WEBVTT 00:00:00.000 --> 00:00:02.439 Welcome to the deep dive. Have you ever looked 00:00:02.439 --> 00:00:04.960 at your to -do list and just felt, well, like 00:00:04.960 --> 00:00:07.919 you're stuck in Groundhog Day, copying and pasting, 00:00:07.980 --> 00:00:11.339 checking one app, updating another. It's manual, 00:00:11.500 --> 00:00:13.759 it's repetitive, and let's be honest, it's pretty 00:00:13.759 --> 00:00:16.280 graining. It really is. So much time gets lost 00:00:16.280 --> 00:00:19.079 there. Exactly. So what if you could literally 00:00:19.079 --> 00:00:23.539 hire an army of tireless digital assistants powered 00:00:23.539 --> 00:00:27.980 by AI working 2047 in the background? Imagine 00:00:27.980 --> 00:00:30.829 how much time that would free up. time for the 00:00:30.829 --> 00:00:35.630 truly strategic, creative, fulfilling work you 00:00:35.630 --> 00:00:37.789 actually want to be doing. That's the dream, 00:00:37.789 --> 00:00:40.869 right? Right. So today, our mission is to explore 00:00:40.869 --> 00:00:43.729 how you can create your very own AI agents. We'll 00:00:43.729 --> 00:00:46.149 use a platform called NEN, and you don't even 00:00:46.149 --> 00:00:48.530 need to write code. Yeah, no coding needed. And 00:00:48.530 --> 00:00:50.270 we want to understand the profound difference 00:00:50.270 --> 00:00:52.229 these agents can actually make in your work, 00:00:52.310 --> 00:00:55.070 maybe even your life. And what's truly transformative 00:00:55.070 --> 00:00:57.109 here, I think, is that we're moving beyond just 00:00:57.109 --> 00:00:59.590 asking AI questions, like you do with ChatGPT. 00:00:59.649 --> 00:01:01.890 We're talking about building entire systems, 00:01:02.390 --> 00:01:05.349 systems that operate autonomously. So by the 00:01:05.349 --> 00:01:07.569 end of this deep dive, you'll not only get what 00:01:07.569 --> 00:01:10.090 differentiates these AI agents, but you'll probably 00:01:10.090 --> 00:01:13.489 have a few aha moments. Aha moments, nice. Yeah, 00:01:13.530 --> 00:01:15.629 as we connect the dots on how you can scale your 00:01:15.629 --> 00:01:19.680 cognitive output. exponentially. This isn't just 00:01:19.680 --> 00:01:23.060 about saving an hour here or there, it's fundamentally 00:01:23.060 --> 00:01:25.760 restructuring how you can operate. That distinction 00:01:25.760 --> 00:01:27.739 is really key, isn't it? Because when most people 00:01:27.739 --> 00:01:30.280 hear AI, they immediately think of tools like 00:01:30.280 --> 00:01:34.680 chat GPT, or Claude, or Gemini. You pipe a punk, 00:01:34.760 --> 00:01:36.780 you get a response, and that interaction, well, 00:01:36.780 --> 00:01:39.680 it typically ends right there. But AI agents 00:01:39.680 --> 00:01:41.700 are, as you said, a completely different beast. 00:01:41.980 --> 00:01:44.459 So what's the biggest misconception we need to 00:01:44.459 --> 00:01:47.060 clear up right away? What makes these agents 00:01:47.060 --> 00:01:50.180 so fundamentally unique? That's a great question, 00:01:50.219 --> 00:01:52.680 because it's a critical difference. Think of 00:01:52.680 --> 00:01:56.189 AI agents more like your Digital employees. Digital 00:01:56.189 --> 00:01:58.950 employees, okay. Unlike a simple prompt response 00:01:58.950 --> 00:02:01.730 system, these digital employees possess some 00:02:01.730 --> 00:02:04.750 truly remarkable capabilities. First, they can 00:02:04.750 --> 00:02:07.870 reason. Reason. How so? Well, it means they don't 00:02:07.870 --> 00:02:10.590 just follow a script. They can actually analyze 00:02:10.590 --> 00:02:13.490 a complex problem, break it down into logical 00:02:13.490 --> 00:02:16.550 sequential steps, and then execute those steps 00:02:16.550 --> 00:02:19.409 to achieve a final goal. Oh, okay. It's like 00:02:19.409 --> 00:02:21.189 giving them a mission and letting them figure 00:02:21.189 --> 00:02:24.250 out the best path forward. So it's not just internal 00:02:24.250 --> 00:02:26.490 logic then. You mentioned they also have tool 00:02:26.490 --> 00:02:29.370 use. Exactly. Meaning they're not confined to 00:02:29.370 --> 00:02:31.789 just one application. They can interact with 00:02:31.789 --> 00:02:34.650 multiple apps simultaneously. Like what? We're 00:02:34.650 --> 00:02:37.030 talking Google Docs, your CRM, that's your Customer 00:02:37.030 --> 00:02:40.169 Relationship Management System, Slack, even internal 00:02:40.169 --> 00:02:42.710 databases. They can pull information from one 00:02:42.710 --> 00:02:44.710 place and use it in another. Okay, that's powerful. 