WEBVTT 00:00:00.000 --> 00:00:01.500 You've probably used ChatGPT, right? But what 00:00:01.500 --> 00:00:04.080 if I told you most of us are barely scratching 00:00:04.080 --> 00:00:06.120 the surface, like using a race car just to pop 00:00:06.120 --> 00:00:09.320 down to the shops. We're leaving like 99 % of 00:00:09.320 --> 00:00:12.859 its actual power just sitting there. Beat. Today, 00:00:12.960 --> 00:00:14.939 we're going to dive deep into how businesses 00:00:14.939 --> 00:00:18.019 are actually building, you know, automated workforces 00:00:18.019 --> 00:00:21.250 using AI. That's exactly it. I mean, imagine 00:00:21.250 --> 00:00:23.609 turning an 80 -hour work week into something 00:00:23.609 --> 00:00:26.370 way more efficient, more profitable, too. All 00:00:26.370 --> 00:00:29.329 just by connecting chat GPTs, let's call it the 00:00:29.329 --> 00:00:32.810 brain, to a powerful automation engine like N8n. 00:00:32.950 --> 00:00:36.289 Think of N8n as the hands and feet for that AI 00:00:36.289 --> 00:00:38.630 brain and lets it actually do things in the real 00:00:38.630 --> 00:00:40.369 world, connect to other apps. And that's really 00:00:40.369 --> 00:00:41.990 the core of our deep dive today. We're going 00:00:41.990 --> 00:00:44.310 way beyond just the chat window. We're talking 00:00:44.310 --> 00:00:47.149 about commanding an autonomous workforce. We'll 00:00:47.149 --> 00:00:49.219 unpack, I think it's 10... Actionable blueprints. 00:00:49.259 --> 00:00:52.259 Yeah, 10. Everything from creating AI art directors, 00:00:52.359 --> 00:00:54.979 scoring leads, even giving your AI voice. It's 00:00:54.979 --> 00:00:56.899 really all about building a system that works 00:00:56.899 --> 00:00:58.939 for you, not the other way around. OK, so kicking 00:00:58.939 --> 00:01:01.659 us off, let's talk visuals, visual content. If 00:01:01.659 --> 00:01:03.399 you're running an e -commerce store or managing 00:01:03.399 --> 00:01:06.189 social media, you know. generating professional 00:01:06.189 --> 00:01:08.370 product shots, that's a huge cost. It can be 00:01:08.370 --> 00:01:11.250 a massive bottleneck. But with AI now, you can 00:01:11.250 --> 00:01:13.590 generate these really stunning product photos 00:01:13.590 --> 00:01:16.709 with, well, pretty simple text prompts. Like, 00:01:16.750 --> 00:01:19.590 imagine asking for a lush soap product shot with 00:01:19.590 --> 00:01:21.909 perfect, you know, professional lighting. Or 00:01:21.909 --> 00:01:24.329 maybe a body wash beauty shot that looks totally 00:01:24.329 --> 00:01:26.310 magazine quality, water droplets and all. It 00:01:26.310 --> 00:01:28.090 feels kind of like basic magic. It really does 00:01:28.090 --> 00:01:30.349 feel like magic sometimes. But the real magic, 00:01:30.409 --> 00:01:32.329 like the next level stuff, that happens when 00:01:32.329 --> 00:01:34.489 you build what the guide we looked at calls the 00:01:34.569 --> 00:01:37.849 AI art director. This is where it gets cool because 00:01:37.849 --> 00:01:40.650 you're chaining AI abilities together. So the 00:01:40.650 --> 00:01:43.730 output from one AI step feeds into the next one. 00:01:43.829 --> 00:01:47.290 The workflow basically starts by sending separate 00:01:47.290 --> 00:01:50.530 prompts to chat GPT. Say, generate two individual 00:01:50.530 --> 00:01:53.090 product images. Then it automatically takes... 