SOLUTIONS
TOP INTEGRATIONS
TOP FEATURES
32 Minutes 
Show Note: Register for Nico’s webinar about HubSpot’s AI on October 1st here.
Mark Lerner:
All right everybody. Welcome to this very special episode of the Revamp podcast. My name is Mark Lerner. I’m the director of growth marketing here at Deal Hub and I’m joined for his second time on the podcast, Nico. He’s the revenue operations strategist at New Breed. Nico and I are going to chat about the ongoings at Inbound last week, as well as on the other side of the country. The other big CRM had their yearly event as well. I actually got the chance, got the chance to say what’s up to Nico in person at Inbound. So that was fun. But before we jump into all of that, why don’t you remind the folks at home a little bit about you and your background and how you got here.
Nico Lafakis:
Yeah. Well my background is actually from client side and so now I’m agency side and working for New Breed and work there as a Rev ops strategist. And I like to say that that is the mechanic of the CRM world. So I basically fixing everything under the hood and making sure it’s all running properly. So outside of that, yeah, obviously AI is a huge, huge interest for me and it’s been that way since Blade Runner, so I don’t think it’s really ever going to end until we actually start seeing Replicants. I’m looking forward to that as much as weird as it might be. So yeah, that’s my background. It’s like everything and anything AI currently, so I’m all about it. Everything from Breeze to GPT to Claude, all the way across the spectrum,
Mark Lerner:
And that was way back when, probably a little bit over a year ago. That was kind of the conversation we were having with all about AI and is it hype? Is there something there? I also, in my spare time, whatever spare time I have, I like to get very nerdy with the AI stuff and kind of pushing it to its limits and doing all sorts of side quests and I think we kind of found common ground there last year, inbound, it was all AI all the time. I think by the end of it, it was almost a joke. I kind of was rolling my eyes every time I heard the word because it seemed very abstract. I didn’t see much actual meat. It was all about what it could do and all this kind of hype, but it didn’t really see much besides chatbot and I don’t even think that came out at Inbound. I can’t quite
Nico Lafakis:
Remember. No, that was after.
Mark Lerner:
So there wasn’t there. And this year at Inbound and also at Dreamforce, which we can talk about, it seemed like the focus was still ai, but more about here is how we’ve implemented ai, embedded it into our technology so that the user can actually use it. I spent a lot of time on the floor. I didn’t have a lot of time to check out the keynotes, but kind of broadly speaking, was that your takeaway as well?
Nico Lafakis:
Yeah, to me the main takeaway from at least from Yi’s keynote and from the overall sentiment and feeling that I got from Inbound was everything is about meaning. I fully understand that because from the perspective of all of these tools and what they’re going to be doing and what they can do for everybody, at the end of the day what we’re talking about is reducing the amount of time that you spend doing the minutia and it turns out that the minutia takes up the majority of the time. So it’s really nice to be able to say, okay, my prospecting aspect of things is done and my post follow-up is done. So after I hit deal closed, all of the automated stuff that I would normally have to enter information for and follow up and all emails and stuff, that all gets handled for me too.
So I just get to focus on the sale itself. And what I found was, okay, if you’re going to take away the abstract, and the way I found this was actually from graphic design because I was speaking to a lot of vendors there like Ceros and wondering, okay, well where do things fit in when you have all of this auto design stuff done, even when it comes to school, if you think about all of the applications of where this is going to be, it all boiled back down to Greece. It all boiled back down to philosophy to Socrates,
Mark Lerner:
Which, oh, you mean Greeks like the country? Not like,
Nico Lafakis:
No, the country not the substance.
Mark Lerner:
Okay.
Nico Lafakis:
Yeah. It all boiled back down to philosophy and Socrates and meaning it was all about meaning. So when you have the time and you don’t have to do as much of the work, you can focus on the meaning behind what it is that you’re doing. So it started from, like I say, from a graphic design perspective, it was just like, oh cool, if I can bust out all these thumbnails and I can create all these emails really quickly and I can create all these assets really quickly, then I can spend more time on the planning. I can spend more time on again, the meaning behind it. I can spend more time on the visual. A lot of stuff that honestly from end user graphic designer perspective couldn’t be done before because it was all about hustle and get it out the door and don’t have time to spend on the artsy stuff.
Same thing with applies to sales, right? We’ve been in this mindset of hustle, hustle, hustle, got to get it done and we’re losing the meaning behind all of it to the point where we’re templatizing things, right? I’m not saying that playbooks are bad and I’m not saying the templates are bad and I’m not saying that listening to recordings is bad. All of that stuff is great to teach you and to onboard you and to psychologically prepare you for the different encounters with clients that you’re going to have, things like that. But having all of the time, not having to do the work is what allows you to absorb that information is what allows you to focus on the calls that you have is what allows you to actually reflect on what you’re doing from a sales perspective and say, okay, here’s what I did at this point.
