This AI Agent prospects on your behalf and books meetings while you sleep

Hayk is the Founder of AgentGrow, an AI-powered SDR tool that automates LinkedIn prospecting while increasing meetings booked by 35%. We discuss how the platform's unique approach of commenting before connecting leads to 2.5X higher acceptance rates.

The conversation explores AgentGrow's browser control agent that works hands-free to navigate LinkedIn, conduct research, and qualify prospects before booking meetings. Hayk explains their free-forever tier and why domain expertise is crucial for effective AI prompts.

We also dive into Hayk's passionate defense of sales professionals, the future of AI in sales, and strategies for AI startups to avoid being cannibalized by model providers. Whether you're carrying a quota or building AI tools, this discussion offers practical insights on the evolving sales automation landscape.

TimeStamps to Jump in:

• (00:00) Intro and AgentGrow overview'

• (03:39) Free forever business model

• (07:19) LinkedIn safety and automation

• (10:07) Live demo of browser control agent

• (13:51) The science of effective LinkedIn outreach

• (16:41) Enterprise pricing and scaling

• (21:59) AI's evolution toward passing the Turing test

• (26:12) Strategic approach to research and messaging

• (29:00) Results: 35% increase in meetings booked

• (32:10) Human-in-the-loop collaboration

• (35:40) Making salespeople's lives better

• (39:40) The AI landscape and competition

• (42:55) Fundraising and future outlook

• (45:14) Wrap-up

Referenced:

• Check out AgentGrow

• Try AgentGrow’s free forever tier

Find Hayk on LinkedIn

Transcripts

Find interview & all prior episodes here.

Hayk [00:00:00]:

Those people could never survive doing a sales role. The people who publicly roast salespeople are like, oh, they cold called me while I was in a meeting. Come on. Like, we have a million dollar annual quota. We have families, we have mouths to feed. And by the way, any deal we ever close is closed consentingly by the other party. It's not coercive, right? So they are happy we are solving a business problem for them.

Benjamin Smokovich [00:00:31]:

Great guest we have on today. Welcome to the AI Home Base podcast. And really what we're trying to do here is just decide, hey, are you on the chopping block or not? Is, is, is your, is your technology going to take over human, human jobs or is AI on the chopping block? So let's start with that, man. What do you think? Or do you. Is your technology that you're going to talk about today going to replace humans or is it going to enhance human intelligence? Or are, you know, or is your company going to be on the chopping block, buddy?

Hayk [00:01:09]:

Arch, we are staunch opponents of genocide. We don't want to get rid of humans. I think humans are the most important thing. And so, no, our application is an AI sdr. It's AgentGrow. It does the prospecting for you. It's not replacing SDRs. If I was an SDR today and I had the ability to sign up, I'd sign up, I'd put my feet up, I'd focus on something more strategic, I'd do account research. And so I think we're really supplementing humans and human productivity, not getting rid of them.

Benjamin Smokovich [00:01:42]:

Now, what aspect of the SDR does it handle? I mean, does it help you with targeting?

Hayk [00:01:48]:

So it helps you with deep research, not with targeting. You Pass us a LinkedIn sales nav URL. So you've probably done your targeting. I think best of breed method for targeting today is Clay. Like waterfall enrichment tools, Clay has competitors, I'm sorry, to just name them. And then you'll just pass that over to us. And then we'll do the actual outreach, we'll do the social warming, we'll comment with people because we found that there's a 2 and a half x higher conversion rate for LinkedIn acceptance if you comment prior to sending a direct message. We'll do discovery in the DM so you don't just book meetings with God knows anyone that's completely unqualified. And then we'll actually send your calendar booking URL, so we'll do the entire thing end to end, automated for you. And then we'll integrate more platforms in the future. So Reddit x email, things like that. But we're really starting on LinkedIn because it's just the obvious place to start.

David Stepania [00:02:41]:

Dude, the thing is, what's interesting and why I wanted to talk to you, there's this guy and he just, he's commenting on everyone's posts and like he produces some content, but his account is growing so fast just from commenting and people don't understand how important that is. Like, if you just comment on people's stuff and if any of those comments like kind of end up in like, you know, most like comments, like, you'll be golden.

Hayk [00:03:11]:

Exactly right. I always tell people it's like being the top comment on a Mr. Beast video. The top comment on a Mr. Beast video on YouTube. You're really famous. And so, yeah, commenting is underappreciated.

Benjamin Smokovich [00:03:23]:

So.

David Stepania [00:03:23]:

Okay, tell me that. Like, so, so there's so much hate about commenting on LinkedIn in terms of like, hey, all this AI generated comments need to go. Like, can we see your product? Like what? Like, can we see what it does? And so we can show it to people and good.

Hayk [00:03:39]:

I will show you guys a demo and you can sign up for free without talking to sales. And it's free [email protected] no, free, free and free forever. Okay. We have paid tiers and we have enterprise paying customers. We have six figures in revenue, but we offer free forever. So if you're a sales rep, you don't need your CRO's permission to go and get yourself some more pipeline. You know, you can go, you can self serve. You can actually use this. No other AISDR offers that functionality today. You don't have to sit through a sales call. I don't have to ask you 30 questions about your budget and your buying process. You can just go try it if you like.

