Wait, wait. Sam, you got to clickbait it up. How a $5 billion founder is using AI to change his life.
Well, are you a five? What are you guys worth now?
You know, you know, you you.
All right, we should do a little intro. So, first, do you... Is it... I've been saying Zapier my whole life. Is it Zapier or Zapier? Zapier makes you happier is the trick.
It reminds me... happier. Wow, dude. How is that not your slogan?
Here's your... here's what your slogan is on your site: 'The automation layer for Agentic AI.' It doesn't really flow the same way as 'Zapier makes you happier.'
It doesn't. We could... So there's two things from the name. One, we just couldn't afford the second P when we started; the domain was available. But the other, like we were too clever for our own good, is it has API in the name, which we always thought like, oh, that's neat, cuz it's all connecting all these APIs and all that good stuff.
I didn't even notice that.
Honestly didn't even notice it was missing.
But also if you played 'Zapier makes you happier' I would have read it as 'Savior makes you happier.'
Just gone on with your day.
But we're happy you're here.
The short story is you're this guy who's built this really cool... started off as a very simple idea where you can make different apps connect with each other. What does that really mean? Let's say I have a business and people submit a form over here on my form thing, but then I kind of need that form thing. I need all the people who submitted the form. I need that data to go over here in this other unrelated app. And you guys built those like universal plugs and you could just plug in anything to anything else. And it was just incredibly useful. Sam used it. I used it. We didn't realize, or I didn't realize, how monstrous of a business this had become. I believe you guys are worth several billions, maybe five billion was a number that I had seen at some point in time, and I think you had bootstrapped a very long way. I don't... is it purely bootstrap or you at some point took some money?
I think the word people call it now is seedstrap. So we did go through YC and we raised about a million and change, but that's all the primary capital we've ever taken.
Yeah. So just incredibly to raise a million bucks and then get to 5 billion in value is just pretty, pretty wild. So super cool. But you're here not to necessarily tell the whole story of how Zapier makes you happier, but to show us what you're doing with AI. Can you give me a public sales pitch real quick? I'm requesting a sales pitch. And what I want is, I know how OG V1 of Zapier worked where you connect one app to another. You're saying, 'Hey, we have this new AI product that you use with Claude presumably, makes your Claude or ChatGPT usage more powerful, whatever.'
I haven't been using that. I also, here's the other caveat, am kind of dumb. So, like, imagine talking to somebody with half the IQ of yours.
He's like, 'Can you give me a pitch to someone who's sort of dumb and doesn't listen,'
but they are smart at the same time. The confidence of a smart man.
I have really good news for you, though. Like I think AI is this incredible gift to people who are kind of dumb because you can just talk to it and say like, 'Hey, I didn't understand that answer, explain it to me like I'm five.' And so you can keep like going down the dumbness stack, I guess, for like lack of a better word, until you understand it and then go, 'Okay, now like teach me back up to where I need to go.' So, okay, the sales pitch, you want the sales pitch. In the old world, you know, Zapier, you know, come in, build integrations, connect apps. The new world has a lot of the same things, but it's different. Nowadays, you're hanging in Claude, you're hanging in, you know, Codex or Cursor or, you know, OpenClaw or whatever it is that you like using, right? And you're just talking to it going back and forth. Most people start with that. AI as a chatbot is valuable, but it's only generally smart. It knows everything that the internet knows. It doesn't know anything about your business. So the next step that you need to go to is to start connecting your tools. You start connecting your CRM, your help desk, your email, your team chat, all this stuff, and you feed that context into the AI. Now when you start asking it questions about My First Million and all this other stuff, you get not just generally smart answers, you get hyperspecific answers around how your business works.
Okay. So if you install Zapier into this, you can hook up any of the tools that we have, right? So that's the first thing. The first bit of value and the first thing that people sort of experience is like, oh, connecting stuff makes sense. But look, you can do native connectors and all that other stuff. So, you know, use Zapier, use the native connectors, do whatever you want. The next layer and where this gets really powerful is not where you're sitting having this like back and forth conversation, it's when you're actually deploying automations and agents to actually go operate for you while you sleep. Like that's where you start to get the real magic. And the reason you want to use Zapier to do that versus like any of these generic agent builders...