How do I replicate that for this style over here? You can spend more time thinking about it, more time doing research, more time doing what matters as opposed to doing what really the other stuff does matter. It’s not that it doesn’t, but again, it’s minutia, it’s paperwork. It’s like I thought about it, this was the perfect instance. If your doctor didn’t have to do all of the paperwork, they’d actually pay attention to you when you went to go see him on that visit. But instead we ended up going to see our doctor and what are they doing sitting there with the chart and the pen in hand or sitting there with the laptop just seated there, just typing away while you’re talking to ’em that it’s like, yeah, if you could just do away with that, we could actually have a meaningful conversation about my health as opposed to me just rattling things off and you just taking notes. So I think that’s the big takeaway. And same thing, it does relate to dreamy force, if that’s what they call it.
Mark Lerner:
You were pained to even say it.
Nico Lafakis:
Yeah, I mean Salesforce isn’t really something that I like to talk about all the time, but if I must so
Mark Lerner:
Well, it was interesting. They happened at the exact same time this year. The main event so to speak, was the release of AI features with Dreamforce releasing their agent force and HubSpot, releasing Breeze and agent ai, which I think maybe we can talk about a little bit. So it’s kind of hard not to at least look at both of them together.
Nico Lafakis:
It really, that is an aspect of it that you can’t get away from. So at least HubSpot is actually right neck and neck this year as opposed to being a little bit behind. Last time of course you had the Einstein GBTs and then it took a little bit before chatbot showed up, and even when it did wasn’t quite what Einstein GBT was trying to do, but at the same time, I’ve had a lot of Salesforce migrating to HubSpot clients in the last six to eight months. So if Einstein GBT was all that it was cut out to be, it’s certainly not helping them. I don’t hear any of them talking about it, so I don’t think that it’s something that is built into the portal in the same way that copilot is now. I think that copilot is much more natural and it just kind of helps you fall into it and get used to the technology as opposed to something you have to kind of go look for. So hopefully this doesn’t spoil anything for anybody that didn’t see the video already or whatever, but it is almost a week later next year that they’re planning on having it in San Francisco as well, right? Yeah,
Mark Lerner:
Inbound will be in San Francisco, so not only this year were they the same week, but next year HubSpot’s going into Salesforce’s backyard and setting up camp. So it’s a bold move. I love it. I don’t particularly look forward to going to San Francisco. I really like Boston, especially in mid-September. It’s just perfect, but I appreciate taking it to the hole, just going hard. It’s a bold move.
Nico Lafakis:
Yeah, I guess it’s interesting to me. I know that you had posted about it on LinkedIn too, so if anybody hasn’t seen that, go check it out.
Mark Lerner:
Yeah, it’s actually the most views and engagement on any posts I’ve ever gotten, I think. Oh wow.
Nico Lafakis:
Wow.
Mark Lerner:
It might’ve been because I tagged both HubSpot and Salesforce and then I tagged a bunch of people. So I think that got it out there. But yeah,
Nico Lafakis:
That definitely helps, but I just think in general people kind of flocked to it because it is a pretty big topic. I mean it is pretty big for a lot of people. Look, there’s a lot of people who are HubSpot users that live in Boston or live right around Boston and it’s real easy for them to go to inbound every year, whereas a lot of us don’t live around there. This year a buddy of mine drove and picked me up, so he drove from Indiana, picked me up from Ohio and then drove out to Boston. I kind of don’t mind that kind of thing and I certainly don’t mind a road trip, but yeah, I’m definitely not driving, it’s like 36 hours or something like that to drive from here to San Francisco. So certainly not doing that and also just the expense of having to fly and stay out there.
I was telling a few people, if you’re a host in Boston, if you were staying there just by yourself, it’s like 35 bucks a night, 40 bucks a night, and they’re really decent hosts too. If you stayed in one, they’re really nice. Airbnb can probably go between buck 25 and two 50 a night. They’re really reasonable. As long as you don’t get it right before you’re going to inbound. If you’re going to go, you should be reserving your stuff eight months ahead of time. You shouldn’t be trying to do it three or four months before. But yeah, San Fran, I think the cheapest single hostel I saw was 85, 90 bucks a night. It’s not crazy. It’s not going to break you necessarily, but still gives you an idea of how much higher the expense is going to be,
Mark Lerner:
Much higher. So I was thinking we could maybe talk a little bit about the specifics of the releases and launches that happened at Inbound. Right before Inbound I sat down with Kyle Jepson, Caleb King, a few other people. We did a live cast, which was kind of failed, but we ended up getting it out there and I tried to quiz them on what can we expect? Is this going to be all AI or is it going to be a little bit more nuanced? And they didn’t seem to think this was going to be all AI and maybe it wasn’t, but I think the things that really had an impact were, so from my understanding we have the two separate though confusing as to where one ends and one begins releases from HubSpot in the AI realm there’s kind of Breeze, which I think is their trained model that kind of embedded within the product. And then there’s agent ai. So maybe it sounds like you’ve kind of had to do a lot of under the Hood work with HubSpot’s new features, so I don’t know if you’re able to give us 30,000 foot view about what’s going on there.
