is we wanna get to a hundred million in revenue with less than 150 employees.

And, a lot of that is we believe a lot in the power of small teams. [00:40:00] So we tend to hire people that are already 10, 15 years into their career, who really love work from home. I think more than folks early in their career. I think there's probably a selection bias happening here.

I think if you're earlier in your career, I actually think you should try to go to an office. That's where you're gonna build more kind of communal bonds and find people you might start companies with. If you're late in your career. You've got, a partner, you've got children,work from home sounds pretty good, cutting out that commute.

So I think 'cause of our philosophy on hiring and the selection bias you get from remote, we probably don't have as many folks early in the career. I actually, you know,from Sidal perspective worry a little bit about, I think if I was starting off now with,AI, like a lot of those entry level jobs have now been taken by AI, right?

And so there is a lot of opportunity for humans to do great work, but it usually is more experienced Folks, that,

Anyways, long wayed answer. I don't have a great answer to managing Gen Z folks, but yeah, we'll leave that open-ended.

Juliet: And what would your question be for our next guest?

Richard: I'm always just curious about, I think there's so many different ways to run a business, there's so many different ways to be [00:41:00] motivated. I think most people think most founders are motivated by money and fame, and I've actually found that's almost never true, or very rarely Sure. Also very strongly correlates with not the most successful people, right?

The most successful people I know, like they just love to build stuff. Same thing with me, so my question would be like, what is the thing that deep down, what is the thing you love to do that gets you out of bed in the morning?

Because for me it's honestly, we, we've got a channel internally Slack called Boom, or people like, we post people's positive comments about Fathom online or in our sports channel. That's what gets me outta bed in the morning. Maybe I'm too externally validation driven, but that's what feeds my Soul cup.

 Revenue number's great, but what I really love in Diamond, it's are people loving the product? And so I'd be curious to ask the next person what gets them out of bed because it might be something different. 

Juliet: I completely agree. And the money's a means to an end, but what is the end? What's the other thing that actually matters? No, I really love that question. Thank you so much and thank you, Rich, for your time today. It's been great having you on How to Start Up and best of luck with Fathom.

Richard: Thank you so much. It's been a lot of fun.

Juliet: 

[00:42:00] If you'd like to contact Richard, you can find all of his details in the show notes along with a recap of the advice he has so kindly shared.

And if you've enjoyed the episode, please rate and review it. No one ever does. We know you are listening and we would really love the feedback. So thank you in advance.