Episode Transcript
[00:00:00] Speaker A: Foreign.
[00:00:25] Speaker B: Hello, and welcome to another episode of Street Level Marketing. I'm your host, Mark Lamplugh, and we have a great show for you. Planned this this afternoon.
I want to get in talking about AI agents and how they can help out with your marketing.
This has kind of been a fascinating subject for me over the last couple weeks because it's something I'm rolling out here, you know, where I work full time, and I'm learning about it. And I ran into our guests Instagram. His name's Chase Hannigan, and he runs a forum or like a. Like a membership to help with AI, I guess, implementation. And I was reading a lot of stuff that he does, and he really seems to know what he's talking about. So I wanted to bring him on the show, you know, to kind of see what we can pick his brain and see what we can learn on implementing these agents for your business.
So welcome to our guest, Chase Hannigan. Chase, thanks for coming on the show.
[00:01:26] Speaker C: Yeah, of course, Mark. Thanks for having me.
[00:01:28] Speaker B: So, Chase, can you explain to the viewers, like, what an agent is in, like, plain language?
[00:01:35] Speaker C: Yeah. So I think the easiest way to think about it is imagine if Chat GPT could actually do things for you.
[00:01:42] Speaker A: Right.
[00:01:42] Speaker C: I think we use ChatGPT a lot as almost like a glorified Google search.
[00:01:46] Speaker A: Right.
[00:01:46] Speaker C: We just ask it questions. But imagine if Chat GPT could actually, like, send emails for you, do LinkedIn posts for you, respond to comments on LinkedIn for you, schedule meetings for you. So AI agents essentially become like miniature digital assistants for you that can actually execute tasks.
[00:02:04] Speaker B: Got it. Now, what are some of the core components that make an AI agent function autonomously?
[00:02:10] Speaker C: Well, I think the biggest thing is the fact that it's able to take in data proactively. So imagine if I had, like, an AI agent that was hooked up to your email and could, like, automatically qualify leads that were coming in and then respond to them, tag them, or let a human know that, hey, this one's super important, or I need an actual human to come here and respond. So the ability to proactively grab information and then respond to the information based on whatever parameters or logic you give it, it has what's called an LLM, like a large language model. Again, think of, like, ChatGPT running in the background as, like, the brain, and you give it rules and you give it parameters for how you want it to respond and logically, you know, deal with certain situations.
[00:02:53] Speaker B: Now, a lot of these agents, they're kind of plugged into, you know, like, chat GPT and, and other large language models.
Can you give like a real world example of like an AI agent in action?
[00:03:10] Speaker C: Yeah, I think a really good one is like Klarna. So Klarna is like a big bank. They first, first started using AI as just like a customer assistant type deal. So imagine something that just kind of responded based on frequently asked questions, but they stepped it up and they actually gave it the ability to be an agent so it could actually do things when customers responded to it.
[00:03:30] Speaker A: Right.
[00:03:30] Speaker C: And actually make changes to their accounts and solve problems. And I think the quote they gave it was doing the equivalent of like 700 actual human agents workload.
So that's a huge company worth billions of dollars that's actually using agents for real in real life scenarios.
[00:03:49] Speaker B: How do these agents learn and adapt to new tasks over time? Once you start giving it a task.
[00:03:55] Speaker C: To do well, I think that's where you as the human come in because you'll see how it responds to certain situations. That's when you can come in and say, hey, I don't really like how you responded to that task. Let's change up your internal logic or your rules. You can give it rolling feedback of how you think it's performing and then actually go in and make changes.
[00:04:17] Speaker B: Got it.
What is the role of the large language models empowering these AI agents?
[00:04:23] Speaker C: So the large language models are the brain of the operations, right? So that's what actually allows us to talk to these essentially machine algorithms like humans through natural language. So it's the brains of the operations, it's taking in your data, it's taking in what you're asking for and then using its own internal logic to then like do machine tasks, like to actually send a post to LinkedIn, to actually read the email. It's the brain that's actually telling all like the lower level, like coding stuff, all the scary coding stuff in the background. The brain is directing all that. That's where the large language model comes in.
[00:04:56] Speaker B: So are these agents, are they tied to just one platform or do they.
[00:05:01] Speaker C: Use multiple, they can use multiple platforms. Usually you have like a singular like large language model brain in the background. And that could be something like OpenAI's GPT, like that's what runs ChatGPT or something like Google's Gemini. That could be any one of those you could use depending on the task and then you can connect it to a bunch of different tools and that can be very wide ranging.
[00:05:23] Speaker A: Right.
[00:05:23] Speaker C: Stuff like go high level stuff like LinkedIn that we talked about, your email Google Docs, all sorts of stuff.
[00:05:31] Speaker B: Yeah. And coming up, the next question as a, you know, as it applies to small businesses and it's something I'm kind of in the process and what kind of drove me down this rabbit hole is I was research, I'm in the automotive business, marketing and I was researching DMS systems to start communicating with potential customers.
And there's a, there's a platform out there that ties in all of your data and your, your CRM and, and creates agents to start marketing to these, you know, the customers based on the data that they is provided through your websites and what they're looking at and all. Well, they were like 20, $30,000 a month to roll these things out and into our dealerships. And I'm thinking why can't I do this myself, you know, what they're doing myself. So now I'm in the process of kind of using Google cloud services to, you know, tie everything and, and then roll out our own agents.
So are the agents assess, you know, accessible to small businesses or are they just for like bigger companies like myself?