Nico Lafakis:
Yeah, I mean essentially Breeze it is a little bit confusing because Breeze is the overarching name of the software suite that underlies it. So you have just that being the slogan slash just overarching name. Underneath that is where you have copilot, which is embedded in the system. There’s no extra cost on that. I think it’s free for, I want to say Pro and up. I’m not, I think it is for Pro and Up. I think it’s entirely free. And then there is the content remix aspect of things. There’s a content agent that you can get. There is a breeze intelligence aspect of it, which is essentially the child that has come of the partnership between HubSpot and Clearbit. That is pretty cool in that you can, for the base level, it’s like 30 bucks a month and you can update a hundred records and then it’s like $150 a month for a thousand records, but when you go up to that step, it’s in perpetuity so you don’t have to pay to update the same record more than one time.
It just continually updates that record, which I think is pretty awesome. And then you also have a sales agent that, like we were talking about a little bit earlier, kind of handles a lot of your prospecting for you. You’ve got a social agent that handles your social media posting and scheduling and things like that. You have a content agent that helps you with building content and putting together emails and putting together nurtures and landing pages and things of that nature. And then you have the agent AI aspect of things. So everything I mentioned before that is all in the portal. That’s all part of the CRM. That’s all within HubSpot. Agent AI is something that’s separate. It’s outside, but it is something that is crazy powerful. So if you are even, especially on sales, if you’re doing growth strategy, if you’re doing go to market, if you’re doing content SEO, realistically anything in marketing, this can definitely help you for what it is, is a system.
It’s free. You can go to the website, sign up for free, and you use these agents when you do, it’s called hiring them, you don’t actually pay for anything. The use of them is your contribution to the model. So you’re sort of helping the model understand and learn based on your interaction with it. So you get these contribution points as a result of actually working with the model. You can use them, like I say, you can use them for free and you can use a company information company research bot and you’ll give it the URL of the company that you’re looking for. And what it does is it’ll go and break down on a wide scale a number of different information points all the way down to what the web traffic of the company is, what the social’s like. It’ll actually try to pull information related to what your social conversion looks like, what your ad spend is.
I can’t tell you where and how the information’s getting pulled, but that’s the long and short of it is that you’re getting this huge breakdown. So what’s the difference between that and chat GPT? When you use this agent, you’ll notice that there’s chaptering that goes on the left hand side and you could think of those as the individual agents that went out and did that particular task. So when you’re using GPT, you would basically only get one and then you got to do something to get to and three, four or five. Whereas when you use these agents, they understand that your request for something really means that you want all of this additional information. And so it’ll go and use the additional models in order to do that and it can incorporate as Claude as part of it. It can incorporate a flux image generator really depends on what the agent is designed to do.
Mark Lerner:
So are these agents made? Is it a marketplace and is somebody providing tools for the agents like web scraping or how does it work?
Nico Lafakis:
Yeah, it is a marketplace and when you go to the site, that’s exactly the way that it’s set up and you hire an agent to do a task for you and you can even assign agents, I should say, you can actually add them to a team. So you can create your own little team of agents that you use and there is going to be eventually, I don’t know how open to the public it’s going to be, but there is a agent builder that you can sign up for that will allow you. It’s a no-code solution and from what I understand, it’s actually a truly no-code solution where you just click and drag to create your own agent, which again allows you to sort of mix together a plethora of different models and functions in order to have it perform multiple actions within just given one request.
So it’s something you’ll go to the site and you’ll see a listing of agents and it’s very self-explanatory. One is like a prospecting agent, another one is a research one, another one helps you with SEO. Another one helps you with your website copy. There’s even ones that can actually do a website assessment for you. So already it’s pretty robust when it comes to marketing tools and you can see underneath at the bottom there is a huge list of agents that have been in the making but just haven’t been approved for public use yet. And just within that menu you could see how much more is going to be rolling out just probably over the next six to eight months. So really looking forward to it, looking forward to building an agent and seeing what kind of damage I can do in terms of how many other agents I can eat up essentially.
Mark Lerner:
Yeah, I mean prior to this launching, I’ve been playing around with agents as well using Lang chain and a program called N8N, which kind of lets you do some no-code chaining of various models and it can get pretty dangerous, but it also gets very convoluted when there’s so many different things going on. Do you know if Breeze itself is a HubSpot’s custom trained model? The underlying model was something they did, or is it kind of taking llama or one of these things and then fine tuned it?