David Stepania [00:04:18]:

So, Benjamin, I was telling hike just now earlier that a few years ago I used an AI tool and by the way, Hike, you can prepare the screen share or whatever, but I was using this AI tool to do outbound email stuff and so it would write up its own templates and I kid you not, one of them called like a big time CEO homeless, because he had a description in his LinkedIn that said he volunteered at like one of the like homeless centers or whatever. And so the CEO responded to me and he's like, where the. Did you see this? It's not often you get a response from someone like that. So that, that was a good thing. But like the.

Hayk [00:05:00]:

I, I promise. Yeah, Maybe that works, right? Pattern interrupt is the, I think, the jargon of the day, but I promise we won't, we won't, you know, offend any Fortune 500 CEOs on your behalf. Maybe we should build that feature. Maybe we should have a ludicrous mode. You know, we're, we're testing out the Grok 3 API currently, and they have a, you know, they have their ridiculous vulgar mode, and so maybe, if you.

Benjamin Smokovich [00:05:24]:

Want, that's an easy recovery, though, man. Oh, I'm dyslexic. I read that wrong. Sorry. But while we're talking, we have many.

Hayk [00:05:33]:

Such examples and, and our unsophisticated users are just like, oh, you ruined it. And our sophisticated users just always are really smooth and save the day and turn it into an opportunity. So.

David Stepania [00:05:47]:

Okay, okay. I got big hopes for this product. So, yeah, I would love to see a demo. Like, when I kind of glanced at it last time, I thought it was amazing, but, yeah, show us what you got.

Hayk [00:05:58]:

I'm surprised you haven't signed up and tried it prior to this call, David.

David Stepania [00:06:02]:

I, I am, I'm, I'm like, I'm. The thing is, like, I'm, I'm keeping a list of.

Hayk [00:06:08]:

I'm a little bit disappointed, to be honest.

David Stepania [00:06:11]:

I'm keeping a list. Like, first of all, like, it's not going to be me. It's going to be like, one of the, some, some of the people that work at our company. I'm, I'm telling them, like, hey, we need to implement this. And so we have a long list of things to implement to modernize our company, to be AI powered, and you're on the list. But I wanted to have, like, you know, see a demo today, and then I, I have a question, like, and this is an important one, I'm afraid to use it on my account because, like, my whole company depends on us having our LinkedIn account not restricted. Because, like, we own a recruiting company and we have recruiting licenses, this and that. And I have gotten banned from LinkedIn for, like, 30 days once. It was crazy. I have no idea why they said I use tools. You know, I did not. I really have not used any AI tools in years. There used to be some things installed that I tried, but then I stopped just because it's such a big risk. So to answer your question, I need to do, like, a deep review to see if I can use this tool because otherwise, like, we could go out of business. So. Would love your feedback on that.

Hayk [00:07:19]:

I can respect that. We have not had a single user ever be warned or banned by LinkedIn to date. That doesn't mean that it could never happen. But there's actually some interesting stuff in the news, which is that Apollo got banned, but that's, it's actually clickbait. So if you go install the Apollo extension right now, you can use Apollo on LinkedIn still, it was just their company page that got taken down. And so Apollo, the technology still completely works on LinkedIn. I think it's LinkedIn just kind of sending a message and saying like, hey, maybe please don't scrape our entire database. If you look at how many users Apollo has tens of Thousands, each running 20 to 30 prospecting queries a few days a week quickly adds up to all of LinkedIn's monthly active users. They've got about a billion users total and some fraction of monthly active. So Apollo probably has, if you think about LinkedIn's moat, gigantic business. It's data huge part of its competitive advantage. Apollo probably has a mirror image of LinkedIn's moat just hanging out in an S3 bucket in the cloud somewhere.

David Stepania [00:08:26]:

Oh yeah, everyone's ripping off LinkedIn. Like even I'm looking to get some, some of the data, right? And there's all these data providers out there and there's all these startups popping out that are saying like, hey, we have a 700 million databases. Like, where'd you get the database? The only place you can get it from is LinkedIn.

Hayk [00:08:49]:

If you look at these model providers, OpenAI and all these other people, their, their level of copyright infringement is the most historic Guinness record book breaking amount of copyright infringement in all of history. Right? All of these models are trained on. I know you didn't invite me here for my hot takes on other people's businesses like LinkedIn and Apollo and things like that, but hopefully it answers your question, like our users haven't had any trouble yet.

David Stepania [00:09:17]:

Yeah. All right, well, let's see it.

Hayk [00:09:19]:

But let's go. I, that reminds me, my, my team, my team, we were celebrating some wins recently and I asked them and they're like, many of them are, are speak Russian. And I was asking them, like, how do, how do I get you hyped up in Russian? I don't speak Russian, I'm Armenian. And they're like, yeah, we don't really have good hype. Words, phrases. Most of it, all of it is just cursing. And so the second video of Elizabeth Olsen, she's from the Marvel movies, cursing in Russian on Conan O'Brien. And I watched and I practiced it for them. And that's become our. I won't curse in Russian because I know this is a. A pg.