00:01:53.319 --> 00:01:55.500 Those image files, right? And feeds them into 00:01:55.500 --> 00:01:57.879 a second chat GPT node. And this time you give 00:01:57.879 --> 00:02:00.200 it a compositing prompt. Something like, create 00:02:00.200 --> 00:02:02.599 a luxury spa scene featuring these two products 00:02:02.599 --> 00:02:05.260 placed elegantly on a marble countertop. Maybe 00:02:05.260 --> 00:02:07.579 add some premium typography. Okay, chaining AI 00:02:07.579 --> 00:02:10.120 like that for custom composites. That's genuinely 00:02:10.120 --> 00:02:11.960 exciting. So what's the biggest shift you think 00:02:11.960 --> 00:02:14.539 this brings to how marketing teams actually work? 00:02:14.620 --> 00:02:16.840 I mean, beyond just saving money on photo shoots, 00:02:17.000 --> 00:02:20.919 what new sort of strategic things does it unlock? 00:02:21.479 --> 00:02:23.340 That's a great question. I think the core of 00:02:23.340 --> 00:02:27.240 it is it fundamentally changes the pace of creative 00:02:27.240 --> 00:02:29.340 iteration. Right. You're no longer waiting weeks 00:02:29.340 --> 00:02:31.280 for a professional photo shoot. You can prototype 00:02:31.280 --> 00:02:33.599 ad campaigns, generate tons of visual options 00:02:33.599 --> 00:02:36.759 like daily, get immediate feedback. It just collapses 00:02:36.759 --> 00:02:38.740 that whole feedback loop. So you can do rapid 00:02:38.740 --> 00:02:41.340 A -B testing of visuals like never before. That's 00:02:41.340 --> 00:02:43.539 a massive competitive edge. And then there's 00:02:43.539 --> 00:02:45.379 the ultimate upgrade they mentioned, the brand 00:02:45.379 --> 00:02:48.360 consistency engine. This is clever. You basically 00:02:48.360 --> 00:02:50.159 feed it a text file with your brand specific 00:02:50.159 --> 00:02:52.500 details, color codes like hex codes, hashtag 00:02:52.500 --> 00:02:55.759 003366 or whatever, font names, mood descriptions. 00:02:55.900 --> 00:02:57.479 You put that in your prompt and it makes sure 00:02:57.479 --> 00:02:59.439 every single AI generated image is perfectly 00:02:59.439 --> 00:03:02.159 on brand. Every time. And it's not just making 00:03:02.159 --> 00:03:04.439 the images, is it? It's also about giving the 00:03:04.439 --> 00:03:08.189 AI a kind of creative eye. to, well, to understand 00:03:08.189 --> 00:03:10.990 them. Like, imagine an AI social media manager. 00:03:11.250 --> 00:03:14.169 It analyzes an image, the colors, the mood, the 00:03:14.169 --> 00:03:16.569 composition, and then it generates these perfectly 00:03:16.569 --> 00:03:18.969 crafted captions, you know, complete with relevant 00:03:18.969 --> 00:03:22.330 emojis, good hashtags. So for that skincare duo 00:03:22.330 --> 00:03:25.110 example, it might spit out. Treat yourself to 00:03:25.110 --> 00:03:29.199 the spa day you deserve. This soothing body care 00:03:29.199 --> 00:03:32.099 set is your new go -to. Hashtag self -care essentials. 00:03:32.500 --> 00:03:34.199 Hashtag bath time bliss. You know, that kind 00:03:34.199 --> 00:03:36.300 of thing. Exactly. And this even extends to something 00:03:36.300 --> 00:03:39.400 they call automated competitor ad analysis. You 00:03:39.400 --> 00:03:42.139 can actually build a workflow that, say, scrapes 00:03:42.139 --> 00:03:44.419 competitor ad images from places like the Facebook 00:03:44.419 --> 00:03:47.419 ad library. Then it feeds those images to ChatGPT 00:03:47.419 --> 00:03:49.319 for analysis. So you get these recurring, like, 00:03:49.379 --> 00:03:52.039 automated competitor intel reports on visual 00:03:52.039 --> 00:03:54.180 trends, spotting patterns you'd probably miss 00:03:54.180 --> 00:03:56.810 otherwise. OK, so beyond social media, how does 00:03:56.810 --> 00:03:59.150 this image analysis capability start to impact 00:03:59.150 --> 00:04:01.629 like core business functions? Are there unexpected 00:04:01.629 --> 00:04:03.849 places that could pop up? Oh, yeah, definitely. 