[00:06:43] Speaker C: I think they're totally accessible to small businesses. I think there's very much levels to it. You know, I think there are certain agents that because of the data they need to have access to is kind of behind very, very expensive paywalls. But I think for small businesses, almost disproportionately, I think they can get a ton of value out of this stuff. And oftentimes it's very, very low cost. I mean take kind of like the email agent for example, right? Imagine you had someone, an agent that's able to read all your emails, qualify all the leads that come in, tag them, respond to them based on your internal documentation, your rules, your logic actually sound like you and then bring up kind of like edge cases to you as appropriate.
[00:07:24] Speaker A: Right.
[00:07:24] Speaker C: That's the equivalent of you know, hiring like an intern to sit at your computer 24, seven responding to these things. And instead you can have an agent, a fairly simple one that for, I don't know, $10 a month essentially probably your cost could do that.
[00:07:38] Speaker A: Right.
[00:07:38] Speaker C: And I think there's a ton of value to be derived by kind of the simpler lower level agents that won't break the bank for smaller businesses.
[00:07:45] Speaker B: Yeah, that makes sense. What's the biggest misconception people have like right now about AI agents?
[00:07:51] Speaker C: I think on one hand people think they can do anything, right? And this is still like a very, very early technology. I mean if you remember chat, GPT hit the scene. What Two, two and a half years ago.
[00:08:03] Speaker B: Yeah.
[00:08:04] Speaker C: So this is super new. And while they can do a lot, they can't do everything. And sometimes it's not obvious what they can and can't do.
So people think they can do way more than they can on one hand. And then the second thing is, I think people think it's just not something, something that's, that they can do.
[00:08:19] Speaker A: Right.
[00:08:20] Speaker C: They're not technical, they're not a developer, they're not in any sort of, like, tech space whatsoever. And so the term AI agent seems well beyond anything they can implement, especially on their own. And I think that's not the case. I think there's a ton of, like, stuff you can do even if you aren't technical. And you keep it small because the space is getting better and better for people who aren't technical to come in here and actually start implementing these sorts of solutions.
[00:08:46] Speaker B: Yeah, and it's, it really is. Even with chat gbt, how, how fast it's progressed. And I remember it just seems like not even a year ago when you would have it create a picture, it would create a picture with, like, six fingers and, you know, be all now, I mean, takes you and puts you in, you know, it's amazing how fast it's, you know.
[00:09:09] Speaker C: Yeah, it's, it's wild. I mean, the, the iconic thing is when they had like, hey, here's AI video. And it was like Will Smith eating spaghetti. And it was just like a complete mess. And then just yesterday, Google announced a ton of different new, like, AI upgrades they did, including stuff with video. And now their AI can not only generate amazing video and they use, like, some spaghetti example to show it again. But now it actually has audio, too. So now you're starting to blend two of these things together and it's getting to the point where it's like, all right, this is just two and a half years of this sort of developing. Obviously this was years and years in the making before everyone really thought.
But it makes you think, what is, what is it going to look like in three years, four years, five years? And how much easier is it going to be for people who aren't technical in the space to do it? Because that's, I think, the big gains.
[00:09:53] Speaker B: Yeah. Now, how did the agents balance autonomy with human oversight?
[00:10:00] Speaker C: I think it kind of just depends on how much oversight you want to give it. Luckily, when these things are being built and when you're using them, you can almost always insert some sort of, like, human in the loop is the term that's often used into the picture.
Take again that email agent example we did. We could have it. So where whenever it's responding to someone, it's just a draft.
[00:10:20] Speaker A: Right.
[00:10:21] Speaker C: We can have it. So we never actually send out an email that we as a human actually look at and approve before something gets sent off with our name and signature at the bottom. Right. So and at any point along the chain you can put a human in there to actually review what's happening.
[00:10:37] Speaker B: What's the one thing you wish everyone would understand about AI agents?
[00:10:42] Speaker C: That they don't have to be super complicated and that even if you aren't technical, you can start using them.
[00:10:47] Speaker A: Right.
[00:10:48] Speaker C: There's really, really small, simple applications that I think people can get a ton of use for and I think marketing is honestly one of the biggest ones.
[00:10:55] Speaker A: Right.
[00:10:55] Speaker C: The idea of having some sort of agent that can research a topic for you, generate a draft or a final product and actually post it for you or at the very least get you right one step before posting.
[00:11:07] Speaker A: Right.
[00:11:07] Speaker C: I think could save you a ton of time as like a practitioner.
[00:11:11] Speaker A: Right.
[00:11:11] Speaker C: Because it alleviates all like kind of that low level, repetitive, boring work and frees you up to actually think about the creative stuff and like those high leverage stuff that's actually going to get you money.
[00:11:21] Speaker B: Yeah, yeah, it's in marketing I think the big fear is, you know, it's going to replace jobs and, and you know, and in essence I think the agent, the marketing agent, agency space is going to change in a sense that you would have these big contracts, you might have like five or six, you know, as a smaller agency that were very high costs and I think now it's going to be able, you're going to be able to help more businesses at a lower cost because you're going to be able to streamline a lot of the workload.
[00:11:53] Speaker C: Yeah, yeah, I totally agree. I think we're probably pretty far from AI taking a huge amount of jobs. I mean, who's to say in five, 10 years what that will look like? But I think in the short term it's going to be more like if you're not using AI, you're just going to be losing market share and profit to the guy who is.
[00:12:10] Speaker A: Right.
[00:12:10] Speaker C: It's going to be who's leveraging it the best. Because if you just ignore it and put your head in the sand, you're going to be just taking on all this work for yourself when you really don't have to.
[00:12:19] Speaker B: Well, we're out of time here. We're going to go to a commercial break, you know. So we'll be back with Chase, and when we come back, I want to talk about, you know, AI agents in the new workforce when we come back with a word from our sponsors.
Welcome back to Street Level Marketing with your host, Mark Lamplugh. I'm here with Chase Hannigan and we're talking about AI agents. And I really want to talk about the agents now and how they're reshaping business operations and streamlined processes.