Benjamin Smokovich [00:10:01]:

So you're getting them hyped up by cursing in Russian.

Hayk [00:10:07]:

I can't. That's right. That's right. Okay, so here's a software demo of AgentGrow for you, David. You had some keywords for me, right?

David Stepania [00:10:16]:

AI agents.

Hayk [00:10:17]:

Yeah, that's right. So you can just quickly also determine the settings of what you want it to do on LinkedIn. And this is our Chrome extension that you can try for [email protected] we also have a fully managed cloud version coming out soon if you don't want to fuss with the Chrome extension. And what you'll notice is, am I an icon in this recording as well? Can you see me or my face or just my screen?

David Stepania [00:10:46]:

Your face and the screen.

Hayk [00:10:49]:

Can you see my hands right now?

David Stepania [00:10:50]:

Yeah, I can see your hands.

Hayk [00:10:52]:

So you can see that I'm not touching the keyboard or doing any of this browsing.

Benjamin Smokovich [00:10:55]:

That's right.

Hayk [00:10:57]:

Great. That's because we've built the browser control agent and it navigates. It will target people that are talking about AI agents. It'll even comment on your behalf. But don't be fooled. The goal of our application is to book you meetings. Remember, it will do the discovery. It'll direct message people. It'll send them your calendar link if they're qualified. The reason we start with comments is because, number one, LinkedIn limits the amount of connection requests you can send to about 20 per day. You all know this. If you're sophisticated about LinkedIn prospecting, and what we've observed is that if you comment prior to sending a connection request with someone and they reply to you, you're two and a half times more likely to get that connection request accepted. So, David, to your earlier point about comments and like, why they're powerful and why they're useful, we've observed that statistically your connection acceptance rate is much, much higher with comments than without. So this is a little basic taste of the application. We also support Sales Navigator for more sophisticated targeting. I think it's one of the first questions we get is like, whoa, this is cool. You comment, you DM people, you qualify them, you book meetings. For me, how do I make sure it's the right people? And so we support Sales Navigator as well. So very cool. If you've never used Sales Navigator before, go and buy it, because if you don't have it, you're probably not very serious about prospecting on LinkedIn maybe prospecting at all. And you can build up who it is that you're targeting. You know, are they mid market sized companies that, you know, employees there that work in operations that are very senior at the business, you know, at least director or vp and they're based in San Francisco. Great. You can build out this list, grab these parameter this URL in the browser bar, paste that into AgentGrow here and press start and it'll run against that list as well. Now the most sophisticated is that Sales Navigator will allow you to integrate with your CRM. So if you already have a targeting source of truth being built by your Rev Ops team or your go to market engineer or whatever this role is called nowadays, you can rely on that source of truth as well. So I'll just take my hands off once again, we'll watch the magic happen.

Benjamin Smokovich [00:13:43]:

Dirty fingers off your mouse.

David Stepania [00:13:45]:

Yeah.

Hayk [00:13:51]:

We'Re going out, we're going out to the operations professionals in San Francisco now and we're again, you start with commenting. It has a two and a half times higher acceptance rate than just sending a cold DM to people, which is, which is kind of the standard for LinkedIn prospecting. It's just like cold DM pitch slap people and if you're a little sophisticated, you just, you don't send them any message and then you pitch slap them three days later. It's like it's not actually impressive, you know, whereas this is not only doing the labor for you, but it's personalizing it for every single person.

Benjamin Smokovich [00:14:24]:

You obviously trust this thing enough to allow it to send messages on your behalf. So that's cool because back to David's story about calling a CEO a bum. What a comeback story. What a comeback story. You know, to me that's pretty cool, right? Because it means you trust it. That's a big deal. Means your stuff works.

Hayk [00:14:48]:

This application is for the early adopters. I had an iPhone one. I wore a Patagonia jacket before. It was the thing to do. As a tech bro, you know, I was a, I was one of the first owners of a Tesla in California. So you have to be an early adopter to be using AI agents in your actual workflow. You have to be someone who believes in taking a little bit of risk. Our users give us all kinds of requests and they vary from I love this application. This is great. How do I send more comments and messages with it to how can I prompt it and do all the prompting myself? Now if you're familiar with how difficult prompt engineering is. Yeah, like we do it full time. We have an eval pipeline set up. It's core to our IP at the company. Then you know that the average user picking up a new tool has no business having prompted access to your tool. They have no business. And so we get people all along the spectrum. But if you're risk loving, you tend to be on the right side of the spectrum, which is like, let's let it run, let's see how it goes. Most of those people who let it run remain very active users, which is heartening for us. That's what we want to see. If you run this thing for at least four hours, you're going to see some warm leads come in. You're going to have that eureka moment. That's our magic metric. Facebook's was eight friends connected with you. Our magic metric is run this for four hours, let the user have a eureka moment of getting some warm inbound leads prospected for them and then they tend to stick around.

