WEBVTT 00:00:00.000 --> 00:00:02.299 Most people guess their way through automations. 00:00:02.779 --> 00:00:05.299 They copy templates and feel lost when they break. 00:00:05.799 --> 00:00:09.000 Beat. But today? We are shifting from guessing 00:00:09.000 --> 00:00:12.279 to architecting. Welcome to this deep dive. It's 00:00:12.279 --> 00:00:14.419 really great to be here today. We are exploring 00:00:14.419 --> 00:00:17.839 a master blueprint for reAANM. We'll cover the 00:00:17.839 --> 00:00:20.719 factory mindset and step zero triggers. We will 00:00:20.719 --> 00:00:23.039 also learn to read execution logs like a pro. 00:00:23.280 --> 00:00:26.739 Yeah, and we'll discuss scaling with a VPS plus 00:00:26.739 --> 00:00:29.679 logic routing and exactly where AI belongs in 00:00:29.679 --> 00:00:32.420 your workflow. Beat. So let's talk about that 00:00:32.420 --> 00:00:35.850 factory mindset. People usually see NN as just 00:00:35.850 --> 00:00:38.369 a way to connect apps. They drag a Google Sheets 00:00:38.369 --> 00:00:40.350 node onto the screen, then they try to wire it 00:00:40.350 --> 00:00:43.030 directly into Slack. It feels like jamming incompatible 00:00:43.030 --> 00:00:45.509 Lego blocks together. Right. And the pieces simply 00:00:45.509 --> 00:00:47.409 don't fit perfectly. You need to stop looking 00:00:47.409 --> 00:00:50.969 at isolated digital actions. Instead, see a system 00:00:50.969 --> 00:00:54.060 moving data smoothly through time. beat. So we 00:00:54.060 --> 00:00:56.079 treat our data like a physical object. Exactly. 00:00:56.140 --> 00:00:58.920 Think of it like a massive physical manufacturing 00:00:58.920 --> 00:01:01.920 factory. Let's trace a real world scenario today. 00:01:02.380 --> 00:01:05.140 Imagine a messy lead generation process for a 00:01:05.140 --> 00:01:07.359 growing business. Raw materials enter the building 00:01:07.359 --> 00:01:11.140 at the very start. Yes. And those materials change 00:01:11.140 --> 00:01:13.500 shape at different workstations. The factory 00:01:13.500 --> 00:01:16.420 makes automated decisions about the paths. Finally, 00:01:16.519 --> 00:01:18.680 a finished product leaves the building entirely. 00:01:19.099 --> 00:01:22.379 So a new customer fills out a website form. That 00:01:22.379 --> 00:01:24.560 form submission is the raw material entering 00:01:24.560 --> 00:01:26.879 our building. Right, and it starts moving down 00:01:26.879 --> 00:01:30.400 the factory conveyor belt. This mindset completely 00:01:30.400 --> 00:01:33.780 removes all the visual interface clutter. You 00:01:33.780 --> 00:01:36.980 stop seeing a confusing web of colorful brand 00:01:36.980 --> 00:01:39.640 logos. You stop worrying about what each specific 00:01:39.640 --> 00:01:42.159 app technically does. You just trace the physical 00:01:42.159 --> 00:01:44.739 path of your data. You look at a node and examine 00:01:44.739 --> 00:01:47.890 its specific job. How does this factory mindset 00:01:47.890 --> 00:01:50.790 change the way we look at a messy screen of icons? 00:01:51.049 --> 00:01:53.549 Well, it simplifies everything. You are no longer 00:01:53.549 --> 00:01:55.469 overwhelmed by the software brands. You just 00:01:55.469 --> 00:01:57.849 focus entirely on the data transformation itself. 00:01:58.069 --> 00:02:00.870 So we just ask, how is this node changing my 00:02:00.870 --> 00:02:03.450 data? Exactly. And that brings us to the start 00:02:03.450 --> 00:02:06.109 of the line. A lot of folks completely misunderstand 00:02:06.109 --> 00:02:08.090 the trigger node. They think it's just a simple 00:02:08.090 --> 00:02:11.509 automated start button. Beat. But triggers dictate 00:02:11.509 --> 00:02:14.990 your data's exact shape at step zero. They really 00:02:14.990 --> 00:02:17.750 do. There are three main types you'll use constantly. 