00:04:03.949 --> 00:04:06.050 Think about automated product descriptions based 00:04:06.050 --> 00:04:10.250 on the image or. generating really precise SEO 00:04:10.250 --> 00:04:13.110 alt text automatically for better search discoverability. 00:04:13.349 --> 00:04:15.569 You can even use it for a new layer of automated 00:04:15.569 --> 00:04:17.550 quality control for visual assets before they 00:04:17.550 --> 00:04:20.569 go live. Okay, let's switch gears a bit to something 00:04:20.569 --> 00:04:23.050 really essential. Pulling information out of 00:04:23.050 --> 00:04:25.449 text. We all drown in unstructured data, don't 00:04:25.449 --> 00:04:27.910 we? Emails from potential clients, PDFs from 00:04:27.910 --> 00:04:30.449 partners, messy CSV files, maybe transcribed 00:04:30.449 --> 00:04:33.689 voicemails. It's just soul -crushingly boring 00:04:33.689 --> 00:04:37.250 work, frankly. Imagine having to manually process, 00:04:37.329 --> 00:04:40.670 say, 30 new leads they're scattered across different 00:04:40.670 --> 00:04:43.490 emails maybe some web form submissions key details 00:04:43.490 --> 00:04:45.370 are buried in paragraphs that's what the guide 00:04:45.370 --> 00:04:48.029 calls the soul crushing problem and it's basically 00:04:48.029 --> 00:04:50.610 a process designed to create data entry errors 00:04:50.610 --> 00:04:52.769 right absolutely and the guide offers some pretty 00:04:52.769 --> 00:04:56.149 powerful solutions here so for really basic stuff 00:04:56.149 --> 00:04:58.949 nea den actually has a built -in node called 00:04:58.949 --> 00:05:01.939 information extractor Simple solution, good for 00:05:01.939 --> 00:05:04.500 quick tasks. But, you know, it does have limits 00:05:04.500 --> 00:05:06.660 when you get into complex, messy, real -world 00:05:06.660 --> 00:05:08.839 data. So when you need to handle that variable 00:05:08.839 --> 00:05:11.100 stuff, the heavy -duty solution is using the 00:05:11.100 --> 00:05:14.759 main OpenAI node in 8 .8n. And the key is this 00:05:14.759 --> 00:05:17.259 powerful three -message system they describe. 00:05:17.560 --> 00:05:20.079 You have a system message. that gives the AI 00:05:20.079 --> 00:05:22.519 its permanent rules. Like you're a smart bot 00:05:22.519 --> 00:05:25.720 that extracts lead details from emails to the 00:05:25.720 --> 00:05:27.480 user message. That's where you dump the actual 00:05:27.480 --> 00:05:29.560 messy data, the email text, the transcript, whatever, 00:05:29.720 --> 00:05:31.860 and then the assistant message. This is the secret 00:05:31.860 --> 00:05:34.459 weapon, really. You give the AI the exact empty 00:05:34.459 --> 00:05:36.699 JSON format you want it to fill out, like a template. 00:05:36.939 --> 00:05:38.720 You know, I still wrestle with prompt drift myself 00:05:38.720 --> 00:05:40.720 sometimes, especially when I'm trying to get 00:05:40.720 --> 00:05:43.240 AI to output things in a really precise format 00:05:43.240 --> 00:05:46.040 like JSON. This three -message system, it sounds 00:05:46.040 --> 00:05:47.819 like you're properly engineering the output. 00:05:48.560 --> 00:05:50.500 How much of a game changer is that for consistency? 00:05:50.839 --> 00:05:53.019 It truly is. It makes a huge difference. And 00:05:53.019 --> 00:05:54.800 a little pro tip from the guide, which I love, 00:05:54.920 --> 00:05:57.839 you can actually ask another instance of ChatGPT 00:05:57.839 --> 00:05:59.879 to help you build that perfect JSON template. 