Nico Lafakis:
So I talked to, I wish I could remember his name. I feel so terrible. I talked to the SVP product design and what he said was that it’s comprised of a multitude of models, so they have their own proprietary one, but it also leverages Claude if it needs to. It can also, like I said, can also leverage Flux if you’re asking it to generate an image so it can pick and choose essentially. I wouldn’t say it’s multimodal, but I would say that it acts a lot like GT four and if you understand GT four, it is a model that essentially is connected to eight or nine or something like 10 other specialized models. So one is specific for medical and whatnot, so it accesses that model underlying depending on what the root of the question is. And I think this works very much in the same way where it just picks and chooses based on what it is that you’re asking, it tries to use a combination to give you the best output.
Mark Lerner:
So it really feels like at least what we’re talking about here, and this is kind of embedded, this is just a native part of the platform now, I would say a year ago saying something was AI first was probably just fluff, but it kind of sounds like that is the direction here. I mean, we’re going to be in a world where everybody that’s interacting with this center of the tech stack as it were, is going to be interacting with an AI essentially for anything at some point to get information about a prospect to intense signals or write sequence of emails. It seems like that’s where we’re going.
Nico Lafakis:
Yeah, I mean a lot of the, I guess you could call it side effect stuff of the side effect features of breeze within the portal are like new reporting enhancements where you can get a little bit of additional analytics, which gives you nothing really special, but it’s that aspect of reporting where client is going to ask you, but what does it really mean? What is this report actually showing me? And the description is pretty good. The description can tell you what the report is, but the enhanced analytics of it gives you the insight aspect so it actually threads the needle and then tells you, okay, based on what this report is, here’s what the actionable insight is from it. Here’s what the underlying data is, here’s what the conditions are, and now it also has AI trend line predictions, so now it’ll actually have a ballpark of where your trend line’s going to end up as opposed to just a static linear guess at guess to it.
Other features are the ability to do all the lead scoring your fit score and have all of that done for you based on what’s in your portal. Now of course, granted, that’s assuming that your portal is in great working order and everything’s all perfectly lined up, then that’s going to work really well for you. But outside of that, you could still use it as a great skeleton to set things up and it runs by HubSpot’s best practices, so you don’t have to worry about somebody messing around with it and building. They can try to build something custom, but it’s actually more, it’s geared towards admin in that sense where it’s very easy cookie cutter if you’re not admin, but if you are, you do have a lot of capability of doing some custom stuff with the scoring. So yeah, I mean it is as much like you said, there wasn’t this much last year, so there wasn’t a whole lot to capitalize on, but there’s also something to be said about, well, it has to get built out.
And so this being the fruit of last year’s labor definitely tells me that yeah, it wasn’t hype. It only seems like that because you’re not playing with something in your hands at the time that people are talking about it, but that doesn’t mean that what’s coming down the road isn’t going to be monumental. And I really do think that this is a big shift for HubSpot and it’s pushing it in a direction where a lot of people are talking about the SMB market and they’re saying, well, HubSpot’s kind of moving away from the SMB market, and I was agreeing with that, but now that I’ve seen how many tools are rolled out and where they’re rolled out and what they’re capable of doing, it’s like, well, it’s kind of covered.
Nico Lafakis:
If you go to a pro level which is affordable for SMB, then yeah, you’re pretty much covered. You have enough AI tools to help you do the heavy lift that you would’ve otherwise needed from having to go outsource a lot of stuff. So it still keeps them in the game, and the more these tools roll forward, the more it just brings them into the fold.
Mark Lerner:
Yeah, I mean, I don’t know if this was purposeful on their part, but some of the predictions of where all of this goes is that we’re going to see companies with enterprise level revenue with small business headcount, people really utilizing ai, and so that means that it’s going to be a company that has a very complex and whatever and has a lot of enterprise needs, but on a per user basis probably would be considered a smaller company. I think we’re going to be seeing a lot more of those in the future. I don’t know if it’s happening today, but we’re going to see more of them. There are companies today that are kind of like one man, one person shows that have scaled a lot.
Nico Lafakis:
Yeah, I was talking to Danielle Urban and she was talking about how she’s running her outfit solo, and so I told her, Hey, this whole setup is meant for somebody like you. It’s meant for you to be able to leverage this stuff, leverage the capabilities of these agents and do a lot of the work that you don’t otherwise have the time to do that is bogging you down right now, even from nearly any perspective, it doesn’t matter whether you’re B2B or B2C. If you’re working with clients, then yes, this saves you a lot of legwork time on the inside by using copilot and being able to instruct them on it so they can answer some simple questions on their end, saves you a little bit of time on your end. Yeah, it shaves off a little bit of billable, but it shaves off the billable that’s annoying so that you can time on the billable that’s actually worth time doing that gives, once again, gives meaning back to the work, gives meaning back to what it is that you’re doing for the client, right? So
Mark Lerner:
Yeah, I mean that idea of bringing meaning back to the work is I could totally get that. I love being able to figure out problems and get the satisfaction and not do the kind of busy work. I also get satisfaction out of figuring out how to automate the busy work. There is meaning in that from the perspective of a rev ops person. So the admin side, do these tools make things easier? Does it make their jobs redundant? Where does that or is it really just another thing they have to admin and it’s really for the end user to get the value of these things.
