THE Ultimate Guide to ChatGPT for Business | Best Practices, LLMs, AI Agents, and More

Today I'm going to walk through a complete beginner's guide to AI in business. This video is not going to just stop at ChatGPT because AI is evolving, and to compete in the business world, gotta know how to use AI for more than just prompting with ChatGPT. By the end of this video, you'll be better at using AI than 95% of people when it comes to business. We're going to cover how to get the most out of ChatGPT and other LLMs, how to bake AI into your daily processes so that it provides you with real-time feedback loops, what AI agents are and how to easily build agents that work for you and save you time and money. Tools outside of ChatGPT, because everyone says ChatGPT, that will actually put you ahead. And as we go through this, we're going to cover a lot of different prompts and tools. You can grab the PDF with all of them in the comments. It's important to know the basics of prompting before expanding into other AI models. So let's start there and talk about how to get the most out of ChatGPT and other LLMs. We're going to cover how to train ChatGPT to quote unquote know you, which LLMs to use when, and how to get the best outputs possible. One of the biggest mistakes people make with ChatGPT is not giving it enough information about themselves. What do I mean by that? Well, ChatGPT recently has the ability to remember conversations and it knows a lot more about you versus other LLMs out there which are still catching up. So if you want to give ChatGPT more information about yourself, it's worth talking to it more. It's worth asking it how much it actually knows about you. So one prompt that I've used is simply asking it, hey, on a scale of 1 to 10, how well do you think you know me right now? And it actually answered me a 7 out of 10. So I was like, please ask me clarifying questions so you feel like you can know me at a 10 out of 10 level. The other thing too that I've run it through, because I talked to it about a lot of business problems, I talked to it about a lot of personal problems too, I run it through a CIA prompt. And what the CIA prompt does is it will analyze me for how I am perceived as a threat vector, what my strengths are, what my weaknesses are, and it will actually tier those strengths out. And it says for me, for example, when it comes to business, it's one of the mistakes that I had in the past was Giving loyalty to people who didn't deserve my loyalty and sticking with people much longer than I should have. Trading on potential rather than acting on patterns, right? And what I mean by that is I often would hire more for potential than people that were proven. And while I still believe in that, I probably should aim to hire 80 to 90% based on people that have been there, done that, and then maybe 10 to 20% you're betting on potential, right? I had it flipped the other way around in the very beginning of my career. For it to know this and call me out and for it to know my weaknesses, I can keep working with it as if it's my business strategist and as if it's all also my business therapist as well. Oftentimes I'll also ask it about what is the highest leverage thing I can do right now on a scale of 1 to 10. And so when I'm asking this 1 to 10 questions, and when I'm often asking it to ask me clarifying questions if it feels like it doesn't understand what I'm asking for, that will go a lot further. And so this is important. I'm giving you a couple things right now. So one, I'm helping it understand you more, but two, I'm helping it understand what it is that you actually want to output at a high level, because Oftentimes people would just put in one sentence and they'll expect the perfect output. That just doesn't happen, right? So for example, here's the CIA prompt over here, and the CIA prompt allows you to understand yourself a lot better. It allows you to take actions on potential blind spots without you having to spend thousands of dollars with a therapist. It's also important to realize that yes, there are different LLMs, but there's also different models within different LLMs that are strong at different things, and it's important to know which one to use and when. And so for example, Gemini 2.5 Pro is excellent when it comes to coding. Claude, which is from Anthropic, is exceptional when it comes to coding or writing copy. When it comes to ChatGPT, the O3 model is exceptional for reasoning, and this stuff changes all the time. And if you're interested in this, we'll put up a little box over here if you want to check out the LLM tier models. And this is all changing very rapidly. When it comes to ChatGPT-4.1, that is a really quick way of getting good answers for you very quickly. You can also use the deep research model with ChatGPT and then pair it with O3, and sometimes it might take 30 minutes or so to come up with 12, 15, 50 pages of deep research that might have costed you thousands of dollars working with the management consultant. You just have to understand that there are different use cases. That's what I would say. Also, just keep in mind this stuff changes very rapidly too, where maybe in the past you might have used something like Perplexity, but in my mind now Perplexity is just not as relevant because these models have gotten so much better when it comes to search. And so the key thing here, I believe in my mind, is if you want to stay on