00:02:44.930 --> 00:02:48.069 And crucially, They exhibit autonomous decision 00:02:48.069 --> 00:02:50.550 making. Meaning? Meaning you don't need to hand 00:02:50.550 --> 00:02:52.569 hold them through every single step. You give 00:02:52.569 --> 00:02:54.789 them an initial objective and they can decide 00:02:54.789 --> 00:02:56.789 which tool to use and what action to perform 00:02:56.789 --> 00:02:59.990 next. All on their own. So they adapt. They adapt 00:02:59.990 --> 00:03:02.550 to the situation as it unfolds. And perhaps the 00:03:02.550 --> 00:03:06.229 most impactful part for your everyday life, persistent 00:03:06.229 --> 00:03:09.129 operation. Oh yeah, that's huge. They can work 00:03:09.129 --> 00:03:11.849 continuously in the background. triggered by 00:03:11.849 --> 00:03:15.849 events or schedules, 204 .7. So while you're 00:03:15.849 --> 00:03:19.310 focused on high value work or even sleeping or 00:03:19.310 --> 00:03:22.090 on vacation, they're still grinding away, making 00:03:22.090 --> 00:03:25.030 progress for you. Exactly. To really make this 00:03:25.030 --> 00:03:27.750 tangible, let's consider a common sales process. 00:03:27.810 --> 00:03:30.770 Okay. Manually, you'd be checking emails for 00:03:30.770 --> 00:03:33.919 new leads, right? Copying their info. Yeah, maybe 00:03:33.919 --> 00:03:36.379 googling their company. Yeah, then opening your 00:03:36.379 --> 00:03:39.500 CRM creating a contact right drafting a personalized 00:03:39.500 --> 00:03:42.340 welcome email Setting a reminder to follow up. 00:03:42.439 --> 00:03:44.539 It's a lot of clicks a lot of context switching 00:03:44.539 --> 00:03:47.860 and mental overhead It really is it is now imagine 00:03:47.860 --> 00:03:50.319 an AI agent handling that okay. How would that 00:03:50.319 --> 00:03:53.680 look it detects a new lead email? Autonomously 00:03:53.680 --> 00:03:56.819 researches the company online Analyzes that information 00:03:56.819 --> 00:03:59.719 to draft an outreach email perfectly tailored 00:03:59.719 --> 00:04:02.389 to their industry Wow sends the email creates 00:04:02.389 --> 00:04:05.150 the contact in your CRM, and even schedules the 00:04:05.150 --> 00:04:07.169 follow -up task. All without me lifting a finger. 00:04:07.370 --> 00:04:09.169 Pretty much. That's a game changer. Yeah, no 00:04:09.169 --> 00:04:11.229 kidding. And this is a really critical distinction 00:04:11.229 --> 00:04:14.030 from, say, traditional automation tools like 00:04:14.030 --> 00:04:16.430 Zapier or Make. Right. I use those. They're more 00:04:16.430 --> 00:04:19.310 like, if this, then that. Exactly. So those platforms 00:04:19.310 --> 00:04:22.569 are fantastic. For those, if A happens, then 00:04:22.569 --> 00:04:25.600 do B rules. They follow instructions precisely. 00:04:25.680 --> 00:04:27.500 Or they don't think. Right. They can't really 00:04:27.500 --> 00:04:30.920 adapt if something unexpected occurs. AI agents, 00:04:31.019 --> 00:04:33.740 in contrast, possess these reasoning capabilities. 00:04:34.139 --> 00:04:37.060 They can handle complex, maybe ambiguous situations 00:04:37.060 --> 00:04:39.579 and make intelligent decisions beyond simple 00:04:39.579 --> 00:04:42.319 rule following. Just like a smart human assistant 00:04:42.319 --> 00:04:44.720 would. Precisely. That makes perfect sense. Yeah. 00:04:45.259 --> 00:04:49.560 So with that clarity on what AI agents are, Let's 00:04:49.560 --> 00:04:51.379 look at the tools. How do we actually build them? 00:04:51.439 --> 00:04:54.180 And that brings us to NEN. NEN, yeah. It's an 00:04:54.180 --> 00:04:57.480 open source tool, visual interface. You literally 00:04:57.480 --> 00:05:00.860 drag and drop these nodes like digital Lego bricks. 00:05:01.019 --> 00:05:02.920 That's a great analogy. Digital Lego bricks. 00:05:03.160 --> 00:05:05.319 You connect them to build your automation system. 