How are the AI agents changing the way businesses operate today?
[00:13:04] Speaker C: Well, I think the biggest thing they're doing is they're freeing up time that was spent on low level repetitive work, especially repetitive work that still required a little bit amount of nuance. I think stuff like lead qualification is a great example.
[00:13:19] Speaker A: Right.
[00:13:19] Speaker C: Like it's always not clear what qualifies a lead in a lot of cases. And oftentimes can be repetitive tasks that require someone to just kind of sit there and do it. And instead now you can offload that work to AI.
[00:13:30] Speaker A: Right?
[00:13:30] Speaker C: It can repetitively do it. You can give it the parameters for what qualifies a lead and what doesn't. And so again, this lets you use your human capital in a way that I think is a lot more useful, like doing creative work, doing that high leverage stuff like relationship building, things like let the AI take care of kind of the low level repetitive drone work that you don't want to spend time doing.
[00:13:51] Speaker B: Yeah, yeah. And even the, the agents, the agents that are, you know, answering calls. I did a test call with an AI agent a couple days ago and it's leaps and bounds to where it was a year ago. I mean, it did a really good job. I mean, I could tell it was an agent probably because I'm more like age in depth into, you know, what it is and, and researching and stuff. But I think the, the general or average person probably wouldn't have known, you know.
[00:14:25] Speaker C: Yeah, they're getting the, the call and the audio especially is getting really, really good and nuanced. It'll kind of have like those natural pauses. It does not sound like a robot. And like I said, it's becoming. I think the customer service thing alongside marketing is probably the biggest one.
[00:14:37] Speaker A: Right.
[00:14:38] Speaker C: Like the Klarna example is a big one.
And I think ones like bank of America doing the same thing. Right. Like it's, these agents are actually doing things and completing tasks that otherwise the human would have to do all of it.
[00:14:50] Speaker B: Do you have any examples? You just brought up bank of America of a business that has been transformed by AI agents.
[00:14:58] Speaker C: Yeah, I think bank of America is just like the client example where you had these like chatbots essentially that then got upgraded to full blown agents. But I mean there's other companies that are essentially just selling these agents.
[00:15:11] Speaker A: Right.
[00:15:11] Speaker C: I think Cursor is a great example. It's essentially a company that deals with like AI coding and it could be argued their entire product is an AI agent and they scale to multi billions of dollars with a team that I think at one point when they were still like 20, 30, 40 people, like very small team. But because of AI and how they leverage it, you're able to do the work of what would have taken someone like Microsoft, you know, teams of tens of thousands.
[00:15:38] Speaker B: Yeah, you know, for some of the viewers that maybe have smaller businesses out there that don't really know a lot of the things that these agents can do. Can you share, you know, some of the tasks that AI agents are best suited for, to handle for a business?
[00:15:55] Speaker C: I think the, anything that requires you to get like a bunch of unstructured data and, and then structure it. So think of again marketing and creating some sort of written content. It can take a ton of time to like do the proper research on a topic to create something. So you might think, hey, I can already do that with ChatGPT already. But I think something like ChatGPT projects, which if you never used before, is a great example of this, like a small scale example of an agent.
[00:16:22] Speaker A: Right.
[00:16:22] Speaker C: Like you try to research a topic and it's going to generate it. And now I want it to generate an actual article. And the hit people have on AI a lot of the time is they get tired of like AI slop.
[00:16:32] Speaker A: Right.
[00:16:32] Speaker C: You go on LinkedIn, it becomes very, very clear that this has been written by ChatGPT and so you ignore it entirely. Well, like how do you get around that? Well, you can feed these AI, you know, systems, agents, automations, whatever you want to call them. Like examples of your own writing. Right. Or if you don't like your own writing, writing you like. And you can essentially train these models. So example LinkedIn post, hey, here's how I post on LinkedIn or here's how someone who I really like posts on LinkedIn. I want you to find an article, I want you to find something I can write about and then I want you to create this piece of content for me in my voice. I give it the thumbs up and I want you to automatically post it everywhere. I want you to post it on LinkedIn. I want you to Post it on my personal blog, the website, and then I also want you to reply to comments to it and take it a step further. The comments I get from that or potential leads, I want you to respond to those and go through that whole scenario.
So I think yeah, anything that requires that took like just like the low hanging fruit.
[00:17:27] Speaker A: Right.
[00:17:27] Speaker C: What were you spending a bunch of time doing? You don't like doing AI could probably do it. So you can actually do other stuff.
[00:17:34] Speaker B: Yeah. And it can even enhance and help you like the performance of your website. You know, you can place an agent chatbot to communicate what website visitors who, you know, might not have stayed on there if they didn't start communicating with somebody.
[00:17:49] Speaker C: Yeah, And a lot of this stuff is pretty easy to do. The chatbot one's a great example. There's tons of like third party tools. Well, it'll automatically look through your website, essentially create its own FAQ page and then you just put it on there. And I think it costs a couple hundred dollars a month for these third party tools, which, you know, before a few years ago if you wanted somebody to do that, that probably you'd be out thousands of dollars. So it's also becoming a lot cheaper as time goes on, which is great.
[00:18:15] Speaker B: Yeah. I found a company this week. They have like, I think like 900 pre built agents and it's like 35amonth or something like that. You know, I don't know how good they are, I didn't try them but I mean it's, you know.
[00:18:30] Speaker C: Yeah, yeah, it's accessible which is the, the biggest thing.
[00:18:33] Speaker B: Yeah. What industries are seeing the most disruption right now from these agents?
