Subscribe
Share
Share
Embed
AI First with Adam and Andy: Inspiring Business Leaders to Make AI First Moves is a dynamic podcast focused on the unprecedented potential of AI and how business leaders can harness it to transform their companies. Each episode dives into real-world examples of AI deployments, the "holy shit" moments where AI changes everything, and the steps leaders need to take to stay ahead. It’s bold, actionable, and emphasizes the exponential acceleration of AI, inspiring CEOs to make AI-first moves before they fall behind.
⁓ And part of that ended up being some of the things that we decided to do at Clay, because we originally started with the idea of how do we give the power of programming to more people. And through a windy set of decisions, ended up giving it to go-to-market people. And I think what we're best known for at Clay is actually defining this new role called go-to-market engineer that is the key thing that I think is one of the first AI native roles.
that ⁓ took people who are kind of like doing rev ops or growth marketing and gave them kind of a title for what they're doing, which is engineering systems to help kind of companies grow as fast as they can. And that's what Clay does. It just gives the power for go-to-market engineers to build systems that find you ways to grow your business as fast as possible.
Andy Sack (01:03)
This is AI First with Adam and Andy, the show that takes you straight to the front lines of AI in business. I'm Andy Sack and alongside my cohost, Adam Brotman, each episode we bring you candid conversations with business leaders, transforming their businesses with AI. No fluff, just real talk, actionable use cases and insights for you.
So excited for today's episode. We have a special guest, a good friend of mine, Kareem Amin, who is CEO of Clay, based in New York. And welcome Kareem.
Kareem (01:36)
Thanks. Great to be here, Andy and Adam.
Andy Sack (01:38)
Today's episode, we hope to bring you really like a frank conversation with the CEO of one of the fastest growing companies in New York City about how Clay is using AI in their product development and how Kareem in particular thinks about.
AI as a CEO and its implications to the future of work and the future of organizations. So with that, I'd really love to just get into the conversation. Let's start with basic introductions. I know you, Kareem, but the audience doesn't. Would you just give a brief background on you personally as well as Clay?
Kareem (02:17)
Yeah, sweet. And I'll add just one more thing, Andy. One of the fastest growing companies in the world. We'll just leave it at that. Yeah. In the world. Yeah, you know, we got to keep SF in there too. Okay, but just like the brief background is I'm originally from Egypt and then I moved to McGill to do a degree in electrical engineering and also studied physics.
Andy Sack (02:23)
Not New York, not New York, the world. I mean, thought New York was the world, Kareem.
Kareem (02:43)
And then I actually thought I was going to do grad school. I was going to do grad school in either physics or neuroscience. I neuroscience, also got a job at Microsoft and then decided to go work at Microsoft for a bit first know, there were a couple of things I needed to immigrate and then I was like, okay, I'll go to grad school after that. But I've always been programming and then with my good friend from college, Nikolai, we started a company.
We did Techstars actually with Andy in Seattle. And then fast forward, we sold that company. I worked for a little bit at the Wall Street Journal, which was kind of like a side quest where I was interested in kind of like news, but also interested in like giving tools to people who aren't programmers to build things. And ended up actually letting people build kind of these like mobile, data-driven stories.
⁓ And part of that ended up being some of the things that we decided to do at Clay, because we originally started with the idea of how do we give the power of programming to more people. And through a windy set of decisions, ended up giving it to go-to-market people. And I think what we're best known for at Clay is actually defining this new role called go-to-market engineer that is the key thing that I think is one of the first AI native roles.
that ⁓ took people who are kind of like doing rev ops or growth marketing and gave them kind of a title for what they're doing, which is engineering systems to help kind of companies grow as fast as they can. And that's what Clay does. It just gives the power for go-to-market engineers to build systems that find you ways to grow your business as fast as possible.
Andy Sack (04:16)
Great. Adam, you want to, do you have any questions you want to fire at Kareem to start us off?