What was... what's an example?
I mean like, you know, Claude has routines or ChatGPT has scheduled stuff, like, you know, there's a bunch of these things that can kind of do agents or pseudo agents these days. But the nice thing about Zapier is that if you use us, we are going to make it so that whatever you build is optimized to run more deterministically than agentically. And I don't think people have caught up on this yet, but you've probably experienced it.
Climbing up the IQ scale.
I know. I know. Hang tight. Hang tight. Hang tight. Agents, agents, they kind of just guess what to do, right? And if they're smart enough, they guess right most of the time, but sometimes the agents guess wrong. And the second thing is agents are using tokens, right? We probably heard token maxing, right? And so they're really expensive. But if you use Zapier, we're going to write code for you and we're going to write workflow logic for you, which means that it's going to run like a machine runs. And so you're going to get... it's not going to burn tokens. It's going to be reliable and accurate and it's only going to use AI exactly where it needs to use AI.
And you've got this, um, you can host it on zapier.com. So you don't have to sit here and like keep your laptop open to run these agents in the background. And so you just get this like rock solid infrastructure that Zapier has always been great at and hookup of everything that we've always done. But you can build, you know, inside Claude, which is like really nice because Claude just kind of... the natural language style of building is, I don't know, it's just like way better than the no-code style stuff that we were doing in the past.
Well, you're not the Wolf of Wall Street, but that was a pretty good sales pitch. I gave you the 'sell me this pen' test and you did a pretty good job there, so I appreciate that. I feel like I now actually understand it. So that was good.
Well, and here's the magic. You can take, if you didn't and you're just pretending and being nice, you can take the transcript of this and you can plug it into Claude and say, 'Wade said this weird thing. I can't tell what this guy was talking about. Tell me what he was really saying.'
And that's like the cool part of AI.
You gave us a list of stuff. So we said, 'Wade, what are you making?' And it doesn't always have to be like this crazy game-changing stuff, but what are you making with AI? And you gave us a list of like three or four things. You said you have robots.
You got to clickbait it up.
How a five billion dollar founder is using AI to change his life.
Well, are you a five? What are you guys worth now?
I don't know how the markets behave right now. They're like up and down all sorts of ways.
Like, what would your guess be?
I don't think about it honestly. Like I'm just like, how do we build stuff that makes our customers successful?
I knew you were a sweet guy. I did think you were a liar though because I know you think about it. Podcast under oath. You know, there's the old saying, right? Like in the short term the markets are a voting machine. In the long term they're a weighing machine. I'd rather you just tell me f off than give me these lies cuz I know you think about it. But don't piss on my back and tell me it's rain. Just tell me it's none of my business and I'd accept it.
All right. It's none of your business, right? If that helps.
Since the last time you guys got value, where did the 5 billion number come from in the first place?
That was a secondary number sale that happened in 2021 maybe.
Okay. Are revenues slightly up, way up? Down, way down since then?
Okay. All right. Good. So, we're doing well. All right. The clickbait headline holds. Okay. Continue on. So, how you're actually using AI as a high functioning CEO.
Okay, let me share screen on some of this stuff. So, we'll get into the fun bits. One of the cool things is, you know, because I work with my agent all the time, I can say, 'Hey, I'm going on My First Million, like help me demo this stuff.' And so this is what it came up with. It's like, 'Here's the stuff you should actually go demo.' And so you can see I've got, you know, four potential things we can show off. One is the robot staff. This is kind of like my chief of staff that does like a whole bunch of different automations and agents for me. I've got the co-CEO CRM that I built that helps me stay really close with our enterprise customers. Um, I've got an AI that argues back. I think this is a really important thing that every leader needs to build. I think there's a lot of CEOs out there, a lot of leaders out there that are talking to an AI that just agrees with them all the time. And so you need to, um, like there's a few tips and tricks you can do to force it to scrutinize your thinking rather than just tell you, 'Yeah, you're totally right all the time.' And then I've got a bunch of skills that I run for a whole bunch of random tasks internally. So yeah, I would probably start with the robot staff. I think this is cool to dig into. The first thing, and I'll show the complete version of this, where I think most people start is kind of the morning brief. You know, it's super simple. It's a good way to experience like your first agent. You know, you hook it up to your calendar. You hook it up to your email. You hook it up to any meeting notes you've got, any prioritization, like a to-do list that you have. And then, you know, in this case, it runs at 6 a.m. in the morning. And then it sends me a Slack message.
