(40) The AI Battle Between Incumbents and Startups- Matt Paige - YouTube https://www.youtube.com/watch?v=mIn-O1SijPk

Transcript: (00:00) Uh people will say AI is not going to take our jobs. It's just an enabler. I I almost feel like sometimes that's something we tell ourselves because it sounds safe and nice. I think it is going to to be honest take jobs, but I think it will open and create new things we never thought of. I hate this [Music] podcast with Jonathan [Music] with Good morning, good evening, good afternoon everyone. (00:35) It is I, Jonathan Carford, and this is the GTM AI podcast. I am excited because I have with me and my apologies for the hiccups. Um, I have with me Matt Paige, who I actually met because um, I'm on TikTok a lot and I do a lot of AI stuff, mostly just to consume different AI things. And again, my apologies for hiccups for whoever's listening, but randomly I found you when you had I think I told you two 300 followers, something like that. (00:57) And I've been following you ever since. And just recently, I've seen you blow up, which I love seeing when people are successful like that. And then I found you on LinkedIn. I was like, "Hey, he's got you're a VP of marketing for a for an AI consulting and product development firm." I was like, "This is awesome. (01:11) " So anyways, I didn't know that before your TikTok because I saw your updates there and you know you talk about a lot of different tools and capabilities and so I was like I've got to talk to this guy. So first thank you. Welcome to the show. How are you my friend? Yeah appreciate it Jonathan. Excited to be on. (01:26) Been good chatting with you so far and yeah it's cool that you found me in that mechanism which led to us being here today. That's the whole purpose of it right awareness building and all that. Yeah I love it. So you're Hatchworks obviously you're now Oh good you're a podcast host as well advisor for Noble. um you've been in a trainer and you've been in marketing for a while like you've been in this go to market AI world for a while which I love but what I like to start out with guests is the non-res version of you so like tell me a little more about you what's what's kind (01:53) of your journey what made you get into a and what kind of led you to where you are now yeah I got a weird journey to be honest I started more on the data side of things and I used to work with a team of data scientists and it blew my mind the stuff they were doing like this was before like you you go ask chat GBT a question. (02:11) They're like building models and all this like statistical stuff. So, hey, I had to go like immerse myself and try to learn all these things so I didn't sound stupid. That was one part of it. But I became the guy that was making sense of the data and presenting that to executives because the data scientists would come with slides and slides of all these p values and all these things and the the executives eyes would just glaze over. (02:32) So that's where I got immersed into AI and then I made this kind of winding journey into product did some product marketing did true product development and then strategy and then as a function our organization's around 200ish people very AI focused and uh leveraging AI throughout the whole software development life cycle when we build we we're not big enough to have a separate marketing department so I have marketing too and I've learned a ton on the marketing side positioning all of that kind of fun stuff and the seminal moment (03:02) I think when chat GBT came out it just it was like this epiphany of all this stuff that is possible and I've just incrementally been playing with it every single day and now I'm actually building stuff on my own which is cool to see because I've always worked with developers I understand how code works and all these things but now I can actually do it which is just this huge democratizing enabling empowering thing and I don't think a lot of people yet realize how much you can actually do with AI. (03:36) Yeah, I love it. It's kind of cool to hear everyone's stories when you have these things like co that everyone experienced or chatb like it just blows up your mind with stuff. It's awesome. Hatch, tell me about Hatchworks. Sorry, tell me more about Hatchworks. What are you guys doing? Yeah, Hatchworks AI. (03:50) So if you think about it, we naturally were a product development company and we fully we've always had an AI practice, but in the past few years, we've dedicated our entire business to this, building AI native solutions for our clients. And you know, we think of ourselves as like an AI and data transformation partner, really helping you realize the value of AI through the power of your data because data is the core piece to actually take advantage of AI, you need data. (04:15) So that's the core element there. And it one thing we like to talk about is AI is easy to do, but it's hard to do well, right? It's easy for anybody to go talk to chat GBT, write an email, do this, do that. But to actually build a system within your business, integrate with your business, start leveraging agents, automating