Benjamin Smokovich [00:16:27]:

So at what point let's say I start passing this out to all my SDR buddies at SAP 122,000 employees, right? At what point do I need some sort of higher tier license?

Hayk [00:16:41]:

That is when you would need a higher tier license. So you can visit our website and what you'll find is we have a free forever tier. Like I mentioned earlier, we're big proponents of that existing of not having to talk to sales to use an AI agent and build your pipeline as a salesperson. We have a cloud managed agent. So if you don't want to fuss with the browser extension, you just want to run for you. But then if you want advanced features like brand governance, CRM integration, talk to multifactor, Authentic, all of the classic enterprising things, then you should have a discussion with us about our enterprise tier. And we actually price that tier per meeting schedule. So it is a performance based pricing metric, which means we really align our success with your success perfectly 100%. This is a very deliberate choice of ours.

Benjamin Smokovich [00:17:40]:

That's good. It's interesting man. Iker, you're a smart man.

David Stepania [00:17:45]:

I got a guy who's going to give you the distribution. His name is Amir. Do you know who I'm talking about? Amir Writer. He all owns Cloud Task. You know who I'm talking about?

Hayk [00:17:56]:

Don't know Amir, but he sounds like in the future we're going to be calling each other habibi.

David Stepania [00:18:02]:

He's actually in Colombia, but yeah, he owns like a marketplace where they place SDRs and other Rev ops functions. But it's it's very much a marketplace. So he even lists products like yours as like a, you know, as like one of the options. Right. So I'll connect you to him afterwards. He like. Or you can look at some of my comments. He talks a lot of on LinkedIn.

Hayk [00:18:33]:

I appreciate that, David, but I don't want anything from you, David, until you try our appreciate.

David Stepania [00:18:37]:

Yeah, I. Okay, I have a question. Okay.

Hayk [00:18:42]:

If I give you a hard time, you don't have, you don't have thin skin, right, David?

David Stepania [00:18:46]:

Because I now, you know, now that I see this app, I know there's one guy that's using those comments to leave comments on my, on my LinkedIn and I kind of like it because like it gives you engagement. So it's like any engagement, kind of like oftentimes, hells. But that's one of them.

Hayk [00:19:02]:

That's one of, that's one of ours, David. Like we saw that you were getting 0 comments on your posts and we felt, we feel really bad, we feel really bad for David. Please just send a couple, send a couple at him so the guy can get some impressions, some views and some flick through.

David Stepania [00:19:19]:

Yeah, so I know there's a guy using him, but then that's the problem. It's like, I know he's using something because it comes off a little bit, you know, like robotic. But. So the question is, can you change the tone of voice on this thing? Like, can you tell? Like, you know, like, hey, like, don't use that tone. I want you to sound like angry or I want you to sound like, you know, friendly or like something like.

Benjamin Smokovich [00:19:40]:

That, you know, of course you can. If, if they have prompt engineering at Core ip, they can do that. Right? But from a UI UX perspective, that's doable too, I imagine.

Hayk [00:19:49]:

Let me turn that question on you, which is, what's the right UI UX implementation for that? Given that, you know that the average person has no business prompt engineering because they're not going to have evals. They're not just going to guess and check. They're going to look at the comment after they change the prompt. They're going to look at like two of them, a statistically insignificant amount and then they're going to make a gut check decision. Right? So should I just give you little buttons that say make me sound smarter, make me look better, make me taller?

David Stepania [00:20:17]:

I just like, there's like, I think the main thing is like sounding conversational. Don't talk about you have no expertise in, you know, like you're better off kind of leaving like a snarky comment, I feel are not snarky, but like just kind of like a, a funny comment than talking about, you know, AGI or rags. When it comes off out of context, I think that's the biggest giveaway. Like when it goes. It wants to talk about details, but it's like, you know, like it's, it's not quite there. Not quite there yet, you know, David.

Hayk [00:20:57]:

You'Re hired, you're hired as a product manager on our team if you want the role. That being said, look, we've done a lot of evals on this and we pay close attention to it. And the gist of our comment prompts is that you want to ask someone a question to tee them up to expand on their main argument or hypothesis of their post. The whole point is to make the other person look and feel good. This is Classic Sales 101. Go read any sales book. All of us have read them prior to entering the profession. And it's not about talking about how great you are, how smart you are, what, you know, sales is all about talking about what the other person has done, why they've done it, why that's interesting. How did they make that decision? People love sharing their opinions.

David Stepania [00:21:52]:

That's a great idea. Great idea comes off authentic, then you're gonna like, really hit the nail on his head, right?

Hayk [00:21:59]:

I'll say this, we're already really close to passing the turing test. Like GPT 3.5 beat humans. And then humans got smart. We're smart, right? We like leveled up and we started to like, notice what the AI slop was. We're like, look, AI, you're not going to get rid of us just yet. And then now GPT4.5 and Grok3 and things are coming out where they're just leapfrogging us entirely again. Like, the creative writing is so good. Being part of the context of the conversation is so good. Getting rid of the AI slop has become so good with the stronger models.