top of the LLM models when it comes to business, you have to keep playing with these things. And I think if you're a business owner, it's worth paying for all these different ones. The last thing to note when it comes to LLM prompting is that even though ChatGPT has memory and knows a lot about you, it isn't actually going to always get you the best output. In fact, what you want to do if you're taking my advice from the last portion is you want to pay for different ones. Maybe you're paying for Gemini, maybe you're paying for Grok, maybe you're paying for Claude, maybe you're paying for all of these. I pay for all these right now, right? I think it's worth it because I'm getting pennies on the dollar when it comes to work that used to take me hours or weeks to do. And now what used to cost me thousands of dollars cost me cents to do. So I want Grok to come up with a great 7 cold email sequence to help push my product, Carrot. I want it to come up with a persuasive subject line. I want the body to be very succinct but to the point. And it did a really good job. I basically dictated out, here's what Carrot does, here's what we're looking to push, here's the offer, here's who we're reaching out to, and I want you to come up with a 7-email sequence. And it created the 7 emails. And then I did the same thing. I took that same prompt, I dictated out, I just copied it. I took it over to ChatGPT-03. ChatGPT-03 gave me a good response. Grok gave me a good response. What I did instead was I decided to have them compete. I went over to ChatGPT and I pulled the Grok response. I said, hey, ChatGPT, I thought that you were the model out there. This is what Grok gave me. And I just pasted the prompt. And then ChatGPT gave me something way better than what it initially gave me by also combining the thinking that Grok did. Now I went over to Grok. I took the ChatGPT response. I took the prompt and I said, Grok, I thought you were the best model. This is what ChatGPT gave me. It gave me a much better response. Both of these were production grade and ready. If you wanted to take it one step further, you can have Gemini join the fray as well. You can have Claude join the fray as well. My point is, when they start to compete with each other, these LLMs, you're going going to get better results when it comes to the right prompts. And I also use these prompts, these emails, these cold emails, when I'm carrying them over into the agentic workflow products that I use, such as Lindy. And so I might initially ask it to do something such as, I want you to be a daily stand-up coach and help coach people up on high-leverage priorities that they should be working on and score them as well. And so that might be my initial one, but then I might say, hey, make this for my agentic workflow. I want to be coaching people. I want this bot to coach people. I want to score them. I want a daily leaderboard. I want a monthly leaderboard, and I want to incentivize people. I want to have this tracked in the and it will create the right prompt for you. Because often what we dictate is not the best thing to go with. You want the AI to help you create the best prompt possible so you can get the best output possible. Now that you know how to use the LLMs to become a prompting powerhouse, let's get into AI agents. AI agents are autonomous workers that help you achieve the outcomes that you want. Think of it as virtual employees. Now, I think AI agents are getting a lot closer this year. When people talked about them before, not so much, but now these can work autonomously. You can prompt it and it can just keep doing the thing that you want it to do. I think we're a lot closer than we've ever been. So AI agents might sound overly technical, but you can actually build these yourself pretty easily. Let's walk through how. So here I'm going to show you the different agents that I have within Lindy. And Lindy.ai is my preferred version because what it does is you can see in here I can actually prompt it within different sections and I can have an AI agent— agent can decide what to do based on this prompt. In this case, let's say we have a lead come in and a lead comes in for Carrot, which is one of our softwares. So once a new row is added to a sheet, the AI will look for find years of experience, employee headcount, company location, Twitter profile, profile, recent news, and it'll send all this lever information. And then what we're looking for is we want the ICP to have more than 300 employees. And if it does, it's going to create a draft email that's sitting in my inbox, or more so my draft box. And then I can just automatically send this, right? And you could actually even have it send emails if you want. But in this case, we want a human in the loop. We want the humans to be able to review the emails before sending them out. Now, this takes out a lot of the manual work that needs to happen and makes things more predictable. Now, on the other side, it will also send them a text message too. So if they fit our ICP, if if it's above 300 employees, it's going to create a less than or equal to 160-character SMS, and it's going to try to get them to book a call, right? This is how our enrichment bot works. And you can see I'm using Twilio, I'm using emails, I'm using Slack. We also have a daily standup accountability bot over here, and it's very simple. So basically every day at 2 PM, Monday through Friday, it's going to