things, entire job functions, that's not as easy to do. (04:38) So that's kind where we come in and are working on a lot of cool projects with a lot of cool companies, uh, helping them leverage AI. Are you doing building out, sorry about that, building out native AI platforms for people or like you have existing SAS products or whatever companies that you're bolting on AI too or both? Tell me more about it. (04:55) It's a bit of both. It goes from okay, we want to build this completely new native AI thing, right? Hell, we even had this, it's funny, this like YouTube personality that's wanting to build this whole platform. So it's like it spans from that to working with huge Fortune 500 companies that are wanting to create operational efficiency by automating workflows internally to the LA the latter of what you just said hey we want to integrate AI into our existing solution. So it spans all of that. (05:27) I think one interesting thing though you know I've been thinking about this a lot. there's almost this huge inflection point that's about to happen right between incumbents and new entrance right in the market and I think the incumbents are going to be experiencing this innovators dilemma like they've never experienced before because new entrance they're not going to go build this huge organization they have natural constraints that are going to push them to use AI and they're going to have AI integrated in their business from the (05:55) very beginning I think the one thing that incumbents have on their side is they're sitting on piles and piles of data, right? So, that's the one advantage have that's proprietary to them. So, it's going to be really interesting to see how this all plays out the battle, you know, shift for battle for power between the these different companies over time. (06:15) I want to get your take on something because I'm curious what you think. So, example would be Salesforce. I've used Salesforce for years and years and the original purpose of it was obviously to capture information about customers inside of some sort of management system that everyone could access and deal with, right? Obviously 10 years ago you needed that but now with a AI and rag and all these other type of systems you have the ability of capturing things and put into some sort of snowflake database and then to query it using rag or (06:38) prompting or whatever. So almost I won't say it totally but almost eliminates the need for a CRM. So a core structure Salesforce then become becomes something that used to be a non-negotiable is now like we could do without it. I'm not saying you I'm not suggesting people do that. (06:54) I'm just saying that the world we exist in that's a possibility. And what I'm and the reason I say this because like you said incumbents I think it's one reason why Salesforce went the agent force direction is because they know this is they got to develop other products to keep relevant make sure people stay with them. So do you see that kind of same theme of where like you said democratizing the data and what people are using that's going to come to this world where people can either build inhouse or have some other platform that can do that form like what do you see (07:20) happening in that kind of vein that used to be non-negotiable is now on the table. Yeah, that's a perfect example too, the whole Salesforce and Benoff thing and shout out to Ben off for making this strategic pivot and shift. I think it's going to be hard to execute on to be honest. Like I've heard people that are in the beta program and it's messy right now because you have this just whole legacy not only for codebase all of this but your workflows, your entire business, all the humans involved to support this (07:48) thing that's been created that now is potentially getting fully disrupted. And in that same vein, HubSpot going through the same thing, but Darmsh, the CTO over there, he's a great person to follow, but he started his own like side thing called Agent.ai. He's built this whole network of agents. Think of it like a marketplace for agents. (08:11) It's just interesting to see where that's going. And then to your point, the holy grail, maybe we'll never have to update a CRM again in our life once these new native solutions come out. And I can't think of the name of it, but there are companies that are to the point early earlier the incumbents versus the new entrance. (08:27) They're trying to solve from this starting foundationally with AI from the beginning versus the very hard journey that Salesforce is going to have to take to shift their whole business. Yeah. But again your your point though they have mountains of data that's proprietary that no one has access to. They have their own version of stuff that in or a newbie startup won't ever have for in talking about the whole CRM thing like I work for a company the momentum that's what they do and there's plenty of other competitors do the same thing but it's (08:54) more more of like it's more of understanding because momentum doesn't replace Salesforce it enhances enriches Salesforce but there there's a world where that a tech like that could have