00:05:05.399 --> 00:05:07.879 OK. And what's really powerful about NEN is its 00:05:07.879 --> 00:05:09.800 visual workflow builder. It lets you clearly 00:05:09.800 --> 00:05:12.459 see the entire flow. Data, logic, everything. 00:05:12.680 --> 00:05:16.079 Makes complex stuff much easier to design and 00:05:16.079 --> 00:05:18.399 understand. And it connects to lots of things. 00:05:18.569 --> 00:05:21.389 Yeah, massive connectivity. Supports over 400 00:05:21.389 --> 00:05:23.189 different apps and services right out of the 00:05:23.189 --> 00:05:25.790 box. Plus, you can connect to virtually any API 00:05:25.790 --> 00:05:28.689 out there. Wow. And being open source is a huge 00:05:28.689 --> 00:05:31.189 advantage, too. Gives you maximum freedom, no 00:05:31.189 --> 00:05:33.230 vendor lock -in. And here's a big one for a lot 00:05:33.230 --> 00:05:36.990 of people. The self -hosting option. Crucial. 00:05:37.430 --> 00:05:40.529 You can install innateN on your own. server, 00:05:40.970 --> 00:05:43.750 meaning all your data, all your workflows, they 00:05:43.750 --> 00:05:46.069 stay private, completely secure. It's your data, 00:05:46.310 --> 00:05:48.810 in your control, absolutely. Which makes it incredibly 00:05:48.810 --> 00:05:51.990 versatile. I can see this being used by content 00:05:51.990 --> 00:05:54.689 creators automating social media posts, video 00:05:54.689 --> 00:05:57.189 descriptions. Totally. Or businesses handling 00:05:57.189 --> 00:05:59.529 lid capture customer support. Even for personal 00:05:59.529 --> 00:06:02.100 use, right? tracking goals, summarizing daily 00:06:02.100 --> 00:06:04.220 news, maybe? Absolutely. The possibilities are 00:06:04.220 --> 00:06:08.300 really wide open. Now, to power NEN and these 00:06:08.300 --> 00:06:11.240 tireless digital employees, you need a robust 00:06:11.240 --> 00:06:13.259 engine behind the scenes. Right. You need somewhere 00:06:13.259 --> 00:06:15.699 to run it. And for those looking for a straightforward 00:06:15.699 --> 00:06:18.040 way to get NEN online, especially when you're 00:06:18.040 --> 00:06:21.300 just starting out, a virtual private server of 00:06:21.300 --> 00:06:25.050 VPS is often the best route. We found Hostinger 00:06:25.050 --> 00:06:27.230 provides a great solution for this. Makes it 00:06:27.230 --> 00:06:29.750 surprisingly easy, even for beginners. Yeah, 00:06:29.750 --> 00:06:31.769 they've streamlined it quite a bit. They offer 00:06:31.769 --> 00:06:35.529 a simple one -click N8n installation. One click. 00:06:35.769 --> 00:06:38.089 Nice. Yeah, you don't need to be a server expert. 00:06:38.110 --> 00:06:41.230 You just select their pre -configured N8n option 00:06:41.230 --> 00:06:43.550 during setup, and it pretty much handles the 00:06:43.550 --> 00:06:46.029 rest. Okay, and we should mention all the specific 00:06:46.029 --> 00:06:48.930 plan recommendations and detailed instructions. 00:06:49.069 --> 00:06:50.810 Those are in our show notes. Definitely check 00:06:50.810 --> 00:06:52.709 those out for the step -by -step. Absolutely. 00:06:53.100 --> 00:06:55.040 And it's worth noting, it's generally much more 00:06:55.040 --> 00:06:57.980 affordable than N8n's own cloud service, especially 00:06:57.980 --> 00:06:59.779 if you start building more workflows so they 00:06:59.779 --> 00:07:02.779 get more complex. Right. Cost matters. Plus, 00:07:03.079 --> 00:07:05.439 you get superior performance with dedicated resources, 00:07:05.920 --> 00:07:09.019 your complex workflows run smoothly, and like 00:07:09.019 --> 00:07:11.319 we said, full control. It's your server, your 00:07:11.319 --> 00:07:13.699 data, your security. Makes sense. All the login 00:07:13.699 --> 00:07:16.180 details, IP address, that'll be in your hosting 00:07:16.180 --> 00:07:19.329 or dashboard then. Exactly. OK, so. Anynion is 00:07:19.329 --> 00:07:21.970 up and running on our Hostinger VPS. You're greeted 00:07:21.970 --> 00:07:25.750 by its Canvas interface, ready to build. Let's 00:07:25.750 --> 00:07:28.430 walk through building a practical AI agent. Let's 00:07:28.430 --> 00:07:31.490 do it. A great example is a comprehensive YouTube 00:07:31.490 --> 00:07:34.050 content system. OK, what would that do? This 00:07:34.050 --> 00:07:37.649 system would take a YouTube video URL and automatically 00:07:37.649 --> 