00:02:18.069 --> 00:02:20.810 First, we have the basic scheduled time trigger 00:02:20.810 --> 00:02:23.849 node. It sends basic time data into your system 00:02:23.849 --> 00:02:26.550 regularly. Think of it as a clock ticking on 00:02:26.550 --> 00:02:28.969 the wall. Yeah, perfect for daily reports or 00:02:28.969 --> 00:02:32.389 weekly summaries. Next is the highly versatile 00:02:32.389 --> 00:02:36.189 webhook trigger node. A webhook is exactly like 00:02:36.189 --> 00:02:38.969 a digital doorbell handing off packages. Another 00:02:38.969 --> 00:02:41.259 app rings the bell. and drops off the payload. 00:02:41.659 --> 00:02:43.699 Finally, we have the straightforward form trigger 00:02:43.699 --> 00:02:46.479 node. This captures exactly what users typed 00:02:46.479 --> 00:02:49.300 into specific web boxes. Beat, the crucial rule 00:02:49.300 --> 00:02:51.620 is to always run the trigger once. You have to 00:02:51.620 --> 00:02:53.960 check the output tab before building further. 00:02:54.080 --> 00:02:57.250 Always. That output tab is the absolute foundation 00:02:57.250 --> 00:02:59.930 of your workflow. Google Sheets gives you a nice 00:02:59.930 --> 00:03:03.169 clean list of rows, but webhooks often hand you 00:03:03.169 --> 00:03:05.949 a highly complex mess of text. Let's go back 00:03:05.949 --> 00:03:07.990 to our lead generation example. Someone submits 00:03:07.990 --> 00:03:10.150 a form on our website. Right, and that webhook 00:03:10.150 --> 00:03:11.969 rings the doorbell with a package. You might 00:03:11.969 --> 00:03:13.990 assume you just got a simple email address, but 00:03:13.990 --> 00:03:16.129 look closely at the output tab. The system also 00:03:16.129 --> 00:03:19.129 sent their IP address and browser type. And a 00:03:19.129 --> 00:03:22.469 messy server timestamp. It's a massive pile of 00:03:22.469 --> 00:03:24.830 unneeded digital garbage. Why is skipping that 00:03:24.830 --> 00:03:27.870 first output check so dangerous? Because if you 00:03:27.870 --> 00:03:30.370 assume the data shape, you're gambling. When 00:03:30.370 --> 00:03:33.189 you guess wrong, every downstream node eventually 00:03:33.189 --> 00:03:35.430 fails completely. Without seeing the foundation, 00:03:35.710 --> 00:03:38.330 you're building the rest entirely blind. Exactly. 00:03:38.620 --> 00:03:41.099 That brings us directly to the actual worker 00:03:41.099 --> 00:03:44.740 nodes. Every single node has three specific functional 00:03:44.740 --> 00:03:47.819 parts. The left side is your incoming raw data 00:03:47.819 --> 00:03:50.870 input. beat. The middle is the node configuration 00:03:50.870 --> 00:03:53.409 and specific instructions. And the right side 00:03:53.409 --> 00:03:56.210 is your final clean data output. Raw data comes 00:03:56.210 --> 00:03:59.270 in, gets processed, and clean data leaves. Let's 00:03:59.270 --> 00:04:01.650 look closely at the Edit Fields node. It really 00:04:01.650 --> 00:04:04.830 is the absolute unsung hero of the NAN ecosystem. 00:04:05.069 --> 00:04:07.370 It doesn't send emails or update your CRM database. 00:04:07.750 --> 00:04:09.969 It just takes messy data and makes it perfectly 00:04:09.969 --> 00:04:12.449 clean. Like our lead generation webhook payload 00:04:12.449 --> 00:04:14.689 we just discussed. Perfect example. You have 00:04:14.689 --> 00:04:17.449 this massive payload with IP addresses attached. 