Adam Brotman (04:21)
I sure. mean, so Kareem, let's start with your, your product. when did you start using AI either to build your product or do you incorporate AI into your product or both? And,
I'm not as familiar with your company as much as Andy is. Were you guys on the forefront of AI even before ChatGPT sort of big coming out moment at the end of 2022? kind of give us a little sense of your history with AI and the company's history with AI.
Kareem (04:37)
Mm.
Yeah.
Yeah, mean, no, the company was before LLMs were available.
more broadly, and we started in 2017, actually. So it's quite a while ago. But what was interesting was that the, what we built was, and the intention behind it was how do we give the power of programming to more people? We ended up deciding to give it to go-to-market teams, which are half of the company that have a lot of creative ideas of how to grow your business, but they can't implement them. And what we built was a kind of like, we were like, the spreadsheet's the world's most popular programming environment.
and people are building lists in it. So what if we just connected it to lots of API so you can build lists of companies and people, which is what salespeople are doing to decide whether to reach out to them, who are they, all that stuff. And so because of the way that we built the product and the intention around giving programming to more people, we were really well positioned when ChatGPT came out and APIs came out for GPT-3. And we were able to kind of incorporate
AI
as just another integration in our product and then we built one of the first, I think, agents and now one of the most popular agents in use and production, which is basically an account research tool. So it like helps you find more information about a business or a person. So that was kind of the history of like why we were well set up for it. We just moved really fast on it because we understood what it meant.
⁓ And because we were trying to do some of the things that were doable by LLMs, but now it's possible to do it better.
Andy Sack (06:22)
And when you say you moved really fast, how did you move fast? And how do you think more? I'm sorry?
Kareem (06:26)
Like same day. Same day.
Yeah, the same day that the API came out, was integrated into Clay
Andy Sack (06:32)
same day and sort of why, I mean, guess you saw it, you were just like, let's go with this integration or did you and your team know intuitively that AI was gonna be a significant component of your product?
Kareem (06:34)
Yeah.
No, we knew immediately. Because it's basically like, if you think about...
Andy Sack (06:47)
and why, why?
Kareem (06:50)
Yeah, if you think about what people were doing, were kind of, you like the way the world looked as like you have SDR in your company. Typically, if you're, if you have any kind of like sales team, you have SDRs, going to websites, they're researching it. They're asking questions like, does this website have SOC 2 compliance? Does it have a login button? Does it have like a book a demo button? And now you can answer all of that stuff with an LLM scraping the website. So it was immediately clear that this would replace a lot of this work.
Andy Sack (07:19)
there's been a lot of talk about AI's threat to SaaS companies and how really it undermined the underpinnings of all SaaS. Does Clay fall into that category? Why or why not?
Kareem (07:33)
This is a good, it's funny, I actually might talk about this before I'm talking to the company about it, because I'm writing something about this, but I'll give you the broad outlines. I think the way that I'm thinking about it is you've always had a build versus buy question, And you also have free open source software that you could have just used this there. So actually maybe you can start off by saying is all software going to be built?
And if the answer is yes, then that's one possibility. I don't think that's the case. And we can discuss why, but let's say that the answer is no. If the answer is no, then is your kind of software particularly susceptible to disruption by build? So on one end of that, you might say Stripe is not, right? Because there is risk, there's compliance, you're not gonna go do all of that stuff.
On the other end, maybe it's kind of like a scheduling calendar app, In which case, maybe you can vibe code that. Maybe it doesn't improve with network effects. There's nothing really that is specific about it being hosted that makes it better. And so you want to be on the other side where it's not kind of disruptable in this way. And then the other way to think about it that might be helpful is, well, let's say that you have, let's talk about the technology
technology as it is right now without kind of like fast forwarding into what if it's 10x better. It's probably going to get 10x cheaper. But you'll have to hire an engineer to actually build anything of reasonable sophistication and maybe they're much faster now.