David Stepania [00:22:31]:

And so, yeah, crap, I didn't like, I saw some people talking shit on my post about AI slop and I need to upgrade to GPT5.

Benjamin Smokovich [00:22:43]:

Forgot to download his service. What else is on your roadmap? What's coming up? Is that something you could share about? Like. Like, are you going to get into the whole realm of I'm going to target. I'm going to build intelligent models into my system to be able to say I should be going after this type of profile on this type of company. What. What's Coming up for you.

Hayk [00:23:05]:

I think there could be a future in targeting for us specifically book building lookalike audiences for you based off of your existing customer base. We all know that targeting is really painful and really tedious. Like you're giving it to a new college grad SDR a lot of the time and you're like, here's this filtering ui, the spreadsheet like or bi tool like user interface like good luck Build, build an awesome targeting prospecting list, right? We're like, we're setting people up to fail. So I think there is the room for a lot of AI innovation there. People also ask for a lot of different platforms. Give us X, give us Reddit, leave us email. If you think about what we've actually built, we've built a pipeline, we've put me in a box, right? I'm the principal PM for the prompt engineer and I've been ahead of sales at three venture backed companies all Andreessen Horowitz backed one of them about to IPO soon. So we just, we wanted to give that to people and you take that and you put that on different platforms. It doesn't necessarily matter that it's on LinkedIn. LinkedIn just happened to be the best place to start. But we will definitely build platforms in the future as well. And then something we delivered recently is we just got done shipping a 13,000 token 40 page prompt for our direct messaging. And so we're just focused on how do you put the best salespeople in the world in this magic AI box? How do you, how do you actually get it there? A lot of people think they can just a b test their way to like the best prompts in the world. We tried that approach does not work at all, just generates AI slop. So you just kind of, you do have to. It really pays to be a subject matter expert. It really pays to just write a book about sales and throw it in the context.

Benjamin Smokovich [00:24:57]:

When you're talking about that. David and I were talking about that on our last call. We were talking about the knowledge worker, right? It's going to be the king of AI, right? And so whether that be in sales rev ops operations, you take the best of the best and you try to put their, their brain in a box and that's what you build it around.

Hayk [00:25:18]:

But you have to be like a true expert. You can't be like part of the mean who's just kind of getting by.

Benjamin Smokovich [00:25:23]:

Correct.

Hayk [00:25:25]:

You have to be like such a true expert. And then maybe the irony of is the is it that we just like, obviate ourselves and make ourselves replaceable by AI. Once. Once you give away all the true expertise, but who knows? It's too much. Too much.

Benjamin Smokovich [00:25:40]:

That's cool. Are you going to add any, like, social listening capabilities where you're gathering not only, like, I'm going to respond to this guy, but I'm going to maybe grab some news articles over here from like 4000 RSS feeds and I'm going to summarize the article and then I'm going to throw a little bit of that into my messaging. Or maybe I have a report that's coming from, you know, their 10k analysis report off their website. And now I'm looking at like some of their pain points they're talking about and maybe throwing that in the messaging. Are you doing anything around that?

Hayk [00:26:12]:

Bingo. We do do the deep research for you, but you'll never see it in the comments. That's only for the DMs, and it's very deliberate. It's like what I was talking about earlier, where if you go to someone at a cocktail party or a conference or someone you've just met, you start talking about how great you are and how your day was. That would be pretty strange, you know, so it's, it's really important to just talk about what they're talking about in the comments, tee them up to be successful in expanding their hypothesis of what their post was about. And then later, once you've broken the ice and you've established some sort of relationship, they've let you into their inbox by accepting the connection request. That's when deep research comes into the mix. That's when you get to start asking them questions about their business. Because you're no longer a complete and total utter stranger.

Benjamin Smokovich [00:26:59]:

Correct. So does your tool go all the way from engaging, commenting, doing that, to where you're gathering interest, and now you're having a full conversation, the DMs to booking a meeting. And so I could just be a person, like, whoa, I got a meeting from, from this person.

Hayk [00:27:18]:

That's exactly right. Between DMS and booking a meeting, we do do discovery and qualification. So it's really critical. You don't want to just book meetings with anyone on your calendar. Right. If you end up with a bunch of unqualified meetings on your calendar with people who want to waste your time or pitch you something that is not interesting. So we do build, currently, how we figure out how to do the discovery and qualification is that we do that research ourselves. We've got A number of prompts to learn about your business and understand who your ICP is, what problem you solve and what questions you need to be asking them. But we are, this is an area in which we are launching some custom instructions is just give us some examples of the best discovery questions. It's very cut and dry. Every business with product market fit as them. Every business has three to five great discovery questions and so we will actually intake that coming very soon to just improve the quality of the discovery.

Benjamin Smokovich [00:28:16]:

I just signed up for your tool hike and let me tell you, I talk to third party vendors all the time at SAP, just companies trying to sell to us and I don't really sign up for that much to be honest with you, because I see so much. So I signed up to your product. I am curious. I'm excited about it for sure.

Hayk [00:28:38]:

Thank you. Appreciate that.