00:04:17.529 --> 00:04:19.410 You only want the first name and the email address. 00:04:19.550 --> 00:04:21.889 Use the edit fields node to map those specific 00:04:21.889 --> 00:04:24.870 requirements. Yeah, you literally drag the email 00:04:24.870 --> 00:04:27.629 field into a new box. Then you drop the rest 00:04:27.629 --> 00:04:31.269 of the garbage completely. Beat. Doing this means 00:04:31.269 --> 00:04:34.490 understanding some very basic JSON. A simple 00:04:34.490 --> 00:04:37.670 text format mapping keys to values, like a shopping 00:04:37.670 --> 00:04:40.529 list. I love that shopping list analogy. Let's 00:04:40.529 --> 00:04:43.250 elevate it to a digital shipping manifest. The 00:04:43.250 --> 00:04:45.949 key is the written label on the cardboard box. 00:04:46.250 --> 00:04:48.339 The label might clearly say the word email. And 00:04:48.339 --> 00:04:50.500 the value is what is actually inside the box 00:04:50.500 --> 00:04:54.839 itself. The value is john at example .com. It's 00:04:54.839 --> 00:04:56.920 just a way for computers to agree on labels. 00:04:57.360 --> 00:04:59.839 Why use a node just to drop data we already have? 00:04:59.980 --> 00:05:02.040 Because clean data prevents massive headaches 00:05:02.040 --> 00:05:04.819 down the line later. Extra data makes the system 00:05:04.819 --> 00:05:07.759 extremely difficult to read. It clutters your 00:05:07.759 --> 00:05:10.639 execution logs and uses unnecessary server memory. 00:05:10.819 --> 00:05:12.800 Cleaning the data early makes the entire workflow 00:05:12.800 --> 00:05:15.540 faster and much easier to read. Exactly. And 00:05:15.540 --> 00:05:17.439 this becomes incredibly important when your work 00:05:17.389 --> 00:05:20.110 flows inevitably break down. Professional system 00:05:20.110 --> 00:05:22.509 builders actually love these unexpected system 00:05:22.509 --> 00:05:25.389 failures. Love is a very strong word for a broken 00:05:25.389 --> 00:05:27.810 system. They love them because execution logs 00:05:27.810 --> 00:05:31.009 reveal absolute truth. The execution log tells 00:05:31.009 --> 00:05:34.089 you exactly what failed instantly. Usually, the 00:05:34.089 --> 00:05:37.149 NEON software isn't actually broken itself. Instead, 00:05:37.310 --> 00:05:39.750 you almost always have a simple data mismatch 00:05:39.750 --> 00:05:42.819 problem. A node expected one specific thing, 00:05:42.860 --> 00:05:45.120 but got another thing. Like an empty email field 00:05:45.120 --> 00:05:47.740 causing a massive red air. Right. The amateur 00:05:47.740 --> 00:05:50.300 builder stares angrily at the email node. They 00:05:50.300 --> 00:05:52.660 waste hours treating the symptoms of the problem. 00:05:53.199 --> 00:05:55.379 But the execution log shows the email field was 00:05:55.379 --> 00:05:58.579 completely empty. So you start your detailed 00:05:58.579 --> 00:06:01.060 investigation back at the trigger node. You click 00:06:01.060 --> 00:06:03.579 each processing node, moving forward step -by 00:06:03.579 --> 00:06:06.620 -step. You check output A against the next input 00:06:06.620 --> 00:06:09.060 B. You find exactly where the data started looking 00:06:09.060 --> 00:06:11.420 very weird. You track the bad material, back 00:06:11.420 --> 00:06:13.699 up the supply chain. Why do we fix the sending 00:06:13.699 --> 00:06:16.399 node instead of the failing node? Because the 00:06:16.399 --> 00:06:18.779 node throwing the error did nothing wrong, technically. 