00:06:00.339 --> 00:06:04.019 JSON is just a structured way to get data back, 00:06:04.079 --> 00:06:05.620 kind of like a pre -formatted spreadsheet row. 00:06:05.949 --> 00:06:08.250 Very useful. And taking this even further is 00:06:08.250 --> 00:06:09.990 the advanced application they mentioned, the 00:06:09.990 --> 00:06:12.949 universal inbox processor. So imagine every piece 00:06:12.949 --> 00:06:16.009 of unstructured info, new lead emails, form submissions, 00:06:16.230 --> 00:06:18.170 those transcribed voicemails, it all gets routed 00:06:18.170 --> 00:06:21.810 into one single workflow. First, a classifier 00:06:21.810 --> 00:06:24.790 AI figures out what it is. Is it a lead, a support 00:06:24.790 --> 00:06:27.810 ticket, an invoice? Then it sends it to a custom 00:06:27.810 --> 00:06:30.129 built information extraction prompt specifically 00:06:30.129 --> 00:06:32.389 designed for that data type. You end up with 00:06:32.389 --> 00:06:35.939 pristine structured data every time. Wow. It's 00:06:35.939 --> 00:06:37.699 kind of like Tank from The Matrix, isn't it? 00:06:37.699 --> 00:06:39.939 Looking at all that raw, cascading green code 00:06:39.939 --> 00:06:42.519 and just instantly seeing blonde brunette redheads 00:06:42.519 --> 00:06:44.800 pulling out the important bits. Just pulling 00:06:44.800 --> 00:06:47.139 out the important bits. Exactly. And the real 00:06:47.139 --> 00:06:49.149 world impact. It's profound. You're looking at, 00:06:49.170 --> 00:06:51.470 well, massive ROI potential. You practically 00:06:51.470 --> 00:06:54.949 eliminate manual data entry errors and you accelerate 00:06:54.949 --> 00:06:57.449 lead processing so dramatically. I mean, your 00:06:57.449 --> 00:06:59.209 sales team ends up spending way more time actually 00:06:59.209 --> 00:07:01.810 closing deals, less time just sifting through 00:07:01.810 --> 00:07:04.230 stuff, which actually segues really nicely into 00:07:04.230 --> 00:07:06.910 the next blueprint, the digital Zen master, bringing 00:07:06.910 --> 00:07:10.350 order to chaos with AI classification. This system 00:07:10.350 --> 00:07:13.129 is basically an innate and workflow that triggers 00:07:13.129 --> 00:07:15.370 every time you get a new email. It sends the 00:07:15.370 --> 00:07:17.810 email content off to ChatGPT to classify it into 00:07:17.810 --> 00:07:20.430 categories you defined beforehand, sales, support, 00:07:20.689 --> 00:07:22.769 admin, marketing, calendar, whatever makes sense 00:07:22.769 --> 00:07:24.870 for you. And then based on that classification, 00:07:25.149 --> 00:07:27.269 it automatically applies labels in your inbox 00:07:27.269 --> 00:07:29.350 or even moves the email to a specific folder. 00:07:29.790 --> 00:07:32.670 Tidy inbox. And the next level beyond just sorting 00:07:32.670 --> 00:07:35.529 is the action -oriented inbox. This is where 00:07:35.529 --> 00:07:37.110 it gets really powerful. The system doesn't just 00:07:37.110 --> 00:07:38.790 label your emails anymore. It actually starts 00:07:38.790 --> 00:07:41.350 to work for you. So an invoice email comes in. 00:07:41.430 --> 00:07:43.189 Maybe it triggers a draft payment in your accounting 00:07:43.189 --> 00:07:46.329 software. High urgency support ticket? It automatically 00:07:46.329 --> 00:07:48.790 creates a help desk ticket and sends a slack 00:07:48.790 --> 00:07:52.050 alert to the support team. A new sales lead email. 