00:07:40.990 generate, say, multiple engaging titles, a detailed 00:07:40.990 --> 00:07:44.600 description, SEO tags. Timestamps. Timestamps, 00:07:44.699 --> 00:07:47.459 nice. Yeah, and even in an email newsletter announcing 00:07:47.459 --> 00:07:49.720 the new video. Wow, okay, that covers a lot. 00:07:49.839 --> 00:07:52.500 It does. This workflow basically functions like 00:07:52.500 --> 00:07:54.939 a digital production line. It starts with a trigger, 00:07:55.399 --> 00:07:57.620 usually a form where you paste the video URL. 00:07:57.819 --> 00:08:00.470 Simple enough. Then comes data fetching. A node 00:08:00.470 --> 00:08:02.689 retrieves the video's transcript. OK, it needs 00:08:02.689 --> 00:08:05.550 the content. Right. Next, the AI brain, this 00:08:05.550 --> 00:08:07.850 is where the magic happens, analyzes that transcript 00:08:07.850 --> 00:08:10.689 and creates all the content. Titles, description, 00:08:11.110 --> 00:08:13.029 tags, newsletter, everything. The heavy lifting. 00:08:13.350 --> 00:08:16.589 Exactly. And finally, storage. It saves everything 00:08:16.589 --> 00:08:19.689 neatly into, say, a Google Docs document. OK, 00:08:19.889 --> 00:08:22.649 step by step. Step one, creating the trigger. 00:08:23.149 --> 00:08:25.160 How do we do that? Easy. You just start a new 00:08:25.160 --> 00:08:28.240 workflow on the N8n canvas. Give it a descriptive, 00:08:28.399 --> 00:08:31.360 name -like YouTube content creation system. And 00:08:31.360 --> 00:08:33.720 set up a form trigger. You give the form a title, 00:08:34.019 --> 00:08:37.000 like enter YouTube video URL, add a field for 00:08:37.000 --> 00:08:39.960 the URL itself. And remember to save your work 00:08:39.960 --> 00:08:43.919 often. N8n doesn't autosave. Ah, good tip. Save 00:08:43.919 --> 00:08:46.980 frequently. OK, step two, getting the video content, 00:08:47.299 --> 00:08:50.139 the transcript. Right. For that, we use an HTTP 00:08:50.139 --> 00:08:53.429 request node. This node lets NAN talk to pretty 00:08:53.429 --> 00:08:56.009 much any web service. We'll connect it to an 00:08:56.009 --> 00:08:58.590 external service, like a YouTube transcript or 00:08:58.590 --> 00:09:01.389 API. You can find these on platforms like Rapid 00:09:01.389 --> 00:09:04.009 API. Rapid API, what's that? Think of it like 00:09:04.009 --> 00:09:06.110 an app store, but for software connections, for 00:09:06.110 --> 00:09:08.559 APIs. Got it. You sign up there, subscribe to 00:09:08.559 --> 00:09:10.799 a transcription API. Many have free tiers to 00:09:10.799 --> 00:09:13.559 start. And then you paste your API key into the 00:09:13.559 --> 00:09:16.720 credential section of the HTTP request node in 00:09:16.720 --> 00:09:19.279 AEN. OK. Connect the API. And you configure the 00:09:19.279 --> 00:09:21.600 node to pull the YouTube URL directly from the 00:09:21.600 --> 00:09:23.980 form you just created in step one. Use a little 00:09:23.980 --> 00:09:25.740 expression editor. It's quite intuitive. Right. 00:09:25.799 --> 00:09:28.240 Connect the data flow. And a pro tip. Yeah. Always 00:09:28.240 --> 00:09:30.580 execute and test each node right after setting 00:09:30.580 --> 00:09:32.460 it up. Make sure you're getting the data back 00:09:32.460 --> 00:09:34.700 that you expect before moving on. Test as you 00:09:34.700 --> 00:09:38.649 go. Smart. Okay, now for step three, the AI brain, 00:09:39.110 --> 00:09:41.549 the OpenAI integration. Yes, the core intelligence. 00:09:42.649 --> 00:09:45.990 Every AI agent node in ANN typically has three 00:09:45.990 --> 00:09:47.710 main components you might use. There's the chat 00:09:47.710 --> 00:09:52.429 model, like OpenAI's GPT -40 or GPT -3 .5 turbo. 00:09:52.590 --> 00:09:54.690 Okay, the engine. Then there's memory, which 00:09:54.690 --> 00:09:56.909 is mainly useful for chat box so they can remember 00:09:56.909 --> 00:09:59.309 the context of a conversation. We might not need 00:09:59.309 --> 00:10:01.679 it much for this specific workflow. Right. And 00:10:01.679 --> 00:10:03.860 then there are tools. These are actions the AI 00:10:03.860 --> 00:10:06.899 can perform, like searching the web or, in our 00:10:06.899 --> 00:10:09.399 case, creating documents. Got it. So how do we 00:10:09.399 --> 00:10:12.500 set up the OpenAI part? First, you need an API 00:10:12.500 --> 00:10:15.980 key from OpenAI. You get that from platform .openai 00:10:15.980 --> 00:10:19.200 .com. Quick note. When you generate that key, 00:10:19.659 --> 00:10:21.419 copy and save it somewhere secure immediately. 