But you're still paying them, let's say, $200,000 to build a software product over time that you could buy maybe for less. the easiest question is, you want to wake up at 3 AM and maintain that software? And how long is it going to take to build? So the only reason to build would be if you have very specific needs that are not found in the market or not customizable. So I do think it changes. I think the most likely scenario is there will be
way more things being built that were not worth being built. And then there's going to be a bunch of stuff that was being built that is commoditized and no longer really worth ⁓ kind of doing. And you can probably have an internal version pretty quick. I don't think we fall into that category for a couple of reasons. One is that we have all these deals with a bunch of different data providers. So you can get that data. So we have to go out and do all that. And then you do like what Clay owns long term.
really is who are your best customers and how do we find more like them. And if we know how that's happening across the market and how you found historically your best customers and what predicts that, that's hard and unnecessary for you to rebuild.
Andy Sack (10:04)
Do you think that AI has been bullish or bearish for your business?
Kareem (10:09)
Very bullish.
Andy Sack (10:10)
very bullish and and why is that
Kareem (10:11)
Yeah.
A lot of what we want to do is, so if you think about kind of like components of go to market, it's like, who are you targeting?
So you need really good information about that. And that's a lot of unstructured information, whether from phone calls, from recorded calls like this, or like sales calls, or Slack conversations, or websites. And LLMs are ridiculously good at turning unstructured information to structured information. So that's super helpful. And then it really depends on your own creativity. Like, are you asking the right questions? So what I mean by asking the right questions is let's say you put out a new product. Are you thinking about what
What
is the value of your product? Who could use it? What information could predict that they could use it? And then you try that campaign out. It's also very good at writing emails, website copy, ads. And it's very good about planning these whole campaigns. So in every way, I think it's allowing people to do more.
Andy Sack (11:06)
has your own professional work, like the way in which you work, how do you use AI?
Kareem (11:12)
This is where I think it's interesting because I've been talking to a lot of founders actually and a lot of people
And I would say I'm on the low end of my personal job being affected. So I think there's nothing that I've built myself actually to increase my productivity, but I'm using a lot of things that people have built. So we have our own, what we call internal sales agent where I can ask Clay, for example, who are our top customers who have used this feature, but I know are fast growing, but haven't used this new feature that we've released.
Andy Sack (11:45)
That's an internal tool that you use.
Kareem (11:47)
Yeah, and we'll eventually make that available to all of our customers. But I can ask a lot of questions about the business. How did we price this ⁓ enterprise plan for our customer that's similar to this? So those are things that other people have built that I'm using. And there's a lot of summarization tools that I've been using. So it's very easy to ask questions about calls that people have been doing, sales calls.
We've been using it in the business quite a bit in support. I would say most of the features that I'm using are clay features or summarization of calls and Slack channels. But I haven't done anything fancy to kind of like create my own reports or I was talking to a couple of CEOs who've built things that are like...
it coaching so them as a coach and they give feedback to all of the reports through their own agent. And he said that, you know, it's actually doing a better job than him in a lot of cases. And he only like occasionally corrects it or it looks at like new hires and flags them if they don't have kind of like the requisite AI skills. I think we're at a scale where like let's say we're like 350 people going to 600 this year.
My intuition right now is that I don't want that.
to like, I actually feel like even human summarization is not good enough for me. And I would rather be in some of these details. I don't want to kind of delegate that. And actually I think that's some of the unique value add that I can give. And I'm not trying to scale myself across the whole org. I'm trying to put myself in the places that have the highest leverage and trust my judgment around it. So
I'll give you an example. One time we did this pulse survey of the company.
and we were maybe 140 people. And when I looked at the summary, which was half AI, half humans, like I was like, there's this issue in this area, there's an issue in this area. And then when I looked at the actual responses, I understood that it was isolated to this like particular person. And I remember the context of why they have, you know, like instead of being like, people are frustrated that they're, you know, like how we do promotions. I'm like, I remember we messed it up with this one person and that's why it's there.
I just think that it's good to not let go of control. I'm not even sure if the CEO is the right person to get value out of this right now. But you know, people really disagree with me. Yeah.
Andy Sack (14:14)
So let's actually talk about that, is that all right?
