A.I. automation 101 for Shopify drop shipping in 2025

Hey, welcome back. In this video I want to introduce you to AI Automation 101. So I'll give you some examples, some ideas. Ultimately this is one of those things that you need to like experience it for yourself, go play with it, try different zaps, different creations, test it in real life. And if you spend like 30, 60 minutes on it and you feel like the tech is beyond you, don't forget you can always go to Fiverr upwork, Freelance or some of the resources we've covered. And it just outsources this work for like 20, 50, maybe 100 bucks. Okay, so I'm going to cover Zapier and make now first of all, you should have already your API credits in your account and you're going to need this OpenAI account set up for your API keys that you get up here. And then under organization, you'll get your organization id. So you're going to need those to connect to Zapier. So when you go in here, they have this new like AI screen, I kind of like the classic view. And you can just put in what different apps you want to create. Okay, so you could create automatic content creation, you know, chat, GPT, writes content and then distributes it to Facebook, YouTube, Instagram, you know, so you can create really complex zaps that are multi step, 10, 15 steps. Now you can create simpler zaps with maybe two, three steps. And I wanted to give you this example where I essentially created an auto Reply bot using OpenAI API. So it's going to be using these credits that we got for free. And so each time we get our help ticket, essentially in our help desk, customer sends in a question or a prospect sends in a sales question, or maybe somebody needs tech assistance. Then a conversation is created and we reply to them. With OpenAI as the first step, it's referring back to a database of our FAQ and stuff like that. And essentially if the answer is unsufficient, they can escalate to a human. But so think about it, you know, in a company where there's 10 support reps, it's all humans answering the questions. And oftentimes these questions are the same. Many of these reps are using keyboard shortcuts to like fasten their time. But you know, they're expensive. So with an AI like this, you could potentially reduce the support time required by 80%, maybe they can keep two reps so there's still that human attention. And you know what, if the AI makes a mistake or the answer is incomplete, you still want that human action. But a lot of these questions can be actually answered by AI. Automated Save the business thousands of dollars each month and you just charge a fraction of it, right? So everything else is profit to them. That's why it's very easy to sell these done for you services, right? You don't need a huge name or reputation to go to business and say, hey, give me a thousand bucks a month and I'll save you 8,000 in support cost. You know, just to give you an example. Now this works with anything, right? You could change this app to, to Intercom, right? Which is a very popular like live chat support system or Zendesk is a very popular one, right? And I could go on and on. So first thing is you're gonna select the event, right? So conversation created or assigned, there's different ways to set it up. And then you're going to connect your account. So you're going to have to log in, it's going to connect, then you choose the trigger, right? So they sent an email to this email, this user. You can put in different tags and keywords to say you could even have multiple bots. And you could train them on either tech sales, regular general customer support. And based on different keywords, you know, that's in the text, you could assign them to different people or you could assign, if this actually goes to a human, like say there's a refund request. Well, maybe you want a human to look at that rather than AI. So you can create that separation. For now we're going to leave it blank, status active, and then we continue. Then we did a test. Now we skipped it because it said there's no data, but there's definitely some data, so I'll double check that. Nevertheless, we connected the OpenAI API send prompt and this is just an example over here. So of course we connected our account. Like I said, you're going to need your API key, which you find here, and, and then your organization id, which you find over there. So once that's done, you'll connect your account, you choose an action model. You can use the DaVinci I also like either the GPT 3.5 Turbo or the GPT 4, the preview version. So, and you can give it a prompt like, hey, you're a helpful assistant. You can pull the data, the FAQ from this website, try to formulate a concise helpful answer to the following email by the customer or by the prospect, right? And so here you can take this data from essentially the help scout message, right? That's the whole point of Zapier is it connects different apps, it lets them talk through API. It pulls the data from Help Scout. It has the text like, hey, this is what the customer just said. This was the body. And then the subject of the email. So this is what the customer said. Try to give a reply, sign off as my AI assistant, and you can add a PS at the bottom, like, hey, in case this is urgent, important, whatever you need, Human assistant, reach out to Mark, for instance, right? And there's many ways