Stop Using GPT-5 Like GPT-4 - Do This Instead

Speaker A: If you clicked on this video, you're probably one of two people. Either you just got access to GPT5 and you're treating it like a slightly faster version of GPT4, or you've been using it for weeks and wondering why you're not seeing massive improvements like outliers on social media are. Here's the brutal truth. GPT5 isn't just GPT4 with a speed boost. It's a completely different beast. By the end of this video, you'll know exactly how to leverage GPT5 advanced reasoning capabilities, prompt agentically and unlock things you never thought were possible. So essentially what you'll learn are these core areas. Number one, the three hidden GPT5 features that change everything. Number two, advanced prompt engineering techniques specific to GPT5. And number three, how to use GPT5's extended context window effectively. Okay, These examples include a team of rivals to solve any problem, ejectic workflow and prompting, and how to even tap into your own data for insane results in seconds to grow revenue instantly. Okay, for those of you who don't know me, my name is Ryan Staley and after leading teams to close deals with the largest companies in the world, I started unlocking unfair advantages about three years ago for individuals and executives to become superhuman with AI have helped over 2,300 executives and their teams. And I'm a huge believer that anything is possible right now. So let's get right into it. I'm going to start sharing my screen as I always do and walk you through this. Okay, I have pre built some of these out. Uh, and the reason being is because I want to make sure that you truly have a good sense of what it looks like and it's fast. Okay? However, this is the team of rivals. Okay, this is the one that I mentioned to solve any problem. So this is the prompt. Analyze any business decision from the perspectives of the customer, shareholder, employee, competitor and regulator. Maintain each perspective's internal logic while identifying the conflicts. And so I put in the example how should Shopify approach AI transformation based on the leaked memo of the CEO Toby. Okay, for those of you who don't know this, what I'm going to do is I'm going to re execute this and explain while this is building out in real time so you can see how fast it is. So basically what happened with this leaked memo is Toby was talking about how they are an AI first company and they needed to justify if they wanted more headcount they couldn't do use it for AI basically so that's a massively oversimplified version. Um, but here's what was really good about it. Okay, so it gave us examples from different views. We have the customer right with benefit from AI enhanced tools, uh, more value, possible friction points. If AI push accelerates the cost of human judgment, merchants may lose personal support. Love that. Uh, shareholder example. Favor greater productivity and profitability. So they're going to like that. What's the potential conflicts? Let's see. Uh, rapid shifts may trigger internal chaos or talent drain if not managed well. And the beautiful thing about it is it's highlighting all these different examples as we go through here. Now it hits on the employee, right? Opportunity fluency becomes a core skill. Uh, performance link clarity. It's now part of reviews, fear of redundancy, insufficient empathy, and then work culture pressure. So all of these are examples, competitor pressure as well. Uh, this puts the emphasis on them to accelerate or being seen as laggards and then regular in society. Okay. Beautiful thing is it even creates a little, little cross, uh, perspective tensions, uh, with different examples, strategic recommendations and then a summary and then even sources at the end. Okay, so one of the things that I think is beautiful about this is just a, how fast it worked, which is why I executed it, but B, how, how many different perspectives it pulled from within seconds. So what we're going to do is we're going to go down to the next one. I am going to go and save agentic prompting for next for agent workflows. All right? For step number two, it's how to use connectors. So what you're seeing is the output of a prompt that I put in. Whereas Effectively I asked GPT5 and I used the thinking version to identify all the dreams, fears, frustrations and desires my contacts over the last 90 days have. Now, I'm not going to scroll up and show you the prompt and use case up there because it actually lists confidential information that I cannot share. However, what I want to share with you, uh, is how to access this because it's not just limited to your CRM. In this case, I'm connected to HubSpot, as you could see. But if you look at it, there's a button. If you click the plus button, there'll be all these different modes you can leverage. All right? So the one we want to focus on is use connectors. Now, one of the things you're going to need to do is enable connection to these different tools or systems to be able to do that. And when you do, you'll be able to have an example like I connected HubSpot here where I could toggle it on and off. And if I toggle it, uh, on, I can basically just go in and ask and say, okay, give me like the ICP trends or give me trends on what my best customers are doing or the contacts in there or notes. It's surprisingly comprehensive and you could create content, sales, scripts, messaging, product ideas from all those different areas. Now these other areas I don't have set up yet, but you can connect to Google Drive as well. Box. Um, and you can even connect more through here. Uh, as you could see, you could browse them or you could create your own. All right, so a ton of opportunities available. But what I would say is like these dreams, fears, frustrations, desires are really close to what I would say just from having a lot of hand holding and one, um, on one conversation with my clients in terms of what they're doing. Okay, so what I'm going to do is I'm going to switch to different mode so I could pull this up or I should say a different account that I have because they do have the Pro account. As I've mentioned, I'm going to show the last and final case study about agentic workflow prompting. Okay, so so far we covered the team arrivals and then we tapped into our own data for insane insight. Now what I'm going to do is I'm going