[00:18:40] Speaker C: I think the biggest are kind of marketing with what we talked about, especially when it comes to content creation. I think we touched on it briefly with like how the ChatGPT image generation got so much better recently. Right before like images, there were some image models out there that were really good but they had a problem with stuff like text. Well now you can do really good images with text and like you start thinking about, okay, like Amazon storefront stuff, like all those ads, like those stuff you could generate with AI and make pretty good stuff because you can give it examples of your own products, right. Like here's product A that I love, here's product B in some sort of ad that I really like, can I swap them out and do a couple of changes? Before you have to know how to use Photoshop, you have to hire a designer. Now you can actually get an 80, 90, 95% solution with AI. Which for a lot of people is good enough at the price point. And then I think the second one would be just like the software industry as a whole.
[00:19:37] Speaker B: Yeah, yeah. And even a lot of different, like Canva, you know, they have a plugin for chat GBT that, you know, you can even do it right through there.
You know, I use Canva pretty much every day.
[00:19:50] Speaker C: Yeah, Canva's a big one. Figma is another big one that's starting to get integrated. So yeah, all these like design type products are all getting the AI treatment.
[00:19:59] Speaker B: Yeah.
How can AI agents help businesses scale without massive hiring sprees?
[00:20:06] Speaker C: I think it's not too difficult because I think a lot of this stuff can be done without necessarily having an AI expert. You don't have to be an AI expert to do these things. You don't even have to be very technical to do these things. Oftentimes you're probably one or two YouTube tutorials away from getting enough of a grasp on some of these like lower level AI automation workflow concepts to actually start integrating them.
And so I would think that doesn't require you to hire, it just requires you to kind of have an idea of what you want to apply it to. Because if anything there's almost too much information out there, right? You go on YouTube, you go on Instagram and you type in AI agents, you're just going to get inundated with 10 million things and you're honestly probably going to get overwhelmed and just quit trying to do it. So I think the biggest thing is identifying first, like what if you did have an agent that could do something, anything, like what would it be like, be very specific about it.
And honestly what I tell people is like, also you can use AI to help figure this out, right? Go to ChatGPT, kind of talk about like, hey, here's how I run my business, here's my day to day. I'm trying to figure out where could I implement AI at a small scale, very simply, and I think you'd be surprised at some of the insights it would have. And I think once you identify where you actually want to use it and what success actually looks like, then it becomes much easier to kind of go down the YouTube rabbit hole, right, and actually look for, you know, explicit examples of how you could do it. And again, I think people should be integrating AI along that whole journey, right? Ask AI how can I solve this? And you know, it's, it doesn't just have to be the solution. You kind of use it as a partner to bounce ideas off of. Because it's probably going to give you some good info.
[00:21:43] Speaker B: Yeah, I've even, I've even used AI to, you know, I'll be using one tool, like a video creation tool, and then I'll ask AI to give me the prompts that I should be asking.
It's like answering its own questions. In a sense. It's pretty wild.
[00:21:59] Speaker C: No, it's, it's a great idea being like, hey, I'm trying to come up with an idea for, you know, X or whatever. Like, what would be a good prompt for that? Here's my first ideas. And it's like, oh, yeah, sure, here's a best practice prompt. I mean, like AI is almost like a really, really, really brilliant, yet really, really dumb intern.
[00:22:15] Speaker A: Right.
[00:22:15] Speaker C: Like it's going to go off at a million miles per hour. But you have to give it blinders.
[00:22:20] Speaker A: Right.
[00:22:20] Speaker C: You got to set it on the right path or it's just going to kind of keep hitting the wall over and over.
But that's kind of how you have to treat it.
[00:22:28] Speaker B: What are some of the challenges that businesses, you know, could potentially face when they're trying to integrate these agents into their workflows?
[00:22:38] Speaker C: I think kind of similar to my last answer is not really knowing what they want to do. I think I work with a lot of clients who tend to be smaller businesses and it oftentimes takes a lot longer than you would think to actually nail down. Like, what do you actually want to do here? Like, what problem are you really trying to solve? Because you have to be explicit. Like, these are very, very powerful tools. But if you don't have a clear, like, vision for what you want the end game to be and what task you're really trying to eliminate or automate, it's easy to just feel like you're throwing AI at the wall because you don't want to get behind the AI curve.
[00:23:12] Speaker A: Right.
[00:23:13] Speaker C: But that's the wrong way to do it because that is where you're going to burn money.
[00:23:15] Speaker A: Right.
[00:23:16] Speaker C: If you bring in somebody who's your quote unquote AI consultant or you're spending money on all these tools without really knowing what you're trying to do, I think that's the easiest pitfall to get stuck in. And you'll kind of just be in a loop where you're just kind of trying things over and over and over instead of being very clear about what's the problem and what's success.
[00:23:34] Speaker B: Yeah. And one of the other things that I, I found in doing this is you really gotta Think out the steps that it takes to complete that task because you have to, you have to break it down, you know, for the agent so it does it properly.
[00:23:52] Speaker C: Yeah. And again, I think AI is great at helping with that. Whenever I go into like a flow of where I've been, you know, hired to create some sort of automation or some sort of agent to help someone out, I mean, my first thing I do is I go to AI and say, hey, like, here's what we're trying to solve. Here's my initial ideas of what I want to do. Here's the eventual end game. Like, what are your thoughts on this?
[00:24:14] Speaker A: Right.
[00:24:15] Speaker C: Like it has access to essentially the entire Internet.
[00:24:18] Speaker A: Right.
[00:24:18] Speaker C: Like it's seen someone has done something similar before and you don't have to go with what they want, but it's a good person to bounce stuff off of because you can ask it infinite questions. It's never going to get annoyed and tell you to stop asking the questions.
[00:24:29] Speaker A: Right.
[00:24:29] Speaker C: So I think people should always start there.