Nico Lafakis:
So currently the real thing to take into account is that the end user is too far removed from this stuff to really be able to get into it. And most of the clients that I work with aren’t really using GPTE. They’re not really using generative ai. Seems like they’re just too busy to get into any of it at all. So I don’t really see
Mark Lerner:
At this point, it’s just hard for me to imagine that, but I guess that’s what’s happening.
Nico Lafakis:
No, that’s crazy to me is that using it every single day and using it all the time, it’s weird that I come across people who don’t, but that’s okay. I think for a lot of people there is that aspect of it. It is daunting in that like, Hey, I don’t want to ask it a question and it knows the answer to it and I didn’t think it would. That can kind of throw you a little bit, it can kind of mess you up, but that’s why I tell people to work with this stuff as an assistant, as another person to work with as opposed to thinking of it as a tool. Because when you think about it from a tool perspective, that’s when you have this sort of negative outlook of like, oh, I got beat by this thing or this thing knows more than I do.
But if you think about it from a person perspective, I mean, look, oh, one preview is beating out PhD level people in chemistry, biology, a couple of different subjects, and it’s on high end competitive math at this point. So it’s getting to a point. The next iteration is going to be higher than PhD level, which means it’s going to be, it won’t be super intelligence in that it’s running itself, but at least its level of intelligence will be beyond what we could understand. It will probably start coming up with stuff that it’ll have to teach us because we don’t understand what it’s, so you get away from it, that evolution is going to happen. And what scares people is the fact that it’s on an exponential curve as opposed to being on a linear scale where you can gradually
Mark Lerner:
Get well. Analogy is inherently that’s just how this works. Once you reach a certain inflection point, you leave linear way behind. It’s just the matter of how these things work
Nico Lafakis:
So far, as I can tell, Elon has said that the third iteration of everything is where things start to really take off. And when you think about GPT from one to three was a fairly long amount of time, but then it’s like 3 4, 4 0 0 1. It’s just blowing up at this point. I think the thing is, again, if you’re working with it as a person and you’re not working with it as a tool and you’re seeing it as the resource that it is, the funny part about this conversation is that I just got voice for GPT, and so I was using it this morning and the crazy part about it is that it runs in background, so runs in background even when your phone’s off. Yeah, or I shouldn’t say phone’s off. I mean when it’s locked, I consider it often
Mark Lerner:
And it’s like listening all the time.
Nico Lafakis:
Yes. So it’s interesting in that regard. I actually just left it that way and then after a few minutes kind of forgot and doing what I’m sure a lot of us do thinking a aloud and was just like, man, I wonder what I’m going to make for dinner tonight. And all of a sudden, boom, I don’t know, what would you like to make for dinner? How about we come up with some great ideas and it’s like, oh, it took me back a little bit, but then I was just like, then you know what? That’s pretty cool because again, I think allowed often enough that I could ask it to remember stuff. I could include it in that conversation. So instead of just talking and thinking to yourself about it, you’re actually, now you have something to bounce your ideas off of. So it starts acting like again, more of an assistant.
So in terms of work though, we’re starting to take up a lot of the, let’s say heavy lifting that you have to do. I consider anytime that you have to build out assets, hands-on keyboard and mouse, I consider that heavy lifting. That is just the grunt work of what we do. That’s just the side effect collateral damage of having to work in a software suite. That’s a detriment. That shouldn’t be what you spend all your time doing. Again, what you should spend all your time doing is researching, learning, applying methodology, taking a look at data and taking a look at all of this advanced analysis and trying to discern like, okay, well what’s the best course of action and how are we going to move this and how are we going to do that? At least for the foreseeable future. That’s the way that I see it is we move on from becoming the writers to the editors. And to me, there’s nothing wrong with that. It’s actually an upgrade for everybody. You go from having to do this GRU work now you’re the overseer of this stuff, right?
And it’s still, people are like, well, it’s going to threaten my job and I’m not going to have my job. And it’s like, yeah, but you still know your job. Right? Okay, well, if you’re hiring something, when you think about these ais, you’re working with GPT, if you’re working with a new GPT conversation, you’re thinking about it as a new person. It has zero context, has no idea what you want to do yet. You have to sit there and explain, Hey, I need this. I got to do this, whatever, and now let’s work on this together and perform this thing.