some ability to capture information and then to replace a system of record if it was able to be cate categorized in the same way that Salesforce is with accounts and field records and all that stuff which is a complicated beast to do. (09:14) Yeah. Crazy. Well, that's the thing too. The shift to change from a system of record, whether it's CRM or HIS or things like that. Those are they're very difficult because humans don't want to change. We don't like change natively. But it's like once you reach that tipping point where people are like, "Oh, this is the better way. (09:35) " Then it's like just avalanche, right? So, I think that's the it'll be very hard for people to switch. But once that light bulb goes off, you know, if Salesforce has time to to build the solution and be that option before that hits, but it's going to happen at some point. Yeah, I think it'll be interesting to see if they become a blockbuster or if they become the just Salforce 5. (09:56) 0 or whatever the next iteration is because I I Mark's a smart dude. I don't think it's going to just fall by the wayside. No, I agree. If you look at the moves he's making strategically, like it's it's what he should be doing now. The question is, can you go execute on it? That's the primary question. (10:10) But yeah, I totally agree. I think the fact that he's a founder, he's still in there. It's his baby, so he's still going to live. If it was me, I wouldn't just sit down and die. Like, bring it on. Let's do this. So, I get it. Exactly. Exactly. Tell me this. When you have companies come in, let's just let's start with the people who have an existing something. (10:27) And they say, "Hey, let's build an AI platform." But they come to you and they say, "We don't know." Or I don't know how often they do that, but they come in with an idea or they say, "You tell us what we can do." What happens? usually man we see all flavors and that's typically we so we have a AI roadmap and ROI workshop that's the whole purpose of that and it's we have to tailor it to or I guess the better way to say it is meet the client where they are in a sense because a lot of times I'm about to talk with somebody right (10:54) after this they have 80 use cases so part that part of that is like okay well okay how do we find the first couple that you need to start with and there's standard approaches with like okay let's look at highest value versus effort but then you can start to layer on other things like okay what's the risk associated with this that's a big component with AI like you look at the example of okay is somebody's life at jeopardy when you talk about like medical and health related issues or is it ad copy and okay not a big deal right (11:22) so there's other layers you can start to put on there so we go from that all the way to here's the idea that I have and I need to know the best way to bring it to market create it all those things so there's that angle and then even had folks that just want to upskill their organization on the other side. (11:41) So, it really spans because it's so new and emergent, there's this wide range of where folks are. Yeah, it's it's fascinating. Think of all the different flexible ways you can do it. But what do you think when you work with companies and then you implement whatever you implement through the work workshop, do you find pretty common categories of things that most people do that make a big difference? Like what what are you seeing as a as patterns? Yeah, I think a common pattern is repeat or sorry repetitive mundane tasks that they just want to automate (12:12) and I think automation in a legacy way was very much okay I got to define every instance or very option of this workflow very deterministic in nature with AI you almost give it just guidance there's a really cool tool I was playing with the other day called bland bl a n d their whole thing is like they help to get rid of the bland task because they they take it on. (12:38) But it's like this voice agent and they have this really novel approach. It's almost like a decision tree for the agent, but you're not saying explicitly say this when this happens or do this when that happens, but it's just giving it rough kind of guard rails and guidance and it has an LLM model built into it. So, it just needs that slight direction and it can take it from there. (12:59) So, I think those are interesting use cases. a lot of workflow based stuff um that people are just wanting to automate and then drives cost out of the business is a big part of it. So a lot of people I think start there because it's you're more risk averse you start there versus let me go bring some new thing to market. So I have a hypothetical not hypothetical I'm more of a I'm going to give you some context I want to get your thoughts on it based on what you guys are seeing with businesses. (13:23) So I always compare when I teach about AI that there is the automation, there's workflow, there's process, there's efficiency play, but there's also this idea of getting data or a view on data that was already there but never could be tracked before. An example would be like I wear this order