00:10:21.659 --> 00:10:23.779 Why? Because they only show it to you once. For 00:10:23.779 --> 00:10:25.809 security reasons. Oh, OK. Good to know. Copy 00:10:25.809 --> 00:10:28.230 it right away. Yep. Then back in AI, you add 00:10:28.230 --> 00:10:31.289 an AI agent node, choose OpenAI Chat Model as 00:10:31.289 --> 00:10:33.409 the model, click Create New for Credentials, 00:10:33.509 --> 00:10:35.809 give it a name like my OpenAI key, paste your 00:10:35.809 --> 00:10:37.210 API key in there. It's straightforward. Then 00:10:37.210 --> 00:10:40.149 you select your desired model. Maybe GPT -40 00:10:40.149 --> 00:10:42.809 for the absolute best quality. Or you could use 00:10:42.809 --> 00:10:45.950 GPT -3 .5 Turbo, which is much cheaper and still 00:10:45.950 --> 00:10:48.289 very capable, especially if you're watching costs. 00:10:48.590 --> 00:10:52.070 OK, choices depending on need and budget. Exactly. 00:10:52.509 --> 00:10:54.909 And since we want the AI to save the content 00:10:54.909 --> 00:10:57.470 it generates, we'll connect Google Docs as a 00:10:57.470 --> 00:11:00.590 tool within that same AI agent node. Ah, OK, 00:11:00.690 --> 00:11:03.029 the tool connection. Right. You click Add Tool. 00:11:03.860 --> 00:11:06.240 select Google Docs, you'll need to create credentials 00:11:06.240 --> 00:11:09.460 for Google too. Is that complicated? NEN actually 00:11:09.460 --> 00:11:12.139 provides really clear step -by -step guidance 00:11:12.139 --> 00:11:14.559 for setting this up in Google's developer portal. 00:11:14.759 --> 00:11:17.559 It walks you right through it. Q, okay. Once 00:11:17.559 --> 00:11:19.759 connected, you choose the append to a document 00:11:19.759 --> 00:11:21.919 action. Then you just need the document ID. Where 00:11:21.919 --> 00:11:23.840 do I get that? Just open the Google Doc you want 00:11:23.840 --> 00:11:26.320 to use, look at the URL in your browser. It's 00:11:26.320 --> 00:11:28.299 a long string of letters and numbers in the middle. 00:11:28.649 --> 00:11:31.610 Copy that, paste it into the document ID field 00:11:31.610 --> 00:11:34.789 in NEN. Okay, grab the ID from the URL. Got it. 00:11:35.009 --> 00:11:37.830 And now for step four, which is arguably the 00:11:37.830 --> 00:11:40.009 most important part. The prompt. Writing the 00:11:40.009 --> 00:11:43.289 AI instructions. Comped engineering. The quality 00:11:43.289 --> 00:11:46.509 of your AI agent's output depends entirely on 00:11:46.509 --> 00:11:48.889 the clarity and specificity of your instructions 00:11:48.889 --> 00:11:52.070 here. Like delegating to a real employee. Exactly. 00:11:52.330 --> 00:11:54.029 The clearer the instructions, the better the 00:11:54.029 --> 00:11:55.710 result. You want to build a structured prompt. 00:11:55.710 --> 00:11:59.240 Give it a clear role first, like You are a digital 00:11:59.240 --> 00:12:01.480 marketing mastermind specializing in YouTube 00:12:01.480 --> 00:12:03.860 content. This kind of role -playing helps the 00:12:03.860 --> 00:12:06.019 AI adopt the right tone and perspective. Okay, 00:12:06.100 --> 00:12:08.519 set the stage. Then provide context. It will 00:12:08.519 --> 00:12:11.019 receive a video transcript. Outline the request 00:12:11.019 --> 00:12:14.360 in detail. Specify formats, like using Markdown 00:12:14.360 --> 00:12:18.120 for easy copying. Ask for, say, five distinct 00:12:18.120 --> 00:12:20.379 video title options. Much titles, okay. A full 00:12:20.379 --> 00:12:22.919 video description, maybe specify needing a strong 00:12:22.919 --> 00:12:26.519 hook, detailed content summary, timestamps extraction. 00:12:26.659 --> 00:12:28.720 Oh, it can pull out timestamps. If you ask it 00:12:28.720 --> 00:12:31.039 clearly and provide the transcript, yes. Often 00:12:31.039 --> 00:12:33.259 it can identify key moments. And maybe a call 00:12:33.259 --> 00:12:36.700 to action. Nice. Also ask for maybe 15, 20 optimized 00:12:36.700 --> 00:12:41.039 SEO tags. A draft for a new video email newsletter. 