00:06:18.959 --> 00:06:20.759 You have to fix the node that sent the corrupted 00:06:20.759 --> 00:06:23.079 package. You have to cure the actual disease, 00:06:23.220 --> 00:06:25.480 not just treat the symptom. Insert sponsor read 00:06:25.480 --> 00:06:31.379 here. Beat. Welcome back to our deep dive. We 00:06:31.379 --> 00:06:33.300 have talked extensively about how to build a 00:06:33.300 --> 00:06:35.740 factory. But where you actually build this factory 00:06:35.740 --> 00:06:38.120 matters tremendously. It's the most critical 00:06:38.120 --> 00:06:40.000 infrastructure decision you will make. Cloud 00:06:40.000 --> 00:06:42.480 options are easy, but they have strict execution 00:06:42.480 --> 00:06:45.680 limits. They charge a recurring fee and get expensive 00:06:45.680 --> 00:06:48.480 as you grow. Yeah, and the local laptop option 00:06:48.480 --> 00:06:51.399 stops when closed. The professional way is to 00:06:51.399 --> 00:06:54.660 use a dedicated VPS setup. Let's define a VPS 00:06:54.660 --> 00:06:57.300 right now. a private computer in the cloud that 00:06:57.300 --> 00:07:00.879 stays on 24 -7. Exactly. A standard VPS gives 00:07:00.879 --> 00:07:03.899 you absolute architectural freedom. Let's compare 00:07:03.899 --> 00:07:06.220 this directly to something like a Zapier account. 00:07:06.660 --> 00:07:09.240 Zapier charges you a restrictive fee for every 00:07:09.240 --> 00:07:12.500 single automated task. Every single action costs 00:07:12.500 --> 00:07:15.300 you a valuable workflow credit. It aggressively 00:07:15.300 --> 00:07:17.500 punishes you for being a successful business. 00:07:17.720 --> 00:07:21.329 If our system captures 10 ,000 daily leads, It 00:07:21.329 --> 00:07:23.709 drains money. Your bill will absolutely destroy 00:07:23.709 --> 00:07:26.389 your monthly operational budget. Self -hosting 00:07:26.389 --> 00:07:29.350 NAN eliminates this financial anxiety completely 00:07:29.350 --> 00:07:32.089 for serious users. You pay one predictable flat 00:07:32.089 --> 00:07:35.170 fee for the server power? Whoa. Imagine scaling 00:07:35.170 --> 00:07:38.639 to a billion queries. That kind of massive scale 00:07:38.639 --> 00:07:41.660 is hard to even grasp. How does that flat fee 00:07:41.660 --> 00:07:44.120 change the way you design automations? It changes 00:07:44.120 --> 00:07:46.540 your fundamental philosophy of technical system 00:07:46.540 --> 00:07:49.500 design. You stop trying to cram everything into 00:07:49.500 --> 00:07:52.660 fewer steps. You stop optimizing workflows just 00:07:52.660 --> 00:07:55.639 to save arbitrary task counts. Right. You build 00:07:55.639 --> 00:07:58.839 the absolute best, most robust system possible. 00:07:59.139 --> 00:08:01.779 You design for reliability instead of cheap cost 00:08:01.779 --> 00:08:03.899 -saving shortcuts. It gives you the freedom to 00:08:03.899 --> 00:08:05.959 build massive systems without fearing the money. 00:08:05.800 --> 00:08:09.019 bill. Exactly and when you build massive systems 00:08:09.019 --> 00:08:12.040 you need strict control logic. You must carefully 00:08:12.040 --> 00:08:14.579 guide your data paths through the factory. This 00:08:14.579 --> 00:08:18.040 involves three very specific types of logic routing 00:08:18.040 --> 00:08:21.279 nodes. First we have the incredibly powerful 00:08:21.279 --> 00:08:23.399 filter logic node. Think of the filter node as 00:08:23.399 --> 00:08:25.589 an aggressive security guard. It looks at your 00:08:25.589 --> 00:08:28.389 incoming data and asks for permission. Are you 00:08:28.389 --> 00:08:31.129 actually allowed to pass this specific physical 00:08:31.129 --> 00:08:34.269 point? So in our ongoing