ring to track some of my health stuff and it gives me analytics in the phone app that I never had before. (13:43) So then I have this angle and view into my own health that that was already there but I didn't have access to the data. I feel like the AI does the same thing in a bajillion different ways to where you have access to data to hopefully help someone have a better decision. And I always ask this question to people is like how do you measure the ROI of a better made decision with data you never had before when you have nothing to compare it against? Yeah. (14:03) What kind of so two questions would be what kind of data or again categories of data you think people are now capturing as a whole with business they didn't before and then what's the value of that? Yeah. So I think a couple things and the point you mentioned too on the ROI question that's the thing that's tough too because like businesses are forced to prove out the value of it but sometimes it's very difficult when it is something new at a merchant and you don't have all the factors to calculate it. Yeah but no you brought up an (14:33) interesting point and that was at a data bricks conference which by the way that if they tend to be more for the enterprise but that platform's amazing doing exactly what you did. It combines a data warehouse and a data lake. But in the past, and I experienced this back when I was working with those data scientists, I did a lot of reporting. (14:51) And it got to the point where you'd have to put in a ticket, get a request, a week later, you get the response back, that would trigger another question, you put in another ticket, you send a request, and it's this painful like process. And that's been democratized a because you can ask questions of the data. (15:08) And then what does that then enable when you can conversationally talk with data? it gets to exactly what you're saying have okay this this foundational metadata or aggregated data but now I can help make sense of it and this was probably like a year I don't know a year or so ago is when the whole custom GPT thing came out with chat GPT I built one that did acted as a data scientist doing something called K means clustering so I just I had a like Airbnb data set I dumped the data set in there I said provide logical groups within this thing (15:39) so it's like think of it as like marketing marketing segmentation. So, it defined like four different groups, the attributes about them, and then it defined a marketing strategy to help increase your, you know, what is it? Price per listing or you could trigger what the goal was. You know, do I want to increase my bookings? Do I want to increase my rating? All these different things, and it would group these things on things that you may not know are related patterns, but it's seeing it in the data. So, that's it. It it just it (16:07) enables a ton. And then you add in the democratization effect where anybody can ask a question of data and it's going to change how we operate in a big way. What I know the ROI is a hard question to answer. I know you don't it's hard to have an answer for that question, but what I don't know if it's results, but what impact do you see h happens as a result of having access to the data that's promptable and access accessible in any ways that you never had before? I don't know if you have example internally you could use or with a (16:35) customer. I'd love to hear what you have. There's there's one part of just the human that does that today that no longer needs to do that. And I think there's this whole people say AI is not going to take our jobs. It's just an enabler. I I almost feel like sometimes that's something we tell ourselves because it sounds safe and nice. (16:55) I think it is going to be honest take jobs, but I think it will open and create new things we never thought of. I think there's that fact. So, is there a human doing this that AI would do? That's a component of the equation. I think the time it takes to get to an answer, that's something you could put into the calculation. (17:17) And then there's this just nebulous variable of making better decisions and how do you quantify that? That's the harder one, but you could throw a number on it. Ask, go ask Chad TBT to help you with it. So, that's the meta thing too as well. And I've started to because it's a change in your habit. Whenever you're thinking through these type of questions, a lot of times we just think in our own head and do it. (17:38) Go have that conversation with pick your model. It doesn't really matter. Maybe use a reasoning model for a certain type of question versus a different one. But I found a ton of value in just talking through things with AI as if it's like a partner on something you're trying to work on. Yeah. (17:58) Especially like I'm sure you've played with deep research both chatb and gemini but like I am loving that because that is more of a thought partner for not that chbby wasn't but just is more thorough. I'm loving it. So I haven't dropped the 200 bucks a month for the deep research yet. I've seen a lot of people using it. It's good. Yeah. I'm wondering it's