[00:24:32] Speaker B: Yeah. We're out of time for this segment, but when I come back, I want to, you know, I want to talk a little bit about how, how the agent space is kind of revolutionizing marketing.
So when we come back, we'll be back with, you know, with Chase. Uh, if it were for these sponsors, welcome back to Street Level Marketing. I'm your host, Mark Lamplugh and I'm here with Chase Hanigan. And we're talking about AI agents and, you know, how they're revolutionizing businesses. And you know, I really wanted to get into talking about how they're revolutionizing marketing. You know, this is street level Marketing, in my opinion. I think the agents really have the greatest impact. And you know, from what I'm seeing and what I'm doing, I mean, I'm sure there's areas that I'm not working in where these applications are applicable, but in for like marketing, you know, it's, it's pretty wild.
Chase, how are agents changing the marketing landscape for brands today?
[00:25:39] Speaker C: I think they're letting them create way more content at scale is kind of the biggest thing you're seeing. And I think what you're also seeing is a huge variance in that content.
[00:25:49] Speaker A: Right.
[00:25:50] Speaker C: I think it's very, very clear when you see poor AI marketing.
[00:25:54] Speaker A: Right.
[00:25:55] Speaker C: And it's not clear at all when you see really good AI marketing. Which is why I think people tend to hate on AI marketing and AI driven content.
[00:26:02] Speaker A: Right.
[00:26:03] Speaker C: Because all you see is the bad stuff.
[00:26:04] Speaker B: Yeah.
[00:26:04] Speaker C: Really, really good AI content either is so good you don't notice it or it's been like they did a really good job of doing it in tandem with like human intervention. So Maybe that's like $0.50, 60% AI and then the rest of it's kind of like human generated.
[00:26:19] Speaker A: Right.
[00:26:19] Speaker C: And I think you definitely have to like strike a good balance and figure out where AI makes sense and where it doesn't. And that's also something that's always changing. Like we talked a little bit about Google and how there you have like that brand new video generation model that's amazing and has voice and so it's like stuff that just you couldn't do before. Like actual AI generated video that looked good and looked real. Now just pull as of yesterday kind of is possible.
And so that's going to change it a bunch. But yeah, I think just the sheer volume of content you can create is massive.
[00:26:54] Speaker B: Yeah. I think I saw the other day Volkswagen create a whole commercial, you know, with AI and it was.
You wouldn't even know if you, if you didn't, they didn't tell you.
[00:27:07] Speaker C: Yeah, I know Coca Cola did one not too long ago as well. Full AI generated.
You're starting to see a lot. Especially when it just comes to like the still image marketing. I know like Activision did one for the recent Warzone like re release.
So it's everywhere and these are huge multi billion dollar companies doing it.
[00:27:27] Speaker B: Yeah. I think that basically it's giving everybody an opportunity to make themselves look bigger than they are. You know if they really kind of grasp the marketing angle and the aspect of it, you know they can create content for their, for their base or their friends or followers.
You know that just as good as like a larger company.
Adam. You know, pretty much practically nothing.
[00:27:58] Speaker C: Yeah. 100. You can, you can, you can become a one man marketing army and you can do things that you really couldn't do before if you didn't have very specific skills. Like when it comes to design and actually creating these videos and editing like you had to be like an editor or a video editor or like someone who trained on this for years and years. And like those people are still the best right now.
[00:28:19] Speaker A: Right.
[00:28:19] Speaker C: Like we haven't reached the point where you can kind of prompt your way into that but you can get pretty close and like without having necessarily that background because of AI. So it's definitely like democratize the space for people who before just had good ideas and not necessarily technical know how. Well now you don't necessarily need the technical know how as much.
[00:28:38] Speaker B: Yeah.
How do these agents enable like really hyper personalized customer experiences?
[00:28:45] Speaker C: I think a good example that is what Netflix is possibly doing soon, which is like using AI to generate ads that are based on the videos you're currently watching.
[00:28:57] Speaker A: Right.
[00:28:58] Speaker C: We'll see how good that is and when they roll it out.
[00:29:00] Speaker A: Right.
[00:29:00] Speaker C: But it has the data.
[00:29:02] Speaker A: Right.
[00:29:02] Speaker C: It knows what you're watching. It pretty much knows who you are because they have all this data on you. They know what ads they were already going to run. And by combining those two and using AI to just generate the video with certain assets they already have, that's like as hyper customized of an advertisement as you can get. And before it used to be like, hey, here's ads 1 through 100. We could send you and based on your profile we're going to pick ad 5 and 7.
[00:29:25] Speaker A: Right.
[00:29:26] Speaker C: Based on your shopping history, what you search for, et cetera. But now those specific ads, it's not just that the company is personalized.
[00:29:32] Speaker A: Right.
[00:29:32] Speaker C: The ad itself of that personalized company is now also personalized.
[00:29:37] Speaker A: Right.
[00:29:37] Speaker C: And now is really going to be that good. Are we there yet? I don't know, but I think we will be soon.
[00:29:44] Speaker B: Yeah. And I don't know the technical aspect of this, so you may know this. Do you need more type of. If everybody's seen their own very hyper focus advertising, so essentially you're seeing a different video that's, you know, for you and your neighbor is seeing one that's, you know, different for them. Do you need more, more like bandwidth in, in these lines to move this data back and forth? Like how does that work?
[00:30:14] Speaker C: I, I don't know for sure. I imagine it's not easier. I imagine. Yeah, there probably is some sort of like technical debt accumulated from, from, from stuff like that.
[00:30:23] Speaker A: Yeah.
[00:30:24] Speaker C: I don't think it'd be cheap either to do all this customization, but I think it's also one of those things that gets cheaper and cheaper and cheaper.