lead generation workflow 00:08:34.269 --> 00:08:37.509 example today, we only want valuable leads originating 00:08:37.509 --> 00:08:40.190 through the United States. Right. You set the 00:08:40.190 --> 00:08:42.990 strict filter to match the country field. If 00:08:42.990 --> 00:08:46.450 it's not the US, the data stops dead. It's rejected 00:08:46.450 --> 00:08:49.289 and removed from the factory line entirely. Next, 00:08:49.389 --> 00:08:51.850 we have the highly useful router control node. 00:08:52.360 --> 00:08:55.200 A router acts exactly like a fork in the road. 00:08:55.519 --> 00:08:57.779 Yeah, a fork in the road is a great analogy. 00:08:58.100 --> 00:09:00.120 When would you use a router instead of just a 00:09:00.120 --> 00:09:02.769 filter? Well, a router doesn't reject data like 00:09:02.769 --> 00:09:05.509 a filter does. It simply organizes valid data 00:09:05.509 --> 00:09:08.950 into different distinct categories. Path A sends 00:09:08.950 --> 00:09:11.710 high -value leads straight to a Slack sales alert. 00:09:11.970 --> 00:09:14.710 And Path B sends low -value leads to a general 00:09:14.710 --> 00:09:17.190 MailChimp campaign. Exactly. It routes the traffic 00:09:17.190 --> 00:09:19.330 without destroying any of the vehicles. Filters 00:09:19.330 --> 00:09:21.929 stop data completely, while routers just send 00:09:21.929 --> 00:09:24.389 it to a different destination. Precisely. And 00:09:24.389 --> 00:09:26.809 finally, we have the essential data merge node 00:09:26.809 --> 00:09:29.610 tool. You might have three different paths processing 00:09:29.610 --> 00:09:31.549 different lead information. But you eventually 00:09:31.549 --> 00:09:34.330 want it back in one single list. The merge node 00:09:34.330 --> 00:09:37.009 brings those split streams seamlessly back together. 00:09:37.350 --> 00:09:39.889 You can combine them beautifully for one final 00:09:39.889 --> 00:09:43.289 daily report. Speed. OK, let's get into the topic 00:09:43.289 --> 00:09:45.470 everyone always asks about. People want to throw 00:09:45.470 --> 00:09:47.830 artificial intelligence at every single workflow 00:09:47.830 --> 00:09:50.190 step. Right. And that is a massive architectural 00:09:50.190 --> 00:09:53.710 mistake every single time. AI should never be 00:09:53.710 --> 00:09:56.350 the core engine of your entire workflow. It's 00:09:56.350 --> 00:09:58.759 incredibly slow. and highly unpredictable by 00:09:58.759 --> 00:10:01.039 its nature. You must put it at the very end of 00:10:01.039 --> 00:10:02.960 your factory line. You only let it touch data 00:10:02.960 --> 00:10:05.320 that is already perfectly clean. So we use it 00:10:05.320 --> 00:10:07.759 as a highly specialized factory worker. Exactly. 00:10:07.940 --> 00:10:11.399 An AI node generally has four specific configurable 00:10:11.399 --> 00:10:14.299 parts. First is the underlying predictive model 00:10:14.299 --> 00:10:17.539 itself. You might use GPT -4 .0 for complex logical 00:10:17.539 --> 00:10:19.879 reasoning tasks. Or you might use Gemini for 00:10:19.879 --> 00:10:22.419 faster, cheaper, simpler processing jobs. Right. 