00:12:41.450 --> 00:12:44.129 Maybe even social media posts suggestions for 00:12:44.129 --> 00:12:46.509 LinkedIn and TwitterX. Wow, that's a lot from 00:12:46.509 --> 00:12:49.830 one prompt. It is. And crucially, you tell the 00:12:49.830 --> 00:12:53.250 AI agent node to use the transcript data coming 00:12:53.250 --> 00:12:55.909 in from the previous step, the HTTP request node, 00:12:56.210 --> 00:12:59.090 as the input for this prompt. Use the expression 00:12:59.090 --> 00:13:01.730 editor again to link that data. Feed the transcript 00:13:01.730 --> 00:13:05.350 into the prompt. Okay, truly comprehensive. Finally, 00:13:05.590 --> 00:13:09.929 step five. Test, refine, and activate. The moment 00:13:09.929 --> 00:13:13.220 of truth. You click. Execute workflow, top right. 00:13:13.740 --> 00:13:15.720 Then you open your form trigger, usually via 00:13:15.720 --> 00:13:19.120 its test URL, paste in a real YouTube URL, and 00:13:19.120 --> 00:13:22.080 hit submit. And watch any on end go. Yep. Watch 00:13:22.080 --> 00:13:24.720 the nodes execute one by one. Yeah. Green means 00:13:24.720 --> 00:13:27.559 success. Red means an error. What causes errors 00:13:27.559 --> 00:13:29.860 usually? Often it's an incorrect data reference. 00:13:30.639 --> 00:13:32.639 Maybe you pointed to the wrong data from a previous 00:13:32.639 --> 00:13:35.259 node. Or the prompt might need tweaking if the 00:13:35.259 --> 00:13:37.460 AI output isn't quite right. You just investigate 00:13:37.460 --> 00:13:39.360 the node that failed, check its input and output 00:13:39.360 --> 00:13:41.679 tabs. Debugging. Once you're happy with the output 00:13:41.679 --> 00:13:44.200 in your Google Doc, you just toggle the active 00:13:44.200 --> 00:13:46.220 switch on the workflow editor. And it's live. 00:13:46.440 --> 00:13:48.519 It's live. Ready to run automatically every time 00:13:48.519 --> 00:13:51.899 you submit a URL to that form. That's amazing. 00:13:52.679 --> 00:13:54.360 What's so exciting about this system is that, 00:13:54.360 --> 00:13:56.100 as you said, it feels like just the beginning. 00:13:56.299 --> 00:13:59.129 Oh, absolutely. The true power of NAN lies in 00:13:59.129 --> 00:14:02.370 its limitless scalability. This is just one workflow. 00:14:02.509 --> 00:14:04.909 You can add more advanced features. Definitely. 00:14:05.169 --> 00:14:07.789 You can add things like IF nodes or switch nodes. 00:14:08.110 --> 00:14:10.389 These allow for smart splitting and routing of 00:14:10.389 --> 00:14:13.669 your workflow. Example. Well... maybe based on 00:14:13.669 --> 00:14:16.350 the video topic detected by the AI, you could 00:14:16.350 --> 00:14:19.190 use a switch node to save content about Python 00:14:19.190 --> 00:14:22.570 to a specific programming folder in Google Drive 00:14:22.570 --> 00:14:26.070 and content about marketing to a different business 00:14:26.070 --> 00:14:28.669 folder. Ah, intelligence sorting. That kind of 00:14:28.669 --> 00:14:30.610 intelligence is fascinating. You could even build 00:14:30.610 --> 00:14:32.950 a feedback loop. Imagine setting up another workflow. 00:14:33.049 --> 00:14:35.610 This one runs, say, weekly. It pulls performance 00:14:35.610 --> 00:14:38.149 data views, watch time, likes directly from the 00:14:38.149 --> 00:14:40.289 YouTube API for the videos you process. And get 00:14:40.289 --> 00:14:43.590 the stats back. Right. Then you feed that performance 00:14:43.590 --> 00:14:46.429 data back into another AI agent node, and you 00:14:46.429 --> 00:14:49.049 ask it. Based on these metrics, which titles 00:14:49.049 --> 00:14:51.950 and topics performed best? Suggest five ideas 00:14:51.950 --> 00:14:55.029 for the next video. Oh, continuous automated 00:14:55.029 --> 00:14:57.330 improvement. Exactly. Yeah. That's when things 00:14:57.330 --> 00:14:59.730 get really powerful. And the integration possibilities, 00:14:59.950 --> 00:15:02.110 you mentioned over 400 apps. So you could automate 00:15:02.110 --> 00:15:05.360 creating deals in your CRM. Like HubSpot. Yep. 00:15:05.639 --> 00:15:08.279 Or post your AI -generated content directly to 00:15:08.279 --> 00:15:10.740 LinkedIn or Twitter using their respective notes. 