And just to ground it, what LLM do you use? Do you use multiple LLMs? Do you have a preferred?
Kareem (14:23)
I use all, yeah. Like I use all of them
up. I don't actually because it's, I kind of like switch between things. because of also my role in this, I experiment with everything. So I'm playing with Codex, I'm playing with Claude Code, I'm playing with code and ChatGPT because I need to stay on top of like everything that's happening. And also both OpenAI and Anthropic are customers of ours.
And the thing, the thing that you wanted to dig into like the comment.
Andy Sack (14:52)
Yeah, so let's
dig into and one other just one last question. you using? Have you used co-work at all like something's half something happened. Something happened from Adam and my perspective.
Kareem (15:01)
Yeah.
Andy Sack (15:04)
in December, January timeframe. it's basically 2026 is clearly the era of is the year of the agents. And so I just want to curious, have you had the aha moment of using Claude co-work specifically in your as part of your own professional work?
Kareem (15:12)
Right.
Yes, think the things around like calendar management and actually sending things out and knowing who I'm talking to and a lot of that stuff is getting automated on our end. And I would say I'm like one step.
If there's kind of like the earliest adopters, I'm one step below that for the following reason, which is that normally I'd probably be on the cutting edge, but I don't have time for that. And so I'm just, I'm one step behind the people who are experimenting deeply.
And then I'm getting kind of the best of breed suggestions and then working on that. Like I have a bunch of people who are running open-claw clusters and doing a bunch of things and I'm waiting to see what productivity they get. And because part of it is that it's fun to play around with. ⁓
Andy Sack (16:03)
Yeah.
Kareem (16:04)
But when you're growing from like 300 to 600 people, a lot of my problems are not, you know, that I need to automate my calendar. It's, or I need to send out more emails. Like most of the emails that I get, I don't even respond to, which I hear is a really bad thing to do as a CEO. But most of them are not. Well, it's just like, they're not important. And in fact, my time management skill is drop everything that is not top priority and then apologize later.
Andy Sack (16:11)
Yep.
But that's the case.
Adam Brotman (16:20)
Thank
Kareem (16:31)
See
Adam Brotman (16:32)
lucky that you took the time to talk to us, Kareem. I actually understand your point of view in terms of like as a CEO, it's sometimes, it's not as obvious or easy to just see all the ways AI could, I'll call it, something you're doing. Because a lot of what you're doing is,
Kareem (16:35)
Yeah.
Adam Brotman (16:53)
your job is to communicate and synthesize and, and, and make decisions and like, you're not going to want to. Yeah. Yeah.
Kareem (16:59)
in my own side. Yeah. It's kind of like,
it's like, I think what you should be trying to do is look at the jobs to be done that you have that are repetitive and be able to take some of them out. But if you as the CEO have that.
That's a problem in the first place, right? Like my job is to ideally try to do the thing that uniquely I can do. And hopefully that's kind of like having a broad view synthesis of like what's happening in the company, communicate back to the company in my own style. And I hate reading AI written things, just to be clear.
Adam Brotman (17:32)
Yeah, mean that Andy can tell you I do too. now let's, let's qualify what you just said, Kareem. Uh, you're the guest. I don't want to contradict you, but there are couple of things. One, well, one is first of all, I actually think that you're pointing out two things that are worth pointing out I was actually just having this conversation earlier today with somebody like,
Kareem (17:42)
Yeah, please.
Andy Sack (17:42)
No, I want to, I want to, he's open
to the dialogue.
Adam Brotman (17:56)
we need to be careful, Andy and I do, but just society needs to be careful. They don't just throw the word AI out and assume that everyone understands it the same way and that we all mean the same things by it. And like there's, first of all, AI itself, the same AI system, let's call it, the same model with the same harness might produce very different results if you're prompting it differently, if you're giving it different, if you're not giving it the right context.
if you're not iterating with it. So by the way, it's rare that you one shot whatever you're trying to do with AI. there's an issue of like, what's that meme or that what the kids say now? It could be a skill issue, not on your part, but I'm saying like, when you get presented with AI slop to read and you're like, I don't want to read this. Like this is actually making me, I'm getting dumber, not smarter by reading this right now. Like, which I hate that. That's an AI slop thing. That sometimes that is,
Kareem (18:25)
Yeah. Right.