00:10:22.820 --> 00:10:25.590 The second part of the AI node is memory. which 00:10:25.590 --> 00:10:27.830 you should usually safely disable for single 00:10:27.830 --> 00:10:30.529 data items. Third, you have external tools like 00:10:30.529 --> 00:10:32.509 giving it web search capabilities. But the fourth 00:10:32.509 --> 00:10:35.750 part is the absolute most critical piece, the 00:10:35.750 --> 00:10:38.289 output parser configuration. Why is the output 00:10:38.289 --> 00:10:40.629 parser the most important piece of the AI node? 00:10:40.970 --> 00:10:44.029 Because AI models naturally want to write chatty 00:10:44.029 --> 00:10:46.309 conversational paragraphs. They want to say, 00:10:46.690 --> 00:10:49.289 here is the data you requested today. That extra 00:10:49.289 --> 00:10:51.450 conversational text breaks automated workflows 00:10:51.450 --> 00:10:54.980 instantly. The output parser forces the AI to 00:10:54.980 --> 00:10:58.120 return a strict, structured list. It forces the 00:10:58.120 --> 00:11:00.299 AI to give you clean data for the next step. 00:11:00.519 --> 00:11:02.940 Exactly. You must give it strict, systematic 00:11:02.940 --> 00:11:06.320 points to check. Always. I tell the AI exactly 00:11:06.320 --> 00:11:08.980 what to find systematically. I show it exactly 00:11:08.980 --> 00:11:11.620 what perfectly parsed JSON looks like. Beat. 00:11:12.320 --> 00:11:14.980 I still wrestle with prompt drift myself. It 00:11:14.980 --> 00:11:17.419 feels like hitting a moving digital target sometimes. 00:11:17.620 --> 00:11:19.580 It certainly takes real practice to get these 00:11:19.580 --> 00:11:21.970 instructions perfect. But you also have to consider 00:11:21.970 --> 00:11:25.090 data privacy with AI models. If you use big companies, 00:11:25.149 --> 00:11:27.470 you send your valuable leads away. But you can 00:11:27.470 --> 00:11:30.330 use a powerful private tool called Olama instead. 00:11:30.509 --> 00:11:33.029 A tool to run AI models privately on your own 00:11:33.029 --> 00:11:36.070 server. Yes. This means the local AI reads your 00:11:36.070 --> 00:11:38.730 data right there. It analyzes the specific lead 00:11:38.730 --> 00:11:41.649 and drafts a custom welcome email. This completely 00:11:41.649 --> 00:11:44.669 avoids paying for every single external AI request. 00:11:44.809 --> 00:11:47.009 And it gives you complex answers without sensitive 00:11:47.009 --> 00:11:50.230 data ever leaving. It's beat. We have covered 00:11:50.230 --> 00:11:53.230 a truly massive amount of ground today. Mastering 00:11:53.230 --> 00:11:55.809 ANN isn't about memorizing complex interface 00:11:55.809 --> 00:11:57.789 buttons. You definitely don't need to know how 00:11:57.789 --> 00:12:00.409 to write code. If you know basic Excel logic, 00:12:00.690 --> 00:12:03.250 you're fine. It's fundamentally a massive mental 00:12:03.250 --> 00:12:06.250 paradigm shift entirely. You are treating data 00:12:06.250 --> 00:12:09.210 like a real physical object moving constantly. 00:12:09.370 --> 00:12:11.409 It moves smoothly through a highly efficient 00:12:11.409 --> 00:12:14.179 digital factory. You keep your individual worker 00:12:14.179 --> 00:12:17.159 nodes very small, always. You rely entirely on 00:12:17.159 --> 00:12:20.059 detail execution logs for ultimate truth. You 00:12:20.059 --> 00:12:23.679 use advanced AI only as a specialized final worker. 00:12:23.820 --> 00:12:26.259 Yeah, and you run everything securely on your 00:12:26.259 --> 00:12:28.379 own private server. You are no longer just a 00:12:28.379 --> 00:12:31.259 basic automation user today. You are a professional 00:12:31.259 --> 00:12:33.500 digital data architect right now. You are building 00:12:33.500 --> 00:12:35.700 machines that work for you day and night. Now 00:12:35.700 --> 00:12:37.299 that you understand how to control the exact 00:12:37.299 --> 00:12:39.909 shape of your data through time. What massive 00:12:39.909 --> 00:12:42.250 complex process in your life could you hand over 00:12:42.250 --> 00:12:43.110 to a machine tomorrow?