Benjamin Smokovich [00:28:41]:

I'm going to be testing this.

Hayk [00:28:42]:

I'll send over the enterprise order form. You can get it in the hands of procurement.

Benjamin Smokovich [00:28:47]:

Let me ask you this, out of all your clients that you've got, what's your biggest metric in terms of how you've helped them expand their revenue? Have you had any metric reporting on that?

Hayk [00:29:00]:

Yeah. Meetings, book, meeting.

Benjamin Smokovich [00:29:02]:

Okay.

Hayk [00:29:03]:

Our most successful clients are increasing their meetings booked rate by 35%. That results in pipeline coverage.

Benjamin Smokovich [00:29:13]:

Yeah, absolutely.

Hayk [00:29:15]:

In sales, if you're world class, you actually lose 80% of the deals in your pipeline. That makes you world class. You're snowflake or datadog. You're a world class company. If you lose 80%, you just win 20%.

Benjamin Smokovich [00:29:31]:

Yep.

Hayk [00:29:32]:

And so if you speak to any CRO, any sales leader, they're all after the mythical 3-4x pipeline code and they will hound their reps about that. So if you're a sales rep that's getting hounded by your CRO, who's sitting in their ivory tower, you know, who's who's driving around their really fancy car and just telling you to make more cold calls and send more emails. I highly recommend AgentGrowth even if you have a full pipeline. Nowadays in sales, people are expecting you to have activity metrics inputs. You should be prospecting for the coming quarter. We need to build our pipeline for the coming quarter. This is one of the most distracting things that CROs and VPs of sales do to their reps, by the way. It's like, hey man, I have a full pipeline for Christ's sake. Just let me close it. Yeah, we actually have some users that just use AgentGrow for the activity metrics and then they'll show the activity metrics to their leadership. They'd be like, look at all the prospecting I'm doing. Could you please just say that I'm doing a good job?

Benjamin Smokovich [00:30:40]:

Let me focus.

Hayk [00:30:42]:

So that was a funny use case. I wasn't expecting that. But as someone who's lived in the trenches as a sales rep and a sales leader, I didn't see that one coming.

Benjamin Smokovich [00:30:50]:

Yeah, yeah, that's cool. Hey, well, I'm in love with your solution, David. Do you got any other questions? Because I would say, I would say the humans on the chopping block here.

David Stepania [00:31:05]:

I mean, I could see, you know, this thing freeing up a lot of time for, you know, for like the top salespeople's, you know, time, basically. I mean, like, essentially, like, what do I do? Like I, I hike. How did you hear from me? It's like one of my assistants or like my sales reps or whatnot, like, reached out to you and now we're talking. Right? So that frees up time for them to do something else, you know, like, at least in my case, you know, the way we've been doing stuff, I've retrained everyone to do things that AI can't do. I didn't let go of anyone. Like we're a small, tight knit company and so we're all learning and we're making things happen, but it allows us to do so much more. So at least in my case, we're not firing people. We're just not. We don't need to hire the people that I used to hire. And that's like a lot of people, they don't understand that. They're like, oh, AI is not going to take any jobs. Yeah, it's technically.

Hayk [00:32:10]:

I will interrupt and I will say this. Sorry, David, I interrupt you a lot today, so please forgive me if, if I were a sales rep using AgentGrow today and you're getting a DM conversation started with someone. Other times that someone will log off of LinkedIn and you won't convert, then in there, pick up the phone and cold call them. And this is where the human is in the loop. I cannot do this for you yet. Hey, thought it would be easier to call. Messaging on LinkedIn is kind of tedious. Anyway, wanted to introduce myself and if it's okay, ask you a couple of questions. Right, so. So pick up the phone, make that cold call. That's the next step there. We're really warming up a lot of these conversations for you. If we don't book the meeting outright.

David Stepania [00:32:54]:

Yeah, 100%. I mean, Ben, you guys said.

Benjamin Smokovich [00:32:57]:

Right.

David Stepania [00:32:57]:

Like, you guys did some math and I think the comp.

Hayk [00:33:01]:

The.

David Stepania [00:33:01]:

The phone calls help or the cold calls help like 20. 20% or increase in like 20% of in conversions or something like that. And like, at least for you guys or something like that wasn't it.

Benjamin Smokovich [00:33:16]:

I mean, it helps convert, right? It helps convert. When you get on the phone with someone, you have much better chance of converting. But, you know, even that I've seen. I've seen AI services that are trying to do that, where they're calling customers and saying, hey, I'm an agent. Just want to blah, blah, blah. I want to help you book whatever the use case. I've seen agents that help book weddings for. On behalf of. And there's like, hey, Jen. Which you can make your digital twin and it can match your voice if you give it so many hours of footage. So I see that actually is a good kind of like, next step for you to hike. I don't know if you like it or not, but it's. It's got so much better in the last two years. Because I've been part of like, that whole inception of hey Gen. It came in early to our ecosystem and I was like, nobody's using this crap. We're not bringing it in. And we kept them around on a couple users. Yeah.