Adam Brotman (18:46)
because somebody just A, didn't know how to use the AI and B, didn't read their own work and weren't discerning. However, when you get someone who knows how to use AI, you can't even tell when they wrote. It's like they're just blended with AI. The AI is just making them better. And that's both upstream from whatever they wrote as well as in the writing itself. It could be like the research they did, the thinking they did, the actual output. And so,
Kareem (18:53)
All right.
Well, it was that.
Adam Brotman (19:15)
I'm so am both agreeing with you, but also kind of asterisk in what you're saying, like as in like, you know, yeah, somebody put something in front of me that is just AI garbage. You can just tell. And you're like, I don't get this out of here. And, and it may have nothing to do with like the model or whatever, but, there's other times where you can't tell. fact, that the goal should be if someone's using AI, right. You can't tell like, was, was there thinking augmented by AI? Was it writing a augmented by AI?
I don't know about you, Kareem I'm in the same boat. Like I'll find myself a lot of times jealous of other people with different jobs to be done than me where I'm like, oh my God, if I had that job, I would be having fun. Cause I love using AI. So I'd be having fun all day to your point with like, this date, would be mixing these two data sources over here and I would be like automating this thing over there. So I, but you're right. That's not your workflow as either as a CEO or just as a leader, you
You can't force it, but it's worth keeping in mind, though.
Kareem (20:10)
Boom.
Andy Sack (20:12)
Yeah, I'm going to take a little bit of a stronger position, in part because it's fun, in part because I've known Kareem longer. I mean, what I heard Kareem say, just to highlight, which you can't argue he's got one of the fastest growing companies in the world, not just New York. So based on results, Kareem, you're doing something right. And
Kareem (20:12)
Yeah, no, I agree with you. Go for it then.
Whoa.
Andy Sack (20:33)
And I really like your somewhat contrarian, particularly for being in the tech space point of view that you're at least thinking about what is the unique role that Kareem brings as CEO of Clay and you're going, okay, like I want to be in the details and ⁓ my job is to communicate and I'm going to do it and AI can't do it. The thing that I would say to you is, is I think that
If you took a stronger approach just of being of being more AI first, the argument I would make to you is that it's not about automating the tasks of the calendar tasks. It's actually about seeing things both in the organization and about your leadership style and about your business that there's no other way to see. And AI it's only when CEOs lead their organizations
Kareem (21:18)
Yeah.
Andy Sack (21:22)
and their companies to be more AI first, where all transcripts are being brought in internal and otherwise, that both the speed of iteration and the insight of the organization, which is driven by the CEO, it's only then that you become a more strategic organism.
And that's and I would say you become smarter when when Kareem actually partners with AI to sit on top of the context of clay's data, I'd be like, ⁓ that's a better spot to be even with your skepticism. ⁓ What's your what's what's your reaction to that?
Kareem (21:56)
Yeah, well, I just want to clarify
this. It's not skepticism. I think of it as like, and it's partially probably a reaction to the unbridled kind of like...
Andy Sack (22:09)
Yeah.
Kareem (22:10)
utopianism that I think is happening because I'm like I think people are saying one thing and then kind of doing something different which was a bit bizarre to me and just you know like it's like
Adam Brotman (22:20)
What are they saying?
What are they doing?