to do it, but this should just give you an idea. Then you can go ahead and run a test and you can actually go ahead and test it for yourself. You know, send an email to your support ticket system and then see if you get a satisfactory answer. And then you want to tweak it until that happens. So hopefully that gives you an example, like how relatively easy it is to create something that could save hours and thousands of dollars for a business. Even more if it's a slightly bigger company. And you could easily charge 500 or a thousand bucks or even a couple thousand bucks a month to deliver such a service. And then you just maintain it, make sure the updates are up to date. If there's any technical outage or any issue, then you double check and fix that. So the toughest part is just setting it up right? Especially in the beginning. If you're not very tech savvy or maybe not familiar with Zapier make, it might be a little bit frustrating, but, you know, once you get through that, once you set it up, it's very easy. And then inside your account, you'll be able to just literally duplicate these. Right, with one click. So next time you set it up, it'll be much easier. And then maintaining it is as well quite easy. Now let me show you. Make. So make is kind of like a Zapier competitor, slightly cheaper, very similar, and they do the same thing. They just have a little bit different design. Now, I like them because they have really cool templates, and I'll just give you a couple just to give you an example. So, for instance, this is pretty cool. So the AI can write an article on your blog post, essentially, and or it can take an article. It could do both. Right. You might just need to move some of this stuff around, but it can totally do both. So either take an article, compress it, rewrite it, shorten it, because Twitter has like whatever 140 characters limit, so make sure it's compressed, put it on Twitter, maybe add a link as well. I'll put it on instagram but make sure you add a picture. Like maybe you take the thumbnail picture of the blog post, etc, etc, maybe turn it into a story and put it on an Instagram reel. Like you can even do that. There's different apps that turn these pictures in a video. In a couple of seconds, boom goes up, you can add automatic captions. So again, this video isn't one on one introduction just to repeat myself. So I'm not gonna go and show you how to put all of this together because we'd be here for a couple of hours if I go through all these different templates. But again, just go ahead and play with it. And it works in a very similar way. I mean, I'm actually going to pull up a couple just to show you the basics, but wouldn't go through all of them if that makes sense. Like I mentioned another option, it can write the whole article and then repurpose the content. Now here was another one is a telegram bot, right? So you can either use any telegram bot message to be translated or kind of like how I did it with Help Scout. If there's any message on Intercom, Help Scout or Telegram, you can take that message, analyze it, maybe compare it to like a database of frequently asked questions, and then give an answer to that. You can these days even create GPT bots, which I'll cover in another video. Like for instance here, let me go to assistance. So you can train entire GPTs, give it tons of files, it can access the web, it can access different tools, and then essentially it can give you personalized feedback. So for instance, I created like this business mentor and I can give it my different offers and stuff that I have and it'll give me feedback based on tens of thousands of hours of this business mentor's mind compressed into AI, which is pretty crazy. So I'll cover that in another video. It is somewhat similar to this, but slightly different. But you want to be aware of all of these tools available at your disposal. So I hope that this gives you a basic introduction to AI automation that's possible through tools like Zapier and Make.com and then as well, in another video, I'll cover the chat GPT bots that you can create and then directly in the back end of OpenAI's playground. So there are two different things, but both of them let you create different like GPPs that you can talk to. Okay, and let me see if this is anything relevant because then I'll just show it to you. So you see, they have a guided setup so they'll be able to guide you through it as long as you read these instructions. Like, I just wanted to show you this last thing because it is essentially identical to Zapier. The only big difference you're gonna see is that they have this kind of like water flow set up. So, you know, step one's up here, step two, step three, four, five. It's just a waterfall here. Essentially it's just horizontal, right? So it goes to the right. But the idea is the same, right? You connect the application, you authorize it, you select which action you want it to pull, what data you want it to pull, and then how you want the other app to respond to it. And in a similar way, you can create automations as long as you want two steps or 20 steps. Okay, so I think that gives you the 101. Let me know if there's any questions below. Happy to answer them. Or create another video. And then in the next video I'll see you explaining how to use the OpenAI playground and how to create these GPTs.