Hayk [00:34:08]:

Be honest. Are you guys. Hey Gens right Now, how many Rs in the word strawberry? How many?

Benjamin Smokovich [00:34:14]:

There he goes.

David Stepania [00:34:15]:

There he goes. It.

Benjamin Smokovich [00:34:15]:

It. It. Can't get it right. Still can't get it right. Yeah.

Hayk [00:34:19]:

So those models are getting really good. They don't pass the Turing test. Same thing with Voice, which is why we're not quite interested in cold calls.

Benjamin Smokovich [00:34:27]:

Ah, yes, I see.

Hayk [00:34:29]:

I've gotten some AI cold calls recently. And, you know, we ra. We raised some money and it goes on Crunchbase, and suddenly you're on everyone's ICP target list. And so I've been getting these AI cold calls, and I'm just like, ah, what's going on? Funny story. One of the best calls I got recently that actually converted for the guy, he sounded like an AI. And I was like, hey, are you an AI? And he's like, no, I just talk like this, man. And I was like, oh, that's really interesting. What do you want? So he actually got past, like, most of my defenses by just talking kind of robotically. It might be a tactic that some sales reps should implement.

David Stepania [00:35:09]:

Yeah. By the way, a lot of good insights here on the call I feel like we all understood. I mean, like, I don't know, like, I started out as a BDR or BDR sdr, so that was pretty much like my last job actually. And then since then I've used that knowledge to start a couple of companies and been doing that for the past like 10 plus years. But yeah, man, it's crazy where the technology, how far the technology has come.

Hayk [00:35:40]:

Me too. That's why I built this company. I've really done sales my entire career before becoming a founder. And I have to say, salespeople have it really bad. And if I have kinship with any group, it's salespeople. And I want to make the lives of salespeople better. Like, none of us have enough pipeline. Most everyone's missing their revenue goals. Go look at repview or one of these statistics aggregators on what average attainment rates of quota is for a salesperson. It's horrendous. And so, you know, for the people who punch down and say, like, oh, you're using prospecting tools and automation, like, you're such a bad person. They make a value, they make a moral judgment about, you know, those people, those people could never survive. And those people could never survive doing a sales role. The people who publicly roast salespeople are like, oh, they cold called me while I was in a meeting. Come on. Like, we have a million dollar annual quota. We have families, we have mouths to feed. And by the way, any deal we ever close is closed consentingly by the other party. It's not coercive. Right? So they are happy. We are solving a business problem for them. So there's people that are just so out of touch with the sales profession who are like, don't use AI. It's so unprofessional. And if you go look at our user base and go look at who the adopters are and who the active users are, it's people who carry quotas. And if you go look at who our critics are, it's people who don't carry quotas.

David Stepania [00:37:05]:

Yeah, makes sense. Makes sense.

Hayk [00:37:09]:

Start the call with that clip so that, so people can know that this episode is these people. People can know that this episode is for them if they carry a quota. If you like to stay on this line, I will just keep making like hot takes for the next hour. So you guys should hang up on me. Hang up on me now.

David Stepania [00:37:29]:

Yeah, give us some. Like, what do you see? Like, I, I feel like right now everyone's trying to sell AI services. Do you think companies are going to come up short, what's your thinking or what are you thinking?

Hayk [00:37:45]:

I think it's just a matter of time, guys. I think it's just a matter of time. And the models just continue to improve. And if you look at any of the research on model improvement, we haven't hit any walls yet. And people just continue to discover new architectures for improving the models whether it's larger data sets or test time compute and there's probably more coming. AI moves at the speed of light. Right. So I think that if the vendors fall short right now, which many of them do and they're, you know, there's not great retention in a lot of use cases except for the, the coding agents, maybe enterprise search like glean and people like that. It's just a matter of time until there's tons of new killer AI use cases because the models will continue to get.

David Stepania [00:38:32]:

Yeah, I feel like there's a lot of. There used to be like even looking at my post, I would post some stuff, it will go viral. I mean you'll get some haters. But there was a lot, a lot of people who like that stuff too. Now you're like, you have a lot of AI haters I feel like. And I think they're kind of like having a moment because some companies, AI initiatives are kind of failing. But the way I see things, I think it's kind of like the Internet moment. Like you know, we're going to have new Amazons, Google's come out out of all of this and I think maybe even some big companies will fall. I think there's real risk of that. We'll see.

Hayk [00:39:14]:

I read a statistic that OpenAI is the seventh most visited website on the Internet today. I don't know if that was fake news or what it was. And I don't use Google and I don't use Google Search anymore. I don't use Google Search. And so Google's a tremendous company. They have so much going on. But like searches your cash cow, I'd be a little bit worried.

David Stepania [00:39:39]:

Yeah.