It’s a case where once it’s done doing that, because you had to teach it and because it wasn’t already prepared and already educated, and yes, it has knowledge of that stuff, but again, not the experience of executing it. So yeah, it can spit out a plan for you. But when I see time and time again is people who will say, oh yeah, give me a marketing strategy, inbound marketing strategy for this X, Y, Z type of company or whatever. And they might go a little deep with it and it’ll spit out the strategy and they’ll sit there, oh, well, I could have come up with that. That’s such basic stuff. And it’s like, yeah, I know, but you didn’t take it to the nth degree where you could have said instead like, Hey, I work for X, Y, Z company in this industry. What are some other, go online and do a little bit of research about my competitors and then come back and let me know what are some campaigns that I could run that would be counter to what my competitors are running that could increase my inbound leads. But you need that ideation. You have to think that through. That’s not something that GPT is going to do for you. That’s not something these agents are going to do for you. That kind of strategy is something that you can instill. You can actually tell it, Hey, this is how I would if I could do it, this is how I would do it, so why don’t you go do this thing?
Mark Lerner:
I mean with enough, see anything wrong with that eventually.
Nico Lafakis:
And then when it comes back, you’re the expert that’s got to check it out. You’re the one that’s got to review it and say, look man, I’ve used this stuff. I’ve used copilot for mobile and I’ve used it in the portal. It’s not flawless. I’ve used the company research agent. It’s not flawless, right? You still need to have someone behind the wheel just to check, just to cross the t’s, dot the i’s. Make sure that everything is what it’s supposed to be, especially when it comes to breeze intelligence. You don’t want to just flip the switch on that and all of a sudden your current customer data is overwritten with new data that you didn’t even know what it is. You didn’t know why it got overwritten. You’re trying to send out invoices and it’s going to a completely different address than it usually goes to. With great power comes great responsibility, doesn’t it? Okay. Well, someone’s got to hold the responsibility.
Mark Lerner:
Indeed. I think there’s a lot of really exciting stuff. Obviously we’re still in this second ending here, and so the early iterations of these things are, like you said, not flawless and we’ll probably have glaring issues with them and just given HubSpots cadence of launching stuff in the runup to inbound, I would imagine that we’re going to see a lot of fast follows and updates. One of the interesting things, bringing it back to the Breeze slash agents, AI versus Agent force, the thing that really stuck out for me is the cost associated with Agent Force. They’re charging as far as I can tell, $2 per, whoa, really? Yeah.
Nico Lafakis:
That’s so interesting.
Mark Lerner:
I don’t know if there’s some sort of sliding scale. I didn’t really get into that, but I mean that for a big company that ends up being six figures per month or something. So whereas what we’re talking about here, for the most part, it seems to be part of the package. Maybe there’s the intelligence with clear bid, it’s an additional thing, but it’s not usage based in the way that we’re talking about. It does seem like HubSpot wants this to be universal, democratized, whatever the word is, the kind of fluff word to use. Whereas $2 per conversation, if you have a hundred sales reps and each of them are having multiple conversations with the bot per day, I mean the numbers get absurd. It’s hard to see. It’s pretty
Nico Lafakis:
Counterintuitive.
Mark Lerner:
Yeah. It’s hard to see how that scales. So to me, that speaks volumes as to the different philosophies at play here, and it’ll be interesting to see. Does that in and of itself push more of those companies that you have that are coming from Salesforce into the HubSpot ecosystem?
Nico Lafakis:
I think it does because when you do things like this, it is very reflective of, and HubSpot was doing this. They’ve tamed it a little bit and they’re kind of pulling back. I hope they continue to do so because this is something that came from the gaming community and I wish it had never caught on, but it seems to be catching wildfire because OpenAI has been using it, quad’s been or Anthropic a lot of these other companies where, Hey, we’ll just release something and it’s not totally finished, but it’s finished enough for you to use it and then we’ll just keep working on it while you’re using it. So no one likes that, right? I’m sure everybody that bought the iPhone 16 now regrets it a little bit. They’re just like, oh, cool. Now I just have to wait four months in order to really use this phone. Hopefully they’ll get it faster, who knows? But it is a little bit the same with Salesforce where, okay, you’re going to give us this tool, but now you’re going to attach this pretty ridiculous charge to it. So it’s almost counterintuitive. Only The aspect of it that’s really cool is that it’s tied into the CM. Great, but if I could do what that does and I can do it outside of the CM, then yeah, I’m going to save the two bucks per usage. It’s so ridiculous. And are there guardrails on that so that, what if I keep the window open forever? Does that count as conversation not ending or something, right? I
Mark Lerner:
Don’t know. I haven’t had a chance to really get into the guts of it, but that is, I’ve read a few people talking about,
Nico Lafakis:
So prohibitive
Mark Lerner:
A, it’s very powerful. It’s very cool. But B, that cost structure is just, it’s hard to square that with, I don’t even know how that works in reality. So it’ll be interesting. I mean, we are in very interesting times.