00:07:52.490 --> 00:07:54.810 Bam. It automatically triggers that lead scoring 00:07:54.810 --> 00:07:56.910 workflow we're going to talk about next. OK, 00:07:56.949 --> 00:07:59.750 what's the biggest like personal tangible benefit 00:07:59.750 --> 00:08:03.310 of having an action oriented inbox, especially, 00:08:03.310 --> 00:08:06.350 say, on a Friday afternoon when your brings already 00:08:06.350 --> 00:08:08.050 kind of checked out for the weekend? Honestly, 00:08:08.129 --> 00:08:09.990 it just transformed your inbox. It goes from 00:08:09.990 --> 00:08:12.689 being this this passive list of messages demanding 00:08:12.689 --> 00:08:15.750 your attention into an active, intelligent system 00:08:15.750 --> 00:08:17.750 that's already working for you. Your mental load 00:08:17.750 --> 00:08:20.370 just drops significantly. It's a huge relief. 00:08:20.589 --> 00:08:23.649 Right. That definitely brings us neatly to automatically 00:08:23.649 --> 00:08:26.689 scoring leads because in any business, you know, 00:08:26.709 --> 00:08:29.089 not all leads are created equal. And wasting 00:08:29.089 --> 00:08:31.129 hours chasing prospects who are never going to 00:08:31.129 --> 00:08:33.269 buy anyway, that's just the fastest way to burn 00:08:33.269 --> 00:08:35.309 out a sales team, isn't it? Oh, totally. This 00:08:35.309 --> 00:08:38.470 automation is designed to be like the sorting 00:08:38.470 --> 00:08:40.370 hat from Harry Potter, but for your business 00:08:40.370 --> 00:08:44.309 leads. It's an AI -powered system, right? It 00:08:44.309 --> 00:08:47.269 instantly analyzes new leads against your ideal 00:08:47.269 --> 00:08:49.899 customer profile, the criteria you set. And then 00:08:49.899 --> 00:08:52.340 it places them into the right house. Hot lead, 00:08:52.559 --> 00:08:56.340 nurture, or maybe just not a fit. So the first 00:08:56.340 --> 00:08:58.820 step is building your scoring rubric. That just 00:08:58.820 --> 00:09:00.740 means defining what makes an ideal customer for 00:09:00.740 --> 00:09:03.080 you. Points for company size, maybe budget range, 00:09:03.340 --> 00:09:05.340 industry fit, whether they mention an urgent 00:09:05.340 --> 00:09:07.740 need. Then you craft the expert analyst prompt. 00:09:08.019 --> 00:09:10.600 This basically gives the AI a persona, like a 00:09:10.600 --> 00:09:12.500 seasoned sales analyst, so it scores consistently. 00:09:12.940 --> 00:09:14.879 And from that score, any end uses what they call 00:09:14.879 --> 00:09:17.440 an automated routing system. triggers different 00:09:17.440 --> 00:09:19.840 actions automatically based on the score. So 00:09:19.840 --> 00:09:22.059 a hot lead might get a high priority task created 00:09:22.059 --> 00:09:24.399 in your CRM immediately. A nurture lead might 00:09:24.399 --> 00:09:26.840 get added to a specific email sequence, not a 00:09:26.840 --> 00:09:28.799 fit, might even get a polite automated decline 00:09:28.799 --> 00:09:31.940 email. And for the pro -level upgrade, dynamic, 00:09:32.100 --> 00:09:34.039 real -time scoring. This is pretty neat. Your 00:09:34.039 --> 00:09:36.139 A &A on workflow could take a lead's email or 00:09:36.139 --> 00:09:38.700 company name, right? And use a Dayton enrichment 00:09:38.700 --> 00:09:41.240 tool, something like Clearbit, to pull in real 00:09:41.240 --> 00:09:43.100 -time info. Things like their latest funding 00:09:43.100 --> 00:09:46.440 status, current employee count. tech stack, maybe. 