00:15:11.120 --> 00:15:13.639 Add users to an email -nurturing sequence in 00:15:13.639 --> 00:15:17.399 MailChimp. Sure can. Or create task cards in 00:15:17.399 --> 00:15:20.720 Trello or Asana for your team based on the content 00:15:20.720 --> 00:15:23.480 generated. It's like building an entire ecosystem 00:15:23.480 --> 00:15:26.200 of digital assistants working together. That's 00:15:26.200 --> 00:15:28.580 the idea. an interconnected system. Okay, now 00:15:28.580 --> 00:15:29.919 when you're working with the data, especially 00:15:29.919 --> 00:15:32.600 connecting all these services and using potentially 00:15:32.600 --> 00:15:36.419 paid APIs, two critical things come to mind, 00:15:36.840 --> 00:15:39.659 security and cost. Very important considerations, 00:15:39.980 --> 00:15:43.059 absolutely. For security, number one rule, always 00:15:43.059 --> 00:15:46.059 store your API keys and any other sensitive information 00:15:46.059 --> 00:15:49.419 like passwords or tokens in NNN's built -in credential 00:15:49.419 --> 00:15:53.000 store. Not directly in the workflow nodes? Never. 00:15:53.179 --> 00:15:55.500 Never hard code them directly into your workflows. 00:15:56.019 --> 00:15:57.620 The credential store encrypts them and keeps 00:15:57.620 --> 00:15:59.440 them separate and safe. OK, use credentials. 00:15:59.580 --> 00:16:02.480 Good practice. And as we discussed, self -hosting 00:16:02.480 --> 00:16:05.840 NAN on your own VPS, like with Hostinger, is 00:16:05.840 --> 00:16:08.600 a critical layer of security for your data itself. 00:16:08.679 --> 00:16:11.179 It keeps everything fully within your control, 00:16:11.519 --> 00:16:13.799 not on someone else's cloud. Right, data privacy. 00:16:14.179 --> 00:16:17.259 And on the cost side, those OpenAI calls add 00:16:17.259 --> 00:16:19.840 up, right? They can, especially with models like 00:16:19.840 --> 00:16:23.570 GPT -4 .0. So... Optimize your prompts. Be concise 00:16:23.570 --> 00:16:26.649 but clear. Get the output you need without unnecessary 00:16:26.649 --> 00:16:28.490 length or back and forth if you can avoid it. 00:16:28.690 --> 00:16:31.370 Makes sense. Shorter prompts, fewer tokens, lower 00:16:31.370 --> 00:16:34.230 cost. Exactly. And consider using the cheaper 00:16:34.230 --> 00:16:37.950 models, like GPT 3 .5 Turbo, for simpler tasks 00:16:37.950 --> 00:16:40.190 that don't require deep reasoning or nuanced 00:16:40.190 --> 00:16:42.629 creativity. Reserve the more expensive models 00:16:42.629 --> 00:16:45.110 for where they really shine. Smart resource management. 00:16:45.389 --> 00:16:47.330 OK. So if we kind of connect this all back to 00:16:47.330 --> 00:16:50.210 the bigger picture, building these AI agents, 00:16:50.309 --> 00:16:52.539 well, It might seem complex at first glance. 00:16:52.820 --> 00:16:54.399 Yeah, looking at that workflow, it has a few 00:16:54.399 --> 00:16:56.220 steps. But as you've seen, it's really about 00:16:56.220 --> 00:16:58.240 connecting relatively simple building blocks, 00:16:58.340 --> 00:17:00.700 those nodes, in intelligent ways. Like digital 00:17:00.700 --> 00:17:04.079 Lego. Exactly. The key to success, honestly, 00:17:04.420 --> 00:17:07.559 is to start small. Build one piece. Test it. 00:17:07.769 --> 00:17:10.809 Add the next piece, test it. Incrementally. Yeah, 00:17:10.809 --> 00:17:12.829 test every single step, and then gradually build 00:17:12.829 --> 00:17:14.930 more complex systems as you get more comfortable 00:17:14.930 --> 00:17:17.049 and proficient with the tool. That YouTube content 00:17:17.049 --> 00:17:19.430 system we built today, it's just one example 00:17:19.430 --> 00:17:22.609 then. Just one instance. With this foundational 00:17:22.609 --> 00:17:26.589 knowledge, you truly can create AI agents to 00:17:26.589 --> 00:17:29.009 handle nearly any repetitive process you can 00:17:29.009 --> 00:17:31.750 think of in your work or life. It's leveraging 00:17:31.750 --> 00:17:34.549 intelligence, not just basic automation. That's 00:17:34.549 --> 00:17:37.660 the core difference, intelligence. So maybe we 00:17:37.660 --> 00:17:39.920 should recap some golden rules for success for 00:17:39.920 --> 00:17:41.579 someone listening who wants to try this. Yeah, 00:17:41.579 --> 00:17:44.980 definitely. Golden rules. One, always start simple. 