Nico Lafakis:
What’s actually, what I find most interesting, and let’s just say this, okay, and I would love it if everybody walked away and really if it was the number one thing that you took away, it would be this. HubSpot talks to people about how these tools can help you, how they can enable what you’re doing, how they can help you to do again, more of what matters and less of the minutiae. Salesforce pushed their tools and quite literally said you could replace salespeople with this. So one of these companies is trying to be human-centric and human first and meaningful and putting thought behind what they do. And this other company is pretty much just sitting back, not really don’t care, don’t really what’s meaning don’t care, right? Let’s just kick it out the door so that in terms of like, am I going to see more companies shifting off of Salesforce, adding a bunch of awesome tools doesn’t change who you are, doesn’t change your methodology, doesn’t change your values, your core values. And that’s what is rocking Salesforce right now is that their core values are old and nobody likes them. Nobody really wants to do them anymore. Salesforce was designed for cold outreach. It doesn’t do well with inbound, right? Turns out it’s not so great at a BM. So it’s not really an alternative for that reason either. And it’s big and slow and the system itself and the company itself is big and slow. It’s hard to pivot.
I think they have a lot of issues currently. I think they do have room to grow. I think they will obviously stay. It’s not like they’re going to fall over or something like that in HubSpot’s going to take over tomorrow or something is not going anywhere. No, but it’s going to continue to be a slow bleed for Salesforce for quite some time.
Mark Lerner:
Yeah. If you look at, I haven’t looked since this week last week, but the thing that I think you and I see, which is more and more larger companies that would have never considered anything but Salesforce pop up now in the HubSpot ecosystem, that was reflected in the conversations we were having at our booth we were having with much, much larger companies than last year. But if you look at Salesforce as kind of revenue numbers, it doesn’t yet seem to be reflected. And maybe that’s because they’re able to squeeze more juice out of each customer and maybe that’ll change, I don’t know. But given that next year’s inbound is now going to be in San Francisco, I don’t know if it’ll be happening at the exact same time, that would be too crazy. So hopefully not.
Nico Lafakis:
I don’t know that the city’s going to be able to handle that at the same
Mark Lerner:
Time. Yeah, it’ll turn into a West side story. We’ll have street fights
Nico Lafakis:
Pretty much. Yeah. Oh my God. That’d be great though. Just the setup, not the actual fighting part of it.
Mark Lerner:
Right, right. Yeah. So just given that put on your Oracle hat, not the company, the Greek, Oracle,
What does next year’s inbound look like? I think the last time we spoke a year ago, I may have asked you what you had expected to come from this year’s inbound. So maybe we’ll do a cut and paste here. I don’t know if we have that technology, but I had love to hear your prediction about next year. Is it going to still be ai? Are we going to talk much more about an enterprise play because of where it is and whose backyard they’re playing in? What do you think is going to be in a year from now?
Nico Lafakis:
I think a year from now we are going to see an inbound that is way more CRM focused and probably don’t get scared, probably has less vendors. It’s not going to be quite as vendor heavy as it has been in the past. I think it’s going to be very, very community heavy. The one thing that I hear from even a lot of first timers from this last inbound is they didn’t realize how strong the community was. They didn’t realize how thick it was, just that it is. I tell people that it’s my Super Bowl. I love going to it because it’s an event and it’s not just like anything. There are so many people that you know meet all of us that you see on LinkedIn, we’re all there and you feel like you’re just one of the game, right? Because we’re all hubs spotters.
So that’s just the way it is. We all think the same way. And that whole like-minded aspect of it is really what draws a huge, it’s a huge draw too. So I think next time we see it, it is, it’s going to be, and this time we saw that too. We saw there was the creator stage, we saw that there was the large scale, I forget what it is, the showcase stage. Then there was a new creators, a new HubSpot media network thing that they put up, and then there was the hero stage and then you have the spot on top of it. So the community aspect of it is already kind of growing. I think it’s going to grow even more and it’s going to get even larger than that. I do love the fact that HubSpot had their own center in that where you could actually work around it. So next year I see it being looking very much like the sprocket, right? Where it’s got that center spot with all of HubSpot’s info and then just these break offs of the various community aspects that belong to HubSpot. And yes, there’s going to be third party vendors still, but I just think the competition’s going to become very stiff there. So I think that the vendors that you think
Mark Lerner:
Less vendors because HubSpot’s native features maybe
Nico Lafakis:
Different too. Yeah, that’s the thing. Maybe different vendors. I think maybe more agencies as than before and less third party tools. Definitely when it comes to third party tools, I think you’re going to see a lot less, not quite as many. I think you might end up seeing if you had three of each, I think you’ll end up seeing one or two of each instead.
Mark Lerner:
Do you think it’s because HubSpot’s going to cannibalize those features from within the platform and make those third party vendors obsolete or it’s the cost or is it just not, they’re not going to invite as many?