Yeah. I call this waking up like the president. Like the president wakes up, gets a daily brief, here's what you have on your plate today. 9:30 you're talking to this person, 10 a.m. this person's here. Here's what they want to talk about. That sort of thing. And so totally waking up like... now everybody gets to wake up with a presidential briefing.
Exactly. And so, you know, these things run like in my case, I have it send me a Slack message. You could have it send you an email. I mean, heck, you could probably have it fax you something if you really want to feel like the president.
I have it set up so it prints, but it doesn't do it all the time. And I also have this woman named Ari, conveniently, because I told... I wanted to make it come from like a person and the person they made...
Is that weird for you that Sam named his agent after you?
The person they selected looked like Ari. So I was like, it's Ari. So I have Ari that sends me like a pump-up speech as well for the day. So I love these daily briefs.
There you go. Now, so the cool thing, I'll actually show you. So I have the link out to this one. This is, you know, this is running on Zapier and you can see what this actually looks like. You know, it kind of looks like, you know, classic Zapier a little bit. You know, runs every day at 6 a.m. It goes and fetches all the calendar events and then here it's writing some specific code and hitting the AI to write to Slack in a very particular way. Now, what is different about this is we can poke over here. This is all code which in the past like you couldn't do this with Zapier. Now, I've not actually looked at any of this code. Like this is actually the first time I popped this open and even taken a look at what this is. So you don't have to know what's going on here. Um, but again it kind of speaks to the value prop of Zapier is it's turning all of this stuff into a deterministic workflow and it's only using AI in the places that you really need to use AI, which means this runs way more reliably, way more consistently, and way more cost effectively than most agents.
But can I use this instead? Like the problem that I have with Claude or Perplexity or whatever is that my computer has to be open for a lot of these routines to run and it's a pain in the ass. Would doing it on Zapier just be like in lieu of using like OpenClaw or having a Mac Mini?
There you go. Exactly the point. You can just deploy these to Zapier and we run them in the cloud for you.
That sounds way better than using OpenClaw or having a Mac Mini.
So way, way simpler, right? You can look up here. It's also agent managed. So, one of the nice things is...
All right, hold on just a second. I know what you're doing. Wade, who started a company, grew it to a $5 billion valuation, he's telling you all these amazing AI prompts that he uses for his business. You're probably taking a lot of notes right now. That's exactly what I was doing. And so, what we decided to do was take all of those notes, all of the scripts that he's talking about, and we put it into a really easy document. So, you just copy and paste everything. So there's the prompts where he talks about AI arguing back with him, where he can stress test different decisions by having a seven-person AI executive committee kind of help him figure things out, and then your own life coach so you can have a better day. So if you want this exact AI workflow again, you can just copy and paste all this stuff. Just click the link in the description or scan the QR code. All right, back to Wade.
While it is running deterministically, if for whatever reason it breaks, like say an API is down or, you know, something doesn't work right, it falls back to having an AI try and troubleshoot for you. And so it'll automatically apply fixes, you know, when things don't work quite the right way. And so there's this really nice blend of traditional software and agentic software here that is just, you know, super well done by the team over here. So I definitely would say, hey, if you're struggling with your OpenClaws, if you're like keeping your laptop open all the time, you should definitely install Zapier into your agent or AI of choice and give this a try.
Okay. What else? So, you get the morning brief. That's pretty standard, I would say.