to go into a complete example. You might remember this one, and I'm going to show it to you again because this is something that I hit on in a previous video. Okay. And this was something around use cases that 95% of people miss when they're leveraging GPT5. So if you didn't watch that, go back and check it out. It's absolute fire with how to prompt specifically in the GPT5. One of the areas that I covered was really how to prompt agentically. And I'm going to go a little bit deeper on that now because I think that is one of the biggest moments, most insane outcomes I've ever seen an AI model do. And I'm going to go all the way to the top and I'm going to show you what happened. Okay. Uh, as you can see, this is pretty extensive. So I started off with this at the top. I said, all right, I want you to design a top 1% of 1% prompt and I need you to do this in a single pass. Okay. Uh, why this is super critical is because a single pass tells the model not to keep stopping and checking. Now what you're going to see here I did, just to get under a little bit deeper is I did a combination of of tasks or basically like a workflow in sequential order. So I said do deep market recon on the web and Reddit resource strategy angle matrix. So resource strategies like lead magnets and hooks, uh, narratives, proofs and platform fed. These are marketing components, full backlog of priorities, outlines, channel, packaging, experiment. Okay, so step one was I created the prompt first. Now I was really impressed with the level of this GTM kind of growth prompt is what it gave. So it gave everything from output, executive summary, resource and then all the details that I mentioned as you could see here. Now the thing that I think is really interesting is I say, okay, run this with a pre filled starter block and it ran it with because it basically mocked up one as an example of a hypothetical scenario. So what it does is it actually went through that and then if you look at this like it's really impressive the results that it Pulled up. Forrester McKinsey. Um, I don't know what this prevost is but um, other examples and it talks about what people are looking for and want talks about assumptions, a deep market recon. Okay. Uh, and then at the same time it goes through all the like triggers and detection signals, objections, resource strategy and how I could leverage it. So like this is all highly, highly relative and detailed data that I could use in my business. It's even got, you know, delivery and backlog how to prioritize these. It's got production pipeline. So I'm going to walk you through another example of it, but I wanted to show the framework for you. All right, so this is another example. Sorry I couldn't screen run it. Okay, so this is another one where I'm like, hey, act like a chief revenue officer, strategic advisor to and I use the company rider, which is a $200 million Series C. Um, and then I basically said systematically analyze the entire revenue funnel from lead generation to customer retention. Only complete your analysis when you have five specific optimization strategies. Now why that's critical is when you're using a personal account like this, this is a personal account. I don't need to worry about credits. I wouldn't put that command in if you are using like an API and you're trying to connect to a solution. Um, but the value in this is I could say only complete the analysis. We have five specific. Okay, I also said don't stop at surface metrics. Reason being if you have like an API connected to a tool it could go in an endless loop and you just burn tons of money with doing that. That's capped obviously with a uh, user account. So here's a real interesting thing. I came up with the snapshot, came up with goal math, came up with their whole basically tofu pipeline, health bill's velocity. It came up with everything, right? And then what it had started to do is walk through examples of different plays and the impact that you could use all the way ranging from uh, basically you know, pql, which is product qualified leads to cloud, uh, marketplace, vertical agent packages, pov, factory. So a lot of different things that it has within here. And then it's like, hey Ryan, this is great, do you want me to put this into a one page ops plan? And so I said yes, right? And then what it did is it built this out. So super useful from that perspective. I'll give uh, uh, one more example. And this is more along the lines of once again another marketing example, but I thought it was really solid as well. So um, I. Different situation, fun little hack here. Um, I saw a prompt that I liked online and so I created a snippet from it, like clipped it and then had it transcribed. And I said all right, basically gave a whole, you know, full data organic LinkedIn content plan, right? So basically it walks through with LinkedIn and it's trying to answer the questions, okay, how do I create an endless pipeline or how do I create content for LinkedIn that leads to results? And so what you're doing is it's breaking down everything from the algorithm to projected impact distribution version as well as budget and what I would need in that 12 week timeframe, right? So those are the three examples. So I'm gonna stop sharing for um, just to kind of round things out and to give you context as we bring this to a close. The one thing I'll tell you is if this video open your eyes, what GPT5 can really do. Here's what I want you to do right now. Number one, like this video, give me a little love man, I'm working hard on these videos. Still wanna get better and better for you. Number two, subscribe and there's gonna be more tutorials that are like this, that are gonna be more advanced coming out. Um, number three is try one of these techniques in your next session, okay? Ah, that's the best way to learn and be a master is to start just implementing and using it, even if it's imperfect. And number four, comment below with your results. Okay? If you want even deeper dive access to prompts like this that I showed in the video today. You can get more of them by clicking on the link below. They'll get access to those prompts as well as weekly updates to what I'm doing in situations like this to better serve you and help you become superhuman. So, uh, once again, thanks for joining me today, and I look forward to seeing you all on the next video.