00:09:46.559 --> 00:09:49.200 And then it adds that fresh data to the information 00:09:49.200 --> 00:09:51.639 the AI uses for scoring, makes the qualification 00:09:51.639 --> 00:09:53.639 much, much more accurate. That sounds incredibly 00:09:53.639 --> 00:09:55.580 efficient. Yeah, like a sorting hat for leads, 00:09:55.720 --> 00:09:58.759 definitely. But are there any risks? Like, are 00:09:58.759 --> 00:10:01.139 you worried about over automating this, maybe 00:10:01.139 --> 00:10:03.940 losing that subtle human intuition that sometimes 00:10:03.940 --> 00:10:06.340 spots an unconventional lead who turns out to 00:10:06.340 --> 00:10:08.559 be really valuable? That's a really valid concern, 00:10:08.659 --> 00:10:11.879 actually. And finding the right balance is definitely 00:10:11.879 --> 00:10:14.809 key. But the strategic shift here isn't really 00:10:14.809 --> 00:10:17.950 about replacing intuition. It's more about amplifying 00:10:17.950 --> 00:10:20.730 it by automating all the grunt work, you know, 00:10:20.730 --> 00:10:23.669 sifting through maybe hundreds of clearly unqualified 00:10:23.669 --> 00:10:26.950 leads. Your human sales team can focus their 00:10:26.950 --> 00:10:29.830 expertise, their intuition entirely on those 00:10:29.830 --> 00:10:32.970 high value prospects. It boosts morale because 00:10:32.970 --> 00:10:35.389 they're not wasting time and it should significantly 00:10:35.389 --> 00:10:37.470 increase conversion rates, too. OK, that makes 00:10:37.470 --> 00:10:40.509 sense. Amplifying, not replacing. So let's unpack 00:10:40.509 --> 00:10:43.049 this. Once you have these qualified leads or 00:10:43.049 --> 00:10:45.169 existing customers, how do you make sure your 00:10:45.169 --> 00:10:47.429 AI actually knows your business inside and out 00:10:47.429 --> 00:10:48.909 so it can answer their questions accurately, 00:10:49.070 --> 00:10:51.309 not just give generic chat GPT answers? That's 00:10:51.309 --> 00:10:53.690 where these RHG systems come in, right? Exactly. 00:10:53.750 --> 00:10:55.970 Building a RAG system that stands for Retrieval 00:10:55.970 --> 00:10:58.789 Augmented Generation. It's like giving your AI 00:10:58.789 --> 00:11:02.490 a Jedi holocron filled with your company's knowledge. 00:11:02.909 --> 00:11:05.049 Basically, it's an AI assistant that has already 00:11:05.049 --> 00:11:07.669 read, understood, and essentially memorized every 00:11:07.669 --> 00:11:10.470 important document your business has. And this 00:11:10.470 --> 00:11:12.789 is crucial because, unlike a general AI that 00:11:12.789 --> 00:11:15.309 might, you know, hallucinate or just guess answers 00:11:15.309 --> 00:11:18.230 based on its broad training data, a RAGI system 00:11:18.230 --> 00:11:20.470 actually retrieves specific, accurate information 00:11:20.470 --> 00:11:22.690 directly from your documents before it generates 00:11:22.690 --> 00:11:25.610 a response. That's the key difference. It ensures 00:11:25.610 --> 00:11:28.529 you get 100 % factual, on -brand answers based 00:11:28.529 --> 00:11:30.850 on your materials. And the three -step blueprint 00:11:30.850 --> 00:11:32.590 they outline is pretty straightforward, actually. 00:11:32.710 --> 00:11:34.990 First, you create your assistant inside the OpenAI 00:11:34.990 --> 00:11:37.990 platform itself. Second, you upload your knowledge 00:11:37.990 --> 00:11:41.500 base. FAQs, product manuals, company policies, 00:11:41.700 --> 00:11:43.320 internal training materials, whatever's relevant. 00:11:43.500 --> 00:11:45.440 Third, you connect that assistant to your innate 00:11:45.440 --> 00:11:47.700 end workflow using the message and assistant 00:11:47.700 --> 00:11:50.299 note. They give a good real -world example. Imagine 00:11:50.299 --> 00:11:52.659 a customer support assistant for, say, a wedding 00:11:52.659 --> 00:11:55.139 photography business. Its knowledge base includes 00:11:55.139 --> 00:11:57.340 a detailed pricing PDF. So when a customer asks, 00:11:57.460 --> 00:11:59.179 hey, what's your pricing for a full -day package? 