So, let's talk about... So, I think the morning brief is pretty standard, but the one that I really like is the evening brief in the scribe. So, you know, I've played with a couple versions of it. One is it runs after every meeting and then I have another version that runs at the end of the day. I actually think this is way better than the morning brief. And the reason why is you can actually have the AI help you do stuff. So, you know, it loops over all of my meeting notes from the day. So, I'm using Granola for all of my meetings. It loops over all of the to-dos that are still in my to-do list for the day. It loops over all of the outstanding emails that are in my inbox for the day and then it prompts me and says, 'Hey, how did the day go?' So, there's two key things that it does for me. One, when I say how the day went, I basically said, 'It was a good day for this reason' or 'It was a bad day for this reason.' So, now all that stuff is getting logged and it's learning how to tune my day to help me have more good days than bad days. But the second thing that it's doing is it's actually taking action items on stuff. So, it says, 'Hey, you know, you were in a meeting with Sam and Shaan yesterday, and they asked for an intro to your buddy over at this company. Do you want me to go ahead and make that intro for you?' Great. Please do that. Right? So, it's drafting those follow-up emails. It's drafting all these other things. And so, this is where AI starts to get really helpful is it's not just giving you a brief and sort of teeing it up for you to do. It's now saying, 'I'll just do that for you.' And this is where I think you start to really feel the power of AI is when you can actually put it to work. So I love Scribe, the sort of post-meeting follow-up or the end of the day wrap-up because it just... I don't know. I used to spend like two hours at the end of my day just like, you know, just trying to catch up on every little tick-tack thing that I sort of got roped into. And now that like two hours at the end of the day is down to like 15 minutes.
So you think this is legit saving you two hours a day almost?
Yes. Yes. And furthermore that like, you know, what's good about my day, what's bad about my day part, it just helps me stay focused on the things that matter. So both the AI is helping me with that, but I also will send... like I remember about 90 days after I started doing this, I said, hey, I had this thought. I was like, you know, it knows so much about like what makes me have good days and bad days. Can you like summarize like what is common about my good days and what is common about my bad days? It was like, 'Sure.' And it did all that stuff. And then I gave it to my assistant and I said, 'Hey, here's like... please make my days look like more good days versus bad days.'
What were your good days and what were your bad days?
So for me, like the single most important thing is I'm a morning person and so I do my best work in the morning, I'm most focused. Um, you know, once after lunch comes around, you know, I'm just not going to get anything serious done. And so the most important thing is I don't want a meeting until 11:00 a.m. And so you just give me that morning time where I can knock out the most critical things and the day is going to go great. Like I will win the rest of the day. But if I have to like, you know, if that time starts to get sucked into like firefighting mode, you can tell I'll just be grumpy at the end of the day where I'm just like, ugh, like I'm not being successful. And you'll notice this podcast started at 11 a.m.
Can you show us how to get the AI to challenge you instead of...
Oh, yeah. Let's do that. So, coming back, AI that argues back. So, this is pretty simple, right? I use Cursor as my daily agent. But if you're using Claude, if you're using ChatGPT, if you're using Gemini, they often have like a system prompt or a way that you can sort of tune how it talks to you. And so in Cursor I have a file that's the AGENTS.md. That's kind of what it uses everywhere. Claude has a CLAUDE.md. And basically I have a file that's like how to work with me. And you know, I have it be direct and honest. I need the truth my co-workers are afraid to tell me. I need you to challenge my assumptions. I need you to poke holes in my thinking. I need you to disagree when you genuinely disagree. And these things are like goal-seeking entities. And so if you tell it like, what I want from you is this, it's gonna say, 'Okay, well that's what Wade wants, so I'm gonna try and give him that goal.' But if you don't tell him this, if you don't make this configuration, it's going to just genuinely try and like please you because it thinks that's what you want. And so you do have to tell it, you know, you sort of guide it a little bit. So like here's an example of a live prompt I ran ahead of time for this. I said, 'Hey, I'm thinking we should kill our entire free tier this quarter because one of our competitors raised prices. Help me draft the announcement.' Right? So, this is like something that it should definitely say, 'Hey, press pause.' Like that's a bad idea. And so, when I do that, here's the pushback that it gives me. It's like, 'Hey, that's a big move off a pretty thin signal. Before you draft anything, I want three things on the table.' So, like, bring me the data, tell me like what the actual problem you're trying to solve is. Pricing announcements are wrong. Like, is there a smaller experiment you can run?