So what kind of might be a gimmick now perhaps soon won't be. And that's like with a lot of this stuff, like it's so new. And the horizon for whether this is going to be good or legitimate or not is constantly changing. I mean, again, I keep bringing up the Google thing, but the videos they showed yesterday, if you'd show those a year ago, you'd be like, no, that's four or five years out. Like it is impossible.
And it is.
[00:30:56] Speaker B: What role do these agents play in optimizing like ad spend and targeting?
[00:31:02] Speaker C: Yeah, I think what they can do is they can do a really good job of evaluating how the ads are doing if you just feed it the data. I think when it comes to like ad spend. Right, like which campaigns are doing well. Okay. Like well based on the data I was given. Well, this is the campaign that's doing the best and perhaps the agent can take it a step further.
[00:31:22] Speaker A: Right.
[00:31:22] Speaker C: Like it now knows which campaign is doing the best. Can it now generate content based on that campaign and make more ads that are similar to the ones that are doing well?
[00:31:31] Speaker A: Right.
[00:31:31] Speaker C: So that would kind of be the next step it takes the data, analyzes the data and then actually starts generating real content based on that.
[00:31:41] Speaker B: Yeah.
When you're. So now we're getting into these agents and you know, creating all this content, what's a good way when we're automating this stuff that brands can, you know, potentially not lose their identity?
[00:32:01] Speaker C: Yeah, I think that's a great question. I think the biggest thing you can do is train it on your own stuff so it has its authentic voice.
[00:32:09] Speaker A: Right.
[00:32:10] Speaker C: If you, for example, making posts on LinkedIn.
[00:32:13] Speaker A: Right.
[00:32:13] Speaker C: If you don't feed it or train it on your own writing and your own style and how it would be written normally by the human beings in your company, then it's going to sound like chat. GPT is just writing for you and it is. Right. So that's inauthentic. So I think on the front end you have to do a really good job of training it on your own material so it maintains your authentic voice. And then on the other side you have to figure out, okay, like where do we need to stick a human in the loop? Do we need to make sure a human being actually has eyes on, on the final product every single time before we post?
[00:32:42] Speaker A: Right.
[00:32:43] Speaker C: I would say normally, yes, you would want to do that. Especially now it's kind of dangerous to just like let it go crazy, especially on public facing stuff like that. But I think the biggest way to do it is in sort of the training phase and giving it its rules and its logic and really figuring out what is our brain and what is our voice. Because you can't train it on your voice if you don't even know what it is in the first place.
[00:33:04] Speaker B: Yeah, that's so true. Yeah. Another thing that I try to do is try not for me, try to create different types of stuff to put out on social media. Don't be posting all of the same type of videos all the time or you know, just articles. You gotta, you gotta be very selective. You Gotta have a different variety of different things.
Some that engage like answers or feedback from potential customers.
Because you start putting the same stuff out there all the time, people are going to stop being interested.
[00:33:44] Speaker C: Yeah, 100%, I think. Again, another great place for AI where you can show it what you have done and just use it for idea generation or coming up with things that you haven't done before, having it realized trends like hey, you've posted this type of content seven of the last ten times. Right. You're not having AI even generate content at that point. You're just using it as essentially sounding board to make sure you're like mixing it up as much as you need to.
[00:34:09] Speaker B: Yeah.
What are some of the ethical considerations marketers should think about or keep in mind when using these agents?
[00:34:18] Speaker C: I think it ties into the authenticity and I think it's a fine line.
[00:34:23] Speaker A: Right.
[00:34:23] Speaker C: Do you want to put this was created by AI if AI touched any part of your marketing? Probably not.
[00:34:29] Speaker A: Right.
[00:34:29] Speaker C: I don't think that would necessarily go over well.
I think it's a little more obvious in other spheres like with customer service, like should they tell you it's an AI agent and have the ability to talk with a human right away?
Maybe. I think people would be less likely to do it if you tell them then they're gonna have to wait 90 minutes to talk with a human.
[00:34:47] Speaker A: Right.
[00:34:49] Speaker C: I think marketing is just tricky, Right. Because I think if you slap on this was AI generated on there, it's probably not going to have the same effect.
So I, I don't know. I think that's a great question and it's hard to really answer as to like do you owe the customer, like an explanation of things are done by AI.
[00:35:06] Speaker B: Yeah.
Another area too is I spent a lot of time in healthcare before the automotive space and like I can see like these phone agents being used for service related calls or sales calls or qualifying. But when you're working in like healthcare, I think there's a fine line to where people are talking about personal health issues to whether they should be talking to, you know, an AI overlaid person. But then I just saw an article, I think it was yesterday where a, a lung doctor was basically saying this. These AIs can predict pneumonia, you know, immediately where it takes, you know, it's, you know, he was basically admitting that it's smarter than they are when predicting some of these medical issues.
[00:36:01] Speaker C: Yeah, yeah. It's a tough, it's a tough thing to balance. I think definitely in like the doctor scenario.
[00:36:09] Speaker A: Right.
[00:36:09] Speaker C: It's being used as a tool. So you do know, at least he's the one signing off and making sure it is correct. And then I, I think you could argue even in the healthcare space like the AI is actually going to make that even better. Right, because you could have AI handle all these like non personal stuff, you know, hopefully stuff like scheduling or just like random questions.
[00:36:26] Speaker A: Right.
[00:36:27] Speaker C: And stuff that does require, you know, the. No kidding, like human touch, like for sensitive subjects. Hopefully. Now this is now like less than the load on everyone else and you can kind of funnel them in that space so they're actually working on the important things that you know for a fact someone needs to talk to a human. And I can offload all these non critical conversations to AI.
[00:36:46] Speaker B: Yeah, we're out of time. But when we come back I want to talk about, you know, the future and you know, what you see down the horizon. So when we come back we'll, you know, with the word from our sponsors, we'll be back here with Chase.