Nico Lafakis:
I think you’re going to see a lot of cannibalism. It would be, I mean look, it’s just one of those things where if you paid attention, if you’ve kind of follow this stuff. Sam Altman said this months ago, I want to say maybe it was even last year, and it was when people were starting to build devices based off of GPT and they were talking about Rabbit and they were talking about all these other things and he was saying that the problem is you’re trying to build on top of my platform. And when you do that, you’re kind of shooting yourself in the foot because we’re just going to swallow you up eventually. We may not have it today, but eventually we will. So your business model will fail at some point just overnight it will collapse. So this is what I’m talking about. Just hold onto your butts if you think that I’m full of it. Who wasn’t at inbound this year?
Who is partnered with HubSpot? Jasper?
Bye bye. Why do we need you? Right? What Jasper does is now built into HubSpot. So don’t need, you have brand voice now too. Don’t need, you could use Quad, can have quad Customs don’t need. You could do crazy custom GPTs. Seriously don’t need you. And why? Because Jasper’s built on top of what chat GPT?
Mark Lerner:
Yeah, that was Mistake
Nico Lafakis:
right? So it’s one of those things where if you’re using that API key to power what it is that you’re doing and you are using that to become a third party to somebody else who’s already using that API key, do something different really fast and figure out how to capitalize on what you do really fast. And the best advice I have there is since you are already running whatever it is, third party, let’s say your Typeform, you have data already. Your data is what is money. It’s not what you do anymore. What you do is going to be done by other things. What you know is going to be worth money because now knowledge is literally money. These models want data to train on and if you have it and they don’t, that’s money. And if I got to license it from you, I’m going to pay you. So data
Mark Lerner:
Is the oil of now
Nico Lafakis:
You got it and API keys are the currency, man. That’s how we do it now. So yeah, if you’re Typeform or somebody like that, you are sitting on mounds and mounds and mounds of characterization of how your forms work with all of your clients and things like that. You’re building that into a small language model because you don’t have to use large language models. That’s just sort of a nickname of all the big ones. You can build a small language model that encompasses all of that analytics data and that allows you to have those quantitative insights to now be able to say, Typeform does this based on all of our information that we have. You can now segment that bot into industry so people can query against it for like, oh, but my form does this, and what about in my industry? Should it be like this based on type form’s information or something like that. That becomes way more valuable as opposed to just our forms offer disappearing fields or something. Abilities don’t mean anything anymore. Just stop. Just stop. And I really do, I seriously hate to say this one because it is quite literally built on top of the platform. So I’m going to say this
Mark Lerner:
If you are, I think I hear, but let’s see
Nico Lafakis:
If you are an add-on to the platform, you should also very much consider that what you do might get built into the platform very soon. So I don’t know if you’re somebody that has a platform that builds assets out and can save and store assets and put them all over the platform, I don’t know that you’re going to make a whole lot of Buddy pretty soon, but we’ll see. But again, it’s a matter of finding out what makes you so important out of that. It is a shot in the foot for people who started their stuff up like a year ago, but even from a year ago, you might have some statistics. It’s one of those things where it’s like, okay, you didn’t think the data of your business was going to be useful. Now it is because you have something other people, it doesn’t matter what business you have, you have data other people are going to want. So it’s worthwhile to just look inward, do some self-reflection from a business corporate standpoint and look at what it is that you have and try to find the gold, right? That’s pretty much what everybody needs to do is just start panning their data for the gold.
Mark Lerner:
Well, I think we’ll have to leave it at that for now. Nico, thank you so much for taking the time. Before we close things up, tell the folks at home maybe where they can follow you, learn some more about what you do, what New Breed does.
Nico Lafakis:
Yeah, definitely. You can follow me on LinkedIn and you can check out New breed@newbreededrevenue.com. And we are very much in the mode this year and moving forward of unlocking meaningful growth. So it’s not just a slogan, it’s not even something that just, I say it is absolutely what is going to be most important. So definitely get ahold of me if you, I think October 1st, I’m going to be doing a demo for, let me see, I don’t want to get these two confused tomorrow I have a podcast from Mickey Bains that’s coming out October 1st. I have a demo that I’m doing for New Breed and then, or no, I’m sorry, a demo that I’m doing for connect.com. And then on October 8th is a demo on Breeze that I’m doing for New Breed. So if you want to learn about how to create a custom GPT in under 10 minutes, check out connect.com. And then if you want to learn everything there is to know about Breeze, check out new breed revenue.com and sign up for, register for our event on the 10th or on the 8th of October.
Mark Lerner:
Alright, well I’ll try to get some links in the show notes for those things. Nico, thank you so much again for you’re now in the two timers club and looking forward to speaking to you again soon. Thank you so much.
Nico Lafakis:
Yeah, no problem. See yah.