Kareem (22:21)
Well, it's just like saying, Hey, AI is automating everything, like it's, it's, you know, or there's a lot of people trying to do that in the org. Like, for example, I can tell you, there's a lot of companies that sell, they're an AI SDR, right. Or an AI account executive, and it doesn't work right now. Hey, maybe someone will listen to this podcast a few months from now and be like, wow, it does work and you didn't know it. And now we have it working. But ⁓ even automating software engineering, which is, I think one of the things that will be
happening
first is not like turnkey, right? It's not like every software engineering work. Like when I'm in meetings with other CEOs, we're talking about like how much productivity are you getting? Is it 20, 25 %? How are you proving that? It's not like I can go now and type something in as a feature in clay and then it magically appears in clay. Maybe you can do that when you're building something completely new or when you're kind of like building a simple product, but definitely there has
has been an increase. And actually the way that I think about it, let's just go back to kind of like how I'm thinking about it for the company. It's like, we should try to automate software engineering. LMs have been trained on coding and are excellent at coding. And if you can turn any problem into a coding problem, LMs are amazing at this, right? So the more you can do that, better. I think LMs will drastically speed up sales by having kind of like all the contacts from all the calls and all
you should be able to analyze kind of all the product usage. You shouldn't be able to analyze what people said they wanted in the future and feed that back into the back loop. So I do think all of those things need to be captured and used to inform your product strategy and your go-to-market strategy. And those are the two top things. Like when I'm talking to other people, they're like, we're trying to decrease costs in other areas. But I'm like, why do like...
Why are we trying to decrease costs? I'm trying to build more things and sell more things faster. And decreasing costs, there's a limit to how many costs you can decrease or having fewer people. The only advantage of having fewer people is moving faster, making decisions faster, like low coordination costs. But if I can get more done, that's what I'm going to do. So those are the two areas that I'm focused on as a CEO. In terms of just the decision making, Andy, you were talking about, it's tricky.
like what I'm thinking about, like being in the field, right? I'm like, okay, should I, for example, we have a bunch of agents that I can ask information, you if I want to get kind of like new data, like, hey, what's our conversion rate? How has it changed since we've made our pricing change? Like, is it gonna go up or down if I make these tweaks, right? That's the kind of things that I'm actually interested in, right? Now I could go and try to talk to the agent and figure that out myself.
And I tend to like to do things myself, but I could also talk to our head of finance and have him talk to our data scientists or do it himself, right? With agents as well. So I'm still not sure if it's like, go out and do that. We're talking about like my specific role, or if it's going to empower some of the other people. now what I want to do actually is ask more questions per minute. Right. like I have a lot of questions on a lot of hypotheticals and similar to when I was talking
to my counsel, for example, for me, good lawyer is number one, if you can answer all of my questions quickly enough and model all the scenarios. Ideally, you actually come to me with, here are all the hypothetical scenarios. I've already thought of them for you, and here's what can happen. But it's hard to do that without having the full context in my mind. So I can imagine that LLMs could be really good at that. What would be amazing is if I start asking all of these questions and these hypotheticals, and then I can get back
well, hey, you didn't think about this scenario. That's it truly kind of like blow my mind. And I do think there's a potential of that. as we get...
Andy Sack (26:06)
I think that's
available today if you pushed on it. It would take you leaning in going, I'm going to use this as a strategic tool to see blind spots either in my organization, my leadership team, or my decisions. And ⁓ I'm going to provide it with
Kareem (26:12)
Yeah.
Yeah, I mean,
truly people are telling me this and I think maybe, but ⁓ you know, like the other version of it is like using it in.
It's kind like people are using it in like therapy sessions or like coaching kind of like things. And I think that is somewhat helpful. But it is a trade off on like, would I rather do that there or would I rather spend my time talking to people? And I'm actually choosing to spend the time talking to people. ⁓
Andy Sack (26:50)
Yeah, valid.
When you're ready to make the transition and you've heard it from enough people, give us a call. We'll help you out.
Kareem (26:59)
Yeah, yeah. No, I'm excited. And also, you know, I'm just giving you some drama, right? So because we're like an AI, we're an AI first company. So we got a, you can have like a headline, the AI CEO that's not using AI for his decision making.
Andy Sack (27:04)
No, we need some...
Adam Brotman (27:13)
No, was actually talking, mean, Kareem, it's worth mentioning on the podcast here. Like I was actually just in a conversation earlier today with a business leader. it was the first time I'd ever talked to him about AI and he was at one of our clients, but was not one of our normal people we talked to. And he was, he was definitely skeptical of AI. Like he was.