Welcome back to Street Level Marketing. I'm your host Mark Lamplugh. You know, we're here with Chase Hannigan and we're talking about AI agents.
In this last, you know, part of the show, I wanted to talk about the future, you know, and what you see going on and, and maybe coming up, what are some of your boldest predictions for AI agents in the next decade?
[00:37:32] Speaker C: I think you're going to see like AI dedicated teams in most of these big organizations, right? Like some sort of team that's kind of in charge of like how these AI agents are deployed in these organizations. And I think you're already starting to see it like the term AI engineer and not talking about someone who's like machine learning educated. This is like people who just use AI tools. Well, is something you'll see on like job boards. If you go to LinkedIn jobs right now and you search AI Engineer, you'll find them and they actually pay pretty decently. And they even use terms like vibe coding. And if you don't know what vibe coding is, that's like using a tool like cursor, some sort of AI tool to code things for you. And vibe coding means like, hey, I don't know how to code, but I'm going to use a tool that does to code things. And it was almost used as like an insult on before like a few months ago and now this was like I saw it for like Citibank or it might have been banking mayor, but like a legitimate organization said we're hiring an AI engineer that knows how to buy code and use these things.
[00:38:29] Speaker A: Right.
[00:38:29] Speaker C: And that's. Right, that term showed up a few months ago.
So I don't think it's going to be wild to see teams that are like, you're the AI team and you figure out how we deploy these AI agents cross functionally so they can talk to one another in different teams. So I think it's going to be more and more seen as like a legitimate spot in, like, the organizational hierarchy, and these AI agents seen as like, members of the team that you're expected to kind of collaborate with or at least push data to.
[00:38:55] Speaker B: Yeah, yeah. The coding is a big one. I canva just roll rolled out a, you know, a coding piece, you know, for their platform. And I just tested it out the other day and I created a.
A timeline of the Corvette, you know, that was interactive and it coded it for me. I mean, it was pretty neat. I mean, you know, so I can't imagine where it's going to be in a year or two years from now if I could do that.
[00:39:27] Speaker C: Yeah, I mean, it's wild. The whole idea that you can even just like prompt something on your web browser and have it create code for you again was basically a pipe dream five years ago. And now it's like, oh, yeah, now it's fine. I can have it do all this stuff and kind of take it for granted already.
[00:39:43] Speaker B: Well, you know, what I'm thinking here now is like, what are some of the easiest way that marketers could start experimenting with these AI agents?
[00:39:53] Speaker C: I think the easiest thing to do is go to ChatGPT, open up chat GPT projects and just dump a bunch of your marketing material in there. Show it marketing material you've done in the past, show it to your own writing and start using that to generate drafts of content for you. Give it topics that you wanted to research about it and tell it, hey, I want you to find stuff out about this topic. Then I want you to reference all my own marketing material, and I want you to come up with stuff that sounds like me that I could post either on my website or my blog or LinkedIn, because I think that's a great way to see how you can start training these training, right, these AI agents and these AI tools on your stuff, on your work, but also leverage everything AI can do to create marketing content. And you don't have to worry about getting posted anywhere automatically.
[00:40:38] Speaker A: Right.
[00:40:39] Speaker C: It's in your chat GPT safe space. But you can kind of see what you could do in theory and right. If you like what that's doing, then you can start thinking of like, how could I automate this?
[00:40:49] Speaker A: Right.
[00:40:50] Speaker C: How could I get this outside of ChatGPT and somewhere else where I can make it a bit more robust and then even like spread it across multiple socials?
[00:40:57] Speaker B: Yeah. What are some of the tools or platforms you recommend for beginners or you know, small business owners that might want to be, you know, you said ChatGPT, but is there any other ones?
[00:41:08] Speaker C: I think Canva is a great one to start with too. Just like you're talking about, very easy and simple to kind of like get started in practicing.
Those two are the biggest.
[00:41:19] Speaker A: Right.
[00:41:19] Speaker C: Especially if you're extremely non technical. If you're more technically inclined and you really kind of want to dive into like the AI agent building space, I would suggest N8N that's a little more technical. You don't have to be a coder to use it. But if you're like, hey, like, I know ChatGPT, I'm confident that I can learn like some sort of more robust tool. That's where I would point people to because that's when you can really start building these things for yourself. Which again, if you're not technical, it sounds scary, but I'm totally of the belief that you're like a handful of YouTube tutorials away from being able to do it if you're just curious about it and want to see how you can do it yourself.
[00:41:56] Speaker B: So how about yourself?
What are some of the things that you're doing for potential businesses or in the agent space?
[00:42:05] Speaker C: Honestly, it's all over the place. The only thing that's common to all of them is they tend to just be smaller businesses.
[00:42:12] Speaker A: Right.
[00:42:12] Speaker C: So this isn't something you have to be some extremely large organization to do. But I've done everything from simple, like, hey, I want.
They were like a mental health organization. They did a lot of town halls. And what they wanted was the transcripts from the town halls to be like synthesized with chatgpt and then automatically chopped up into like 10 different Twitter posts and then to be posted on Twitter.
[00:42:36] Speaker A: Right.
[00:42:37] Speaker C: So they could kind of market their thing and get a conversation going on the other end. I did some work with like a smaller venture capital firm that ran an accelerator.
[00:42:44] Speaker A: Right.
[00:42:44] Speaker C: So essentially like once a quarter they would have 300 people who are trying to get funding for their businesses and they wanted some sort of program, some sort of workflow that would take all these guys information, like their pitch decks, you know, all their like history, their pitch for themselves and then like evaluate it.
[00:43:00] Speaker A: Right.