Yeah. It's just like helping you think through these problems. And then, you know, I have this other skill that I've made, which I call my war council skill. And so, this is like an even more specific version of this that I will often invoke. So, I could say, 'Hey, you know, I'm thinking about hiring this person. Here's all the notes that I've got on them from the interview panel. Like, help me figure out if this person's going to raise the bar or not.' And what it will do, the war council is... actually, let me poke back. I think I have the war council skill in here. Uh, yeah, here it is. Um, okay. It didn't actually pull the skill up. So, if you go to Zapier's GitHub, you can get a copy of the war council, but what this does is it spins up sub-agents, and these sub-agents take on personas. There are seven sub-agents that get spun up. So, seven personas. Some of them are standing members. So for example, I have the wartime COO is a standing member of the war council. There's the ruthless CFO is a standing member. There's a contrarian board member that are standing members of the war council. Then I leave four personas to be generated dynamically. So based on the prompt, it will decide who else do I need to get advice from.
That wartime COO, you created the persona. You described its characteristics or you trained it on like Travis Kalanick transcripts or something? Like what did you... how did you do that?
In my case, it's a generic version, but you could do whatever you want, right? If you wanted to have like...
What does generic version mean though? You just said you be...
I just said these are the traits of... I said this is the traits of a wartime COO and it like... I said actually what I did was said I want a wartime COO, you know, generate that persona and then I looked at what it created and just made some like...
Do you have like a pretty chill, generally happy, positive outlook? Like...
This is... you're seeing the internal competitor inside of me, right? Like this is the Kobe Bryant Mamba mentality like, you know, coming through. I like having this thought partner that is a little bit more like ruthless because that's not my default mode, right? I want to have that kind of checks and balance there. And as a CEO, you're not always going to have people who are comfortable, available, skilled in that way around you, right? Like this is who you would want on your team when you want them to spin up.
Quick stupid question. You said, 'I use Cursor for my AI.' I thought Cursor is a coding tool, but it sounds like you're using it to chat and make decisions. Am I just wrong about what Cursor is or...
Yeah, I mean Cursor, you can use Cursor, Claude Code, Codex. Like these are all ostensibly coding agents but they have this agent mode where you can just talk to it and the thing is it talks to it but because it can write code it can do stuff for you. And so the way I try and talk to people who are not engineers and say, hey, you should try using this, is don't think of it as a tool for engineers to build stuff. Think of it as like you're hiring Cursor or Claude Code or Codex to be the engineer you delegate things to. Where it's like you probably have engineers you've worked with. You like to tell that engineer go build this for me or go build that for me. Like that's why I use Cursor is because it goes and builds stuff for me, not because I'm using it to actually go build a bunch of stuff. I don't really write any code.
Wait, so your default when you just need to like talk to an AI or chat or ask a question, you go straight to Cursor. You don't do Chat or Claude or anything?
I mean, mostly, especially when I'm at my desktop, I'm mostly talking to Cursor most of the time. It has like different models that you can sort of toggle back and forth on, which I like. So, I can, you know, test a bunch of different models and things like that.
Are you a typer or a talker?
I do a lot of voice-to-text these days. So, I use this app, Monologue.
I don't know. I like it. Like, I've tried a few of these. I've tried like Whisper Flow and Super Whisper and Monologue. Um, I don't know. It just gets me. Um, like I don't know a better way to describe it. Quick side note, I think this is going to be like an interesting thing about software for the future. Like when it is so cheap to build software, I think you're going to see more software that has like a personality to it. You know, you think about like consumer goods. It's like how many places can you buy jeans from or how many places can you get, you know, whatever from? There's like not that many differences, but you choose like you might choose one brand and I might choose a different brand and it's just because like I like it better, but it functionally does the same thing. I think there's going to be more software that kind of feels that way.
Yeah, it's... Yeah. No, I mean I think that's a good point. It's like the Adidas and the Nike shoes do the exact same thing, but like, you know, or like t-shirt brands. Like do you remember... I was watching this Blink-182 documentary last night. Do you guys remember Hurley, Sean? You remember Hurley shirts? They were everywhere. And I was watching this show last night. I'm like, why did we love those shirts so much? Isn't that crazy that we got obsessed with it? It was just that stupid logo and why did that matter to us so much? And that's... I think that's a good point. That's how we're going to think about B2B, even B2B boring software.