He was skeptical of it at a visceral level, not around like, there's too many people that hype it. And that's not what you're saying. He was saying, I you actually use AI in your product. You're very aware of AI. And you're consciously making an educated decision about how you want to lead your company and how you want to spend your time. And you're actually making fair points about
like some of the trade-offs with AI and how you're choosing to sort of be on that and whatever. And I actually, it was interesting because it's just a good reminder that for Andy and I anyways, we honestly, we're so like hyper aware of like all the ways that you can use AI to.
make better decisions faster, not just to save time and automate things and reduce costs, but also to amplify your abilities at any position in some ways. But that being said, it's a good reminder sometimes, not just because of what you're saying, but it's just a good reminder to Andy and I too, and to anyone who's listening that
people shouldn't just take this attitude as like everything should be put through AI no matter what. And it's always going to make the situation better. I actually, it's not, in fact, it's not what we teach. When we talk to clients, we're usually like, look, you have to understand what your business objective is. And if you're stuck somewhere
You should be enlightened as the art of the possibility of what AI can do. And there are going to be a lot of times in the course of your day where you're like, I could just have AI do that, or I should be bringing AI to the table with this. And the truth is, in the course of an average day of almost anybody, that's going to be several times a day. However, that doesn't mean you have to do it several times a day,
Kareem (29:14)
Okay.
Adam Brotman (29:18)
And I think there's a danger if people are like the opposite side of the spectrum is that you're just doing it to doing it. Because if you do that, You're going to be worse off. I read a study that came out yesterday that said people actually get fatigued sometimes when they like read AI written stuff, because the AI doesn't just get to the point.
Right. And it's like, it makes you just want to let it's, it's not just always AI slop either. It could be like good AI output, but like it becomes fatiguing that I have to like read this long thing to just get the short answer. So I think we're all learning like how to, when to use it and how to use it and cut through it. So I, you know, I appreciate your perspective, man.
Kareem (29:59)
Yeah,
and also staying open to things as they improve, Like I'm incorporating new things as they get better. But there's like...
lots of little ways in which things improve my workflow. For example, I read a poem usually on our all hands every week. And sometimes I just screenshot it from a book I'm reading. And now I can just turn it into text and put it into my slide deck in one moment. So that provides joy for me. I think on the side of thinking through
Andy Sack (30:30)
For all those
listening, there was a great tip there, was to have a poem a day at the all hands meeting and you might increase your birth rate.
Kareem (30:39)
Yeah, people love that.
Probably 60 % of people love it and 40 % are like, what's happening? But I think, but when I stopped doing it, people wanted it back. And the thing that I would also say is that,
Andy Sack (30:44)
Hahaha!
Kareem (30:52)
Any tool, I mean, you were talking about like, you know, how it can help you write better or I think for me, the way I'm approaching it at least so that we have ⁓ like a unique perspective is I'm writing, right? And I want to be able to, if it can help me think through scenarios faster and other times actually where I found LMS really helpful in scenario modeling is after I've like,
Sometimes I'm keeping in Ram, like all of the ideas, but I, if I can have a conversation with an LLM and then have it produce, for example, a chart that shows like all the possibilities, that's tremendously valuable. So I think all of these like tools are super helpful, in getting kind of like my ideas and refining them and testing them against kind of like possibilities of it. Not working. it's almost like the voice that I have in my head that I'm discussing things with. Now I have many.
of them. And to me, agents and the way that we're building agents at Clay is how do we let you do something that's differentiated? you know, some things should be automated and some things are creative, right? In sales, you actually need to be doing something differentiated, which is why I think you need to constantly find the edge. We call it like the go to market alpha.