[00:43:01] Speaker C: Should they invest in them? Why? Why not do some sort of like validation on what they're actually saying. Go to their LinkedIn profiles, like scrape their data.
[00:43:08] Speaker A: Right.
[00:43:08] Speaker C: So actually pretty dynamic and robust. So everything from simple just I want stuff posted on Twitter to like, I want you to see if I should invest millions of dollars into these guys.
[00:43:17] Speaker A: Right.
[00:43:18] Speaker C: It's all over the place and. But for the most part they tend to be rather simple.
[00:43:22] Speaker A: Right.
[00:43:22] Speaker C: They tend to just be like, I want some sort of AI agent that I can feed my personal data to you and have it act like a member of my team. Whether that's respond to emails or give me new ideas about how I should approach these problems. Something along those lines.
[00:43:35] Speaker B: Yeah.
How can AI agents empower, like entrepreneurs and underserved communities?
[00:43:43] Speaker C: I think it just allows you to do way more as a small team. And so when it comes to underserved communities, right, like you probably, you may have like less funding, just less people on board. And so you need to do more with less.
[00:43:55] Speaker A: Right?
[00:43:55] Speaker C: And I lets you do more with less because it lets you do things that otherwise would require a certain amount of technical expertise or more often just time. Like I keep bringing up like the research stuff to like post articles because research takes a ton of time to do the quality research on whatever you're trying to write and stay authentic and make sure this actually makes sense. That's a huge amount of time you need to dedicate to do that. And AI can do that. And same thing with the marketing stuff, like Canva, right? Like creating these like infographics or interactive graphics like you talked about with Corvette. Again, try to do that on your own a few years ago, ton of time probably wouldn't look good. And now you can do it very, very easily. Right? So it's just, it's a huge, just like force multiplier. If you're a small team and you don't have the resources to just like have someone else do it.
[00:44:42] Speaker B: How are you addressing some of the skepticism, you know, small businesses or people may have about, you know, these AI agents in the marketing world?
[00:44:52] Speaker C: I think you just have to show them real world examples of it working.
Like I said, I think it tends to get a bad rap, especially in marketing, because you only see the bad stuff, right?
You rarely see a lot of the good stuff. And a lot of the good stuff that AI does is kind of behind the scenes. No one can see the research it did or the ideas it created or the past that led you down, that's all invisible to the end user.
So I think the biggest thing is just showing real world examples. And that's why I say do the chatgpt thing, hop into projects, give it your own stuff. And I think when you see AI create something that actually sounds like you and is on brand is kind of wild, right? Like this machine that I just fed, you know, five, six, seven examples of my marketing or writing and all of a sudden it sounds like a little mini me is pretty crazy, right? Like, how long would it take you to train someone to do that themselves? And how much would you be paying them?
[00:45:39] Speaker B: Yeah, yeah, I use a application, I think it's called R Dial or something like that, where I uploaded a three minute video of me just talking and it was able to clone me. I mean, it's fun. It's.
[00:45:54] Speaker C: Yeah, it's, it's crazy. Yeah, the voice, the. Yeah, the voice generation, the video generation. Clones are also. There's a ton of them out there. And again, they're only getting better, which is kind of a little scary, honestly.
[00:46:04] Speaker B: Yeah.
What's one actionable step the viewers can take this week to start exploring ages?
[00:46:13] Speaker C: I would say sit down and ask yourself, if I could have one thing automated by AI, what would it be? Seems like an easy question. And I think you'll find it's harder than it is to really nail that down and then also nail down what would success in this look like? Because if you can do that, if you can accurately define the problem and accurately say, this is what I want it to look like, you can quite literally take that to AI and say, how can we do this? And I guarantee it will generate you a step by step explanation of how you can, which is super enlightening, especially if you keep it small. Because if it's super small and tight in scope, you can probably do it in that same weekend.
[00:46:47] Speaker B: Yeah, well, you know, we're, we're coming out of time, but I want to give the viewers an opportunity and this kind of goes into what we kind of just lastly talked about because I know you have a way that people can get involved or you, you help them, you know, so can you tell the viewers like, you know, what you're doing in the AI space, how they can get a hold of you?
What are some of the things you offer people?
[00:47:16] Speaker C: Yeah, I'd say the easiest way to get a hold of me is just on social media, so Instagram, YouTube, TikTok. If you just look for Chase AI, you'll find me there.
So my whole niche is just trying to explain these AI topics and AI agents in a way that makes sense for people who aren't technical. Like I'm not someone who came from a technical background. My background is actually the military. I was a pilot before this kind of just fell into the AI space. Mostly self taught, so kind of just made it my goal to teach people who aren't technical what AI agents are and how to use them practically, especially in a business sense. So if you find me on social media, you'll see kind of videos in that vein. I also run an AI agency on the side for where I do client work. And then lastly I have a school skool.
There's links to that in my socials where it's essentially a community for people who want to deep dive even further. Where I do like step by step guides of like, okay, we keep talking about AI agents. How can I actually build one and make it work for real? So that's what that's all about. But yeah, just check me out on social media again is the best place.
[00:48:18] Speaker B: Great. Well, thanks a lot for coming on the show and I think we learned a lot here today and I think the viewers, you know, can pro, you know, probably picked up a lot of good information here to, to start rolling these out. And I, one of the things I was trying to talk to other people about is it's either now or later. I mean this is coming so it's probably better to get on board now. So you're ahead of the curve because you don't want to be the one trailing behind when all your neighbors, businesses or your competitors have, you know, a team of 500 employees to a little small mom and pop store that you don't see because, you know, it might put you out of business.
So we had a great show. I want to thank our viewers for watching. I want to thank Chase for coming on the show. And we'll see you next week with another episode of Street Level Marketing.