Well, that's... I think it's a really fascinating point because it always was differentiated on either features or price, right? Those were kind of the two main axes of competition and then design got layered in and I feel like there was like a wave where better design, better UI/UX became a differentiator and design, you know, got a seat at the table. And then, but if you look at like other products, non-software, it'll be either kind of brand or like status, you know, like luxury, it's signaling something about you that doesn't really exist in software. And it sounds like what you're saying is there's going to be a new axis of competition which is personality. Uh, you know, who do I like the feel of when I interact with it versus which model has more parameters and, you know, a higher benchmark score on the evals is not going to be how people are going to choose these things because I'm already doing that. I think you're absolutely right.
Well, and I certainly feel that like, you know, as someone who plays around with a lot of the models, I'm starting to find myself have some preferences for like different tasks. Like if I'm asking it to actually go do an engineering task where I don't need to talk to it, there's models I like for that. You know, like Fable is fantastic. Like you're going to give it a really complex engineering task. I'm like, great, have Fable go take a swing at this. It's going to nail the thing. But man, I do not want to talk to Fable. Like it is not fun to talk to at all.
We all have co-workers like that. They crush it at some tasks, but they're not a great hang.
Yeah. Exactly. But then there's other models where you're like, 'Hey, if I want to sort of go back and forth and have like a discussion and a debate around like what we're doing.' I'm like, 'Okay, I feel like I can tolerate that discussion a lot better than like trying to talk to Fable where I'm just like, oh my god.'
Are all the phrases like when Claude talks to me, it's like what are its phrases? It's like, um...
It's not this, it's that.
Or here's the thing that's quietly really the problem.
Yes. The honest truth, the load-bearing, you know. Yeah, like here's... let's get to the spine.
Are those specific to AI or to the model?
To the model for sure. Like you can tell most models I think are derivative of either the Anthropic series or GPT series. And so like, you know, those two kind of have two distinct personalities. But if you go use any of the open source tools or things like that, you can kind of see like who they've been distilling a lot and like it sort of inherits the personality of the one above.
Dude, I've noticed that my co-workers are starting to say like the Claude... like someone the other day said like let's get to the load-bearing part or like something like that. Like when they're talking out loud, not in writing.
Yeah. Yeah. Yeah. I've noticed that too. I've noticed that where they start saying things I'm like, that's Claude. And it's sort of like...
Yeah. Like or they're trying to talk that way. So then they can like use Claude's writing and be like, 'No, it's really me.'
Who is training who, right? Are we training the AIs or are the AIs training us? And I think you can go a step further where actually...
Let me show you this other use case cuz this is interesting too. So this is a bunch of synthetic data. So, it's not real, but, you know, I have a real version of it. It's got a list of accounts, and you can kind of click into the accounts and see what's going on with it. So, if we click into, you know, Acme Logistics here, you can see a bunch of stuff about it. It's like, oh, you know, usage is down 34%, there's a renewal in 45 days. Here's all the apps they're using. Here's the people in the organization that you need to care about, etc. But then, like, here's the interesting part. It's got the suggested CEO outreach where it's like you should email Jordan and, you know, check in on him, see what's up. And so this is where, you know, I think AI gets really interesting because in some respects, who's the boss now? Like am I the boss or is the AI the boss? You know, every Sunday morning now I wake up with a list of 10 customers that I want to get in touch with and it's got proposals for what I should say and what I should do and why I should do it and all that sort of stuff. And I candidly love it. Like this is stuff that was really hard for me. Like I'm not naturally a person who's just like, you know, ultra networker like you guys are where I'm just like I got to keep in touch with everybody. And so now I've kind of put an agent in the loop and said like, 'Hey, help me be better at this stuff.' And yeah, I just literally wake up in my Sunday morning briefing. It says, 'Here's all the accounts you need. I've already drafted the emails in your inbox. You just need to review them and make edits or press send.' And I think there's like something to this where, you know, if you do a good job of building these tools up, they'll kind of start telling you what to do. And you do have to keep your brain on because if you turn your brain off, you'll probably sound dumb from time to time, but by and large, I think these things make pretty good bosses for certain use cases.