⁓ and then it should self learn. So I think that's what's different about an agent versus, usually I think what people mean when they're saying AI, they're like, Hey, it has memory. can improve as you kind of like give it more information. and the fact that it can write code for problems is tremendously helpful if you know how to break problems into kind of things that software can do.
you know, like one time I was trying to figure out there was a really long list of things and I needed to know kind of the occurrence of like a certain concept, how many times someone talked about AI in this particular list. And I was like, maybe I can write software for that. Or I just had Claude code write that in minutes. So that to me is really, really powerful.
Andy Sack (32:49)
Awesome. Kareem, I'm conscious of time. So I'm going to turn it to Adam and ask him. We just had a conversation with Kareem Amin, CEO of Clay and what jumped out? What's your takeaway that you want to highlight for our audience?
Adam Brotman (33:04)
I mean, appreciate it. Kareem comes to the table as somebody who definitely understands the power of AI and was early to adopt it in his company's products to give that product alpha and an edge and to just be a better product overall. And when we turn the conversation to like, what are you doing internally for yourself or with your company?
We didn't really get into the company stuff, but Kareem himself was saying, look, I understand the power of this stuff, and I was the first to put it into my company's products. But I want to personally have a style of leadership that gets closer to the facts and closer to the people and closer to the problems I'm trying to solve. And I really respect that. it's not coming from a position of
bias against LLMs or misunderstanding of them, but it's a personal decision that For me, my takeaway is just a good reminder of that's a really important part of actually being proficient with AI is knowing when to use it and how to use it. And we usually focus more or less on what Kareem said and more on like, have to understand where it can make mistakes and you got to understand not to produce AI slop and et cetera.
Kareem's coming at it from a similar perspective and saying, I'm actually choosing my own leadership style and not to just like automatically enhance myself through AI. I, you know, that was a kind of a cool perspective to hear.
Andy Sack (34:19)
Yeah, I mean, I'll echo that and also take advantage of the fact that I've known Kareem now over 15 years. Adam, you and I speak often about how AI and agents now in 2026 are not sufficient to winning the market. Here's Kareem winning the market.
Let's assume they didn't use AI at all. Clearly, they're doing something right in the market, and it's being rewarded because there's product market fit. And the company is one of the fastest growing in the world. And A, Karim clearly is smart and knowledgeable about AI, and he's making a decision to wait until the technology is further along and use it strategically.
I would continue to contend that he's under utilizing it, that it's he's under utilizing it as CEO and he's under utilizing it on behalf of clay. But the results speak volumes. He doesn't he doesn't need to it. The company's plowing ahead. So we'll I'm sure.
Kareem (35:16)
Yeah, most likely. ⁓
Andy Sack (35:18)
I'm sure
Kareem will continue to have the conversation. And the reason that I would say that you're under utilizing it just to make the point again is, is that I actually think that there are things that with the right content, and it only happens when you as CEO, you red pill when you dive in further than you have today. Still skeptical.
but you dive in further and you're like, shit, man, like, and it gives you access and an interface to insights for you to make better decisions faster about your company, your executive team, your people, as well as feedback to yourself. There's, there's, there's feedback that you, Kareem are ignoring because you are not being an AI for CEO. You're ignoring about your leadership style.
your decisions, the way you communicate. And the only thing that's going to give you that is Claude or OpenAI because everyone on your management team is playing some form of politics and AI doesn't. AI doesn't, it gives it to you straight.
Kareem (36:13)
You're most likely
right, but we'll see.
Andy Sack (36:18)
Yeah, I mean, but and your results speak for themselves. So I'm pretty sure I'm right about what I just said. And kudos to you to being somewhat contrarian and like making a strategic decision to be a fast follower. And I think it's I know I know that's what you do. Kareem, thank you so much. With that, I want to thank the audience.
Kareem (36:22)
Mm-hmm.
It's just not that.
That's what I do, Andy.
Andy Sack (36:43)
For listening to AI First with Adam and Andy, for more resources on how to become AI First, you can visit our website, forum3.com, download case studies, research briefings, executive summaries, and join our email list. We invite you also to connect with our AI First community, a curated hub and network for leaders turning AI hype into action. We truly believe you can't over-invest in your AI learning. Onward.