look at the messages in our daily standup channel. And what it's going to do is it's This is an accountability coach prompt. And so you are an accountability coach tasked with ensuring our team consistently prioritizes high-leverage activities to maximize impact for our customers and business. And then here's our requirements, and it's going to score everyone every single day. You know, ultimately for your company, you want people working on the highest leverage activities. We do the same thing for our EOS Gong coach, which means these are our internal team lead meetings or team meetings where they're talking about perhaps product issues or client issues and things like that. But we want these meetings to be ran well ultimately. And so, you know, in this case, we're looking for This is for our internal team meetings. Let's say it's on the SEO team or the paid media team. And then what it's going to do, it's going to score these and it's going to save them into a Google Sheet. And then from there, it's going to score it and we can see over time if the person managing this or that meeting is going well over time or not. And so this in effect creates new work that wasn't possible before because it would just cost too much time and effort for any human being to do this. And now this can be done by robots. And so I genuinely think, I mean, these, you can call these agentic flows, but these are actual, these are working autonomously. You're just prompting it and it's constantly moving. Here's the problem with most marketing agencies. They don't understand what is happening with the world of AI. With my agency, Single Grain, we think about agentic workflows. We think about creating agents for the future. When we think about SEO, we think about creating different products such as internal linking products or content consolidation products. And on the other side of things, we think about building products such just care. We just think about making marketing easier overall. If you want to stay at the cutting edge of marketing and at the same time leverage the power of AI, go to singlegrain.com and we'll see you on the other side. Now that you know how to use the best LLMs and how to build agents, you're already ahead of 99% of people when it comes to AI. To bridge that 1% gap, you need to stay updated with the rapid changes happening in AI. To help you get started, here are some other tools and other things that you should be paying attention to. As an example, you can use Lindy or you can use genspark.ai and it can make phone calls for you. It can text and email people for you as well. So imagine The little standup bot that I have, instead of just checking in with people and giving feedback through Slack, you can actually call people to give them that feedback. If you want to do that, you can text people to give them that feedback as well. For example, when we have people reach out to us for our certain products, if they look like they're qualified, the bot can qualify them based on our ICP or our ideal client profile. And maybe you'll send them a text message trying to book a call with them. It'll instantly send them an email as well. And by sending them a text message and email instantly, it's going to do a much better job of getting people booked a lot faster because just because a lead comes in, oftentimes people will not even reach out to that lead. You can use ChatGPT with the different models that I talked about. You can use O3 for deeper reasoning. You can combine O3 with Deep Research for really deep consultant-level research that you'd pay tens of thousands of dollars for. And you can even use it for coding. You can use Super Whisper to talk a lot faster because we talk a lot faster than we type and dictate the prompts that you want and get what you need a lot faster. Cursor has helped make our engineers 10 times faster. It will work with you, create more code. It'll also spot errors with your code as well. You can even use ChatGPT Codex, which is fairly new, but it's also an operator that will look at your code and do the work with you. It'll also function as a different employee, spot errors with your code, and do more than that. You can use Manus to come up with 21 production-grade pieces of content, do long-tail keyword research for you, scan your blog to figure out the writer guidelines, and then output the blog post for you. You can use Superhuman to help draft emails for you. Superhuman does a good job of preparing draft responses in your draft folder. So I'm actually using drafts a lot more than I've ever in my life. And so my point here is that these are a couple of different use cases that you can use, but the only way you're going to get to a level where you're more autonomous with this stuff is that if you're constantly experimenting and you're constantly dedicating maybe 10 to 20% of your time on this, because the best CEOs and leaders that I know, they are absolutely obsessed with the world of AI right now because they know how much more value it's going to add, not only for their business, but mainly for their customers. All right. So congratulations on getting to the end of this video. Continue to invest in yourself and your future and you'll see success guaranteed. If you want any of the prompts from this video, hit the PDF in the comments. For more information on how I'm personally using AI to become a top marketer, you can check out This video over here.