Hayk [00:39:40]:

And they have innovators dilemma like they have if they want to put AI into the search bar they will remove people clicking through and actually monetizing ads which is where they make most of their profits. So it's tricky. It's tricky for Google. I will say one thing that's, that's hard about AI agents is that if you are letting the LLM do all of the heavy lifting, how long is it, how. It's just a matter of time until a model Provider meters their API usage, finds out what the killer use cases are and releases that as a new AI agent and they have distribution and they just crush you. Right. And so that's, that's one concern, that's one big concern because look, they are, they have investors, right? And they are, there are some huge bets being made. And so they, they and, and frontier models are being commoditized frankly. And so they do need to get into more agentic use cases. And we're seeing Anthropic launch a command line tool for writing code as well. And so, you know, they, they probably saw Cursor doing tremendously well, skyrocketing to 100 million in revenue. And, and they just, they just come and compete. And so what does that mean for vertical AI apps like my company or for like the builders out there? Do you, do you have to do your AI, your API calls across like three or four different models of providers to keep them guessing? You know, just do like a bit of a head fake so they don't actually know what's going on.

Benjamin Smokovich [00:41:09]:

That's what we did. We brought in all the models we could and then we also brought in open source models. Right. And we are fine tuning some of those and we're connecting them to probably 10 or 11 different internal APIs so we can get specific on our own content, our own data, our own use cases. So yeah, that's, that's what we've done.

Hayk [00:41:32]:

Maybe it's an argument for only using open source models.

Benjamin Smokovich [00:41:35]:

Say again?

Hayk [00:41:37]:

Maybe that's a great argument for only using open source models. Yeah, it's like completely defend yourself from these people who are eventually going to cannibalize. I mean we saw it with the public clouds. Look at GCP or AWS and what they do to confluent and other open source softwares. Spark, they tried, but databricks is winning. You know, they just, they just take it and we've seen a bunch of people move to a copy not quite open source MIT license something more of a hybrid so that a big cloud provider can't just rip off all of their, all of their hard work and technology so happens in a lot of industries.

David Stepania [00:42:16]:

Yeah, I feel like if you kind of aim for coding is like a massive thing, but if you aim like a smaller niche, I'm pretty sure you're safe. Like which makes it hard, right? Because like if you're building like a startup and you need to pitch your VCs and tell them like how do you get to a billion dollars? Well like Anything that gets to a billion dollars might be on the radar of OpenAI or, you know, anthropic and et cetera, et cetera.

Benjamin Smokovich [00:42:46]:

Or you get to the whole, maybe you don't need to do vcs. Maybe you can just build quick enough in the future where you can just do a bunch of micro niches.

Hayk [00:42:55]:

Well, we've done VCs, so we raised, we raise money and we love our investors and they're great. We've raised seven figures.

David Stepania [00:43:00]:

Nice. What, did you just close the round or when did you guys raise?

Hayk [00:43:03]:

We just closed the round. We have another one opening now.

David Stepania [00:43:06]:

Oh, wow, Nice. Congrats.

Hayk [00:43:10]:

Insert address if you're an investor watching.

Benjamin Smokovich [00:43:13]:

This, insert address to email hike.

David Stepania [00:43:18]:

Yeah, we, we got some VCs joining our newsletter. So yeah, they're definitely watching.

Hayk [00:43:24]:

And so. But to your point, David, is sales a small market? Like coding? Your point was like, coding is a really big market. Is sales a small one? It's like, no, it's arguably the biggest in the world. Like, we're all salespeople whether we like it or not. Right?

Benjamin Smokovich [00:43:40]:

Yeah.

David Stepania [00:43:41]:

I guess where I'm going with this is like, I have, you know, like I, I know some companies where they, they were the, you know, they had the first mover advantage, all this and that. You know, they, they grew or they were growing quickly and then someone just comes in there and copies their product and within like two years, like has a quick exit, not like a huge exit, but like a quick 50 to $100 million, like super profitable exit, and they're out, they're done. The other company is now they don't know what they're doing. Like, they don't know what to do because all this other competition is popping up and it's eating away their like, clientele and stuff like that. So I guess where I'm going is like sometimes just copying something quickly and doing a quick exit, like, could be the way to go. Like, yes, you're not reaching a billion dollar valuation, but you're not watered down or like, you know, like you're not underwater in terms of equity and stuff like that. So I think that could be one way to go.

Hayk [00:44:42]:

I don't know if we get cloned because of your advice. I'm sending you the bill. Nah, I'm sure, I'm sure we'll have competitors and I'm, I'm less concerned about other startups and I'm more concerned about the, the large frontier model providers.

David Stepania [00:45:03]:

Yeah, gotcha. Interesting. Well, Ben, any questions? We're coming up on time here. I don't know if you wanted to add any.

Benjamin Smokovich [00:45:14]:

No, but I'm really. No, no questions. But I, I am really interested to have a follow up with Hike, you know, and you know, maybe three months so we can see the growth, we can see, you know, how the products developed and I, I, I really want this guy to win. So I'm hoping to see explosion happen from now till then. Three months is forever in the AI space, so thank you.

Hayk [00:45:38]:

Appreciate you guys a ton. I want to see you all win as well. This is the bad best podcast that I've been on, my work history and this was a ton of fun, guys, so thank you so much.

Stream the full episode on Apple, Spotify, or YouTube.

Find interview & all prior episodes here.