interesting. (18:16) It's like what are the use cases where you'll use that on a regular basis? And then to the point earlier, maybe there's things that we never even thought of that we would do to where we would leverage that. And then they have the whole operator thing that's tied in with their $200 a month plan. (18:32) But the cool thing is like there's open source options for this something called browser use on GitHub. And this is the cool thing too. If you're sitting there a marketing professional or go to market leader, go to GitHub, go ask Chatbt how you can actually use this. It'll walk you through. It'll probably point you to a tool like Lovable or Bolt or Cursor or whatnot and just go create it. (18:54) Talk with the agent to help you do it. Like it's a cool just side project to do. And that's kind of the rabbit hole I've been going down as of late is like actually building things with AI. Yeah, I've been having I made a Chrome extension with a cloud actually and I double checked it in cursor, but it's this dumb little thing that I use for my own email thing, but it works. (19:15) I'm like I would never in a million years be able to code a Chrome extension and I pro probably could, but I don't really care to learn how to code a That's the thing. But yeah, like I've always had this feeling where I want to do this. I've tried many times but I never get over that hump because to truly become like a skilled software engineer back in the day. (19:36) You you had to really practice at it. Now when you have AI you have that partner that can get you over that hump. But same I think there's this other effect too where in the past if you wanted to a product like actually develop something you needed a market to sell it to to justify the investment to build the thing. Right now you don't. (19:56) You could build for an audience of one being yourself and it's okay because it takes you an hour to go build it. It's like changes this whole shift I think in like how we think about SAS buildver versus buy all these things. Like my wife and her friend were exchanging clothes from each other's closet because they were tired of their stuff and they were losing track of whose stuff was what. (20:19) So, I just built this thing with AI where they could snap a picture of the thing, say who borrowed it, and they were able to track it there. And it's not necessarily something that may be, you know, marketable and you could sell, but I built it in a day or two. So, it's like stuff like that you can just start to build on your own. (20:37) I love I had the other day that someone talked about the same concept of having a team of agents, but they were self-built, not pre-made agents, but like you said, it's like some sort of software or agent or something that does the thing for me only. And the value inside of that is like awesome because then one, it's a competitive advantage and two, it's just in my workflow of what I do, how I think, which is pretty crazy. (20:55) Yeah. And I think the thing about agents too, real quick, go check out the Google white paper on agents. just Google it and find it and I can shoot it over if you want to put it in the show notes but they talk about three components right there's and this I think makes agents they just do a great do job of dumbing it down so there's three components the LLM which think about that as like the brain of the agent there's the orchestration layer so that gives it context for how it should reason and think about what's (21:23) how should I approach my next step it's like the logic layer and then there's giving it access to tools to where it can actually go do something those three things. Yep. Create an agent. Like the other day I had I was playing with a tool called Do Browser. I had it buy me toilet paper just on Amazon for fun. (21:40) Like it's just like just playing around with it, but it can like actually go do things now, which is really cool. That's awesome. Yeah, it's crazy to think about this stuff. And you said earlier about jobs like I'm on camp with you where forever people like it's not AI that's going to replace you. It's the people using AI. (21:54) I'm like that's a bunch of bull crap. It's going to be AI doing it. But I feel like we're in this place kind of like the industrial revolution where we have all these horseshoes. We're horseshoeing with office whatever and we would never imagine a world where you have 5 million people employed by the automobile industry. (22:09) But that's where we are. Exactly. It's like we're in this shift that's just it's instead of automating and speeding up the physical work, it's now the mental work and we just can't conceptualize what it's crazy. Well, tell me this and we have I think we will see the same thing too. And I think I do think though the people that are using AI will be the ones that lead that next wave of what you were just talking about. (22:30) So I think there is it's not like a wo is me just go take your ball and go home and life sucks, right? So I think you still got to be on the front edge of that and it's not as hard as people think. Just go start playing with stuff on a daily