00:17:45.119 --> 00:17:47.160 Expand gradually. Don't try to build Rome in 00:17:47.160 --> 00:17:49.819 a day. Two, test each node independently as you 00:17:49.819 --> 00:17:51.859 build. Seriously, this saves so much headache 00:17:51.859 --> 00:17:54.720 later. Test, test, test. Three, invest time in 00:17:54.720 --> 00:17:57.839 writing clear, detailed prompts. Prompt engineering 00:17:57.839 --> 00:18:00.259 is absolutely crucial for getting good results 00:18:00.259 --> 00:18:03.079 from the AI. Garbage in, garbage out still applies. 00:18:03.339 --> 00:18:06.950 Right. Quality prompts matter. Four, always prioritize 00:18:06.950 --> 00:18:10.529 security and privacy. Use the credential store. 00:18:10.829 --> 00:18:13.069 Consider self -hosting for full control. Super 00:18:13.069 --> 00:18:16.269 important. And five, don't be afraid to experiment 00:18:16.269 --> 00:18:19.269 and innovate. The real magic happens when you 00:18:19.269 --> 00:18:21.289 just start playing around, trying things out, 00:18:21.750 --> 00:18:23.890 connecting different services. Okay, great rules. 00:18:24.309 --> 00:18:27.289 So what does this all mean for you, the listener? 00:18:27.430 --> 00:18:29.450 This isn't just about automation, is it? Not 00:18:29.450 --> 00:18:31.509 at all. It feels like it's about gaining back 00:18:31.509 --> 00:18:35.269 the most precious commodity we have. Time. Exactly. 00:18:35.769 --> 00:18:38.069 Time for strategic thinking. Time for creativity. 00:18:38.490 --> 00:18:40.869 Time for what truly matters to you. It's about 00:18:40.869 --> 00:18:43.609 scaling your impact without necessarily scaling 00:18:43.609 --> 00:18:46.410 your effort in the same way. Scaling impact without 00:18:46.410 --> 00:18:48.690 scaling effort. I like that. And I truly believe 00:18:48.690 --> 00:18:50.970 the future belongs not just to those who know 00:18:50.970 --> 00:18:54.470 how to use AI. Like using chat GBT. Right. But 00:18:54.470 --> 00:18:56.880 perhaps even more so. to those who know how to 00:18:56.880 --> 00:18:59.220 build and direct it, how to orchestrate these 00:18:59.220 --> 00:19:02.539 agents. By creating your own AI agents, you're 00:19:02.539 --> 00:19:04.519 developing what I think is genuinely one of the 00:19:04.519 --> 00:19:07.099 most valuable skills of the next decade. Building 00:19:07.099 --> 00:19:10.119 and directing AI. Wow. So we really encourage 00:19:10.119 --> 00:19:12.960 you. Pick just one tedious task, something you 00:19:12.960 --> 00:19:15.480 do every day or every week that drains you. Break 00:19:15.480 --> 00:19:17.440 it down into steps. Like we did with the sales 00:19:17.440 --> 00:19:20.400 process example. Precisely. and then start building 00:19:20.400 --> 00:19:23.180 your own little digital assistant in ADN to handle 00:19:23.180 --> 00:19:25.480 it. See how quickly you can transform that part 00:19:25.480 --> 00:19:28.200 of your workflow. Your first AI agent awaits. 00:19:28.940 --> 00:19:31.619 It sounds achievable now. It really is. The tools 00:19:31.619 --> 00:19:33.779 are ready, the instructions like we walk through 00:19:33.779 --> 00:19:35.640 are pretty clear, and the possibilities seem 00:19:35.640 --> 00:19:38.400 endless. They really are. And just a reminder, 00:19:38.640 --> 00:19:40.380 you can find all the resources we mentioned, 00:19:40.539 --> 00:19:42.720 the hosting or setup link, with that special 00:19:42.720 --> 00:19:46.059 discount code, Aurelius, for 10 % off your first 00:19:46.059 --> 00:19:48.039 order. It's all in our show notes. Yep, check 00:19:48.039 --> 00:19:49.819 those out. And if you want to take your AI skills 00:19:49.819 --> 00:19:53.319 even further, dive deeper. Check out the AI Fire 00:19:53.319 --> 00:19:56.740 Academy Premium Plan. There's a 14 -day free 00:19:56.740 --> 00:19:59.670 trial. What's in there? Instant access to over 00:19:59.670 --> 00:20:03.690 500 pre -built AI workflows, advanced tutorials, 00:20:04.190 --> 00:20:06.650 exclusive case studies. It's a great resource 00:20:06.650 --> 00:20:09.289 to accelerate your learning. Fantastic. So really 00:20:09.289 --> 00:20:13.650 it's time to dive in and unleash your own digital 00:20:13.650 --> 00:20:15.410 workforce. Go build something cool.