basis. Habits build over time. Uh but I totally agree with you. Old jobs will go away, new ones will emerge and just put yourself on the front end of that so you're not you have choices. Yeah. (22:52) In the last five minutes we got, um, tell me this. What's the best way for people to get in touch with you? Yes, LinkedIn is the best way to find me. But like you said, I'm on Tik Tok. I started posting on YouTube. I basically just started pushing out the things I do across all the different channels, especially with the Tik Tok band that went out. (23:12) So, I was like, "Oh, this may go away." So, yeah, LinkedIn's probably the easiest way. Or just find us. Go to hatchworks uh.com and you can find us there and see what we do. Love to chat with anybody that's trying to build something in AI or anything around that as well. Love it, man. Last question then, which would be what are I know there's the classic chat, Gemini, Claude. (23:32) So, sorry all those other tools, but what are the ones that you're geeking out on? They're like, "Oo, this is fun." That's like newer or newer to you. What are you geeking on? Yeah, the area that I'm deep into right now is building with AI. So, I'm going deep into these AIDS, which are like integrated development environments. (23:49) It's where you build software, right? Yeah. For anybody that's just starting on this, go play with lovable or bolt. And it's very easy. It's very focused on the user just using prompts to build something. Replet, I would say, is the next level where it's a more of an integrated fully functional IDE. And then cursor is the one I'm using mostly now. (24:07) That one is a fork of VS Code, which is what typical software engineers use. So if you're an engineer, that's probably your best bet. So those I'm going deep into. But then I'm just like playing with a bunch of random stuff. I think I mentioned browser use which is like operator open ais operator but it's open source so you don't have to pay 200 bucks a month to use it. (24:27) Uh there's all gland AI is really cool. That's like the voice agent. I don't if you've heard of Hume. That's another neat one. They can read your facial expressions. Uh but there's just a laundry list of cool things. And then Hugging Face just came out with their like, you know, quote unquote app store for AI. (24:46) So if you want to go find a tool, that's a great resource. They call it Hugging Face Studio. So that's a few things. That's awesome. Have you seen yet a place where someone can make an agent team that someone else doesn't have to rebuild, but they could sell that team's abilities and all the user has to do is plug in integrations, they're good to go. (25:04) Have you seen that yet? Yeah, I think that's what Darm mesh is attempting to build to be honest because it has this whole network of people where you can become a builder and you're building agents and then people can use those agents. I don't know if elit leaders really emerged as the go-to agent person. It's funny. (25:21) I saw a leaked video the other day where OpenAI was demoing their sales agent. Like it's an actual thing that's not been released yet. So I could see them starting to make more verticalized agent. Yeah. So I think it's very emer like do browser is a tool I mentioned. It's like 25 bucks a month and it allows you to do things and you can customize agents and then there's like the open source options too. (25:43) Yeah. I joined agent when and I've actually so I it's funny when I had I had a question for the group and Daresh actually responded to me. It wasn't it was on a public forum but I was like yeah it was just cool. He's like the coolest guy super smart down to earth really really and it's funny cuz like I think most people when think about agents think about this advanced thing and sometimes it's just this simple chat to blog or chat to one output type of agent play because that's a lot. (26:09) What currently he has on agent.ai is that type of workflow. The future of having a bunch of those mixed together in a team that does these things in sequence is where it's going to get really powerful. Well, that's the thing. I think it's going to be people think of like us working with an agent directly. (26:22) I think it'll be agents working with agents asynchronously, synchronously, all those type of things. Like right now, we build a lot of that custom because there's just not that like easy button out there. So, I guess that's the other angle with Hatchworks AI. if you want to build an agent for a use case. (26:38) We do those things in a custom way just because there's not that offtheshelf option that really is out there yet. That's awesome. Thank you. I know on I'll say I'll put the link in the in the show notes, but Matt Page P a G E68 on Tik Tok and obviously on LinkedIn I'll make sure we have all the stuff, but thank you so much, man. (26:56) It's been a pleasure talking to you and I'll be liking your stuff on TikTok as you go. So, keep it up. Awesome. Yeah, I appreciate it, Jonathan. Good chat. Yeah. Thanks, man. Thanks for listening to the GTMAI podcast. Please follow, subscribe, like, comment, and come back for more.