00:11:59.950 --> 00:12:03.470 Chef, the RE -powered response is pulled directly 00:12:03.470 --> 00:12:06.570 from that specific PDF, ensuring the answer is 00:12:06.570 --> 00:12:09.960 accurate every single time. No guessing. And 00:12:09.960 --> 00:12:11.700 for the advanced application, you can get really 00:12:11.700 --> 00:12:14.379 sophisticated and build a multi -brain R &D system. 00:12:14.659 --> 00:12:16.799 So you might create multiple specialized assistants, 00:12:17.179 --> 00:12:19.720 maybe a sales brain trained only on sales scripts 00:12:19.720 --> 00:12:22.460 and competitor info, a support brain focused 00:12:22.460 --> 00:12:25.080 on technical docs, an HR brain that knows the 00:12:25.080 --> 00:12:27.960 employee handbook cold. Then use another class 00:12:27.960 --> 00:12:30.080 or AI up front to figure out the user's query 00:12:30.080 --> 00:12:32.139 type and route it to the correct brain for the 00:12:32.139 --> 00:12:34.330 most relevant answer. Okay, so drilling down 00:12:34.330 --> 00:12:36.649 on that, what's the core problem our gag system 00:12:36.649 --> 00:12:38.610 ultimately solves for businesses, especially 00:12:38.610 --> 00:12:41.169 in customer support, that just using a generic 00:12:41.169 --> 00:12:43.690 AI model can't really fix? Yeah, fundamentally 00:12:43.690 --> 00:12:46.629 it prevents AI hallucination, which is a huge 00:12:46.629 --> 00:12:49.129 deal for businesses. It allows the AI to provide 00:12:49.129 --> 00:12:51.590 consistently accurate, company -specific answers 00:12:51.590 --> 00:12:54.090 by grounding every response directly in your 00:12:54.090 --> 00:12:57.549 own validated internal knowledge bases. Trustworthy 00:12:57.549 --> 00:13:00.190 answers. Midroll sponsor. Read placeholder for 00:13:00.190 --> 00:13:03.539 provided script. Okay, we're back. So we've talked 00:13:03.539 --> 00:13:05.480 internal knowledge with Argi. Now let's look 00:13:05.480 --> 00:13:08.200 outward at external research. That manual grind, 00:13:08.340 --> 00:13:10.840 right? Opening dozens of browser tabs, reading 00:13:10.840 --> 00:13:13.659 through walls of text, manually copying and pasting 00:13:13.659 --> 00:13:16.360 key points for sales prep or competitor analysis. 00:13:16.539 --> 00:13:19.320 It's just such an incredible time sink. Oh, absolutely. 00:13:19.600 --> 00:13:22.019 This next automation basically transforms your 00:13:22.019 --> 00:13:25.399 AI into a tireless automated research assistant. 00:13:27.120 --> 00:13:30.960 Now, the guide mentions the old school way, a 00:13:30.960 --> 00:13:34.100 two -step method using N8N's HTTP request node 00:13:34.100 --> 00:13:37.200 to grab raw website HTML, then sending that to 00:13:37.200 --> 00:13:39.799 ChatGPT for analysis. But honestly, that often 00:13:39.799 --> 00:13:41.659 struggled with modern websites, you know, the 00:13:41.659 --> 00:13:43.740 ones heavy on JavaScript that build the page 00:13:43.740 --> 00:13:46.179 dynamically. The much more elegant approach now 00:13:46.179 --> 00:13:48.700 is the one -step solution, what the guide kind 00:13:48.700 --> 00:13:51.200 of playfully calls the perplexity cheat code. 00:13:51.629 --> 00:13:53.889 You use NNN's built -in Perplexity integration. 00:13:54.490 --> 00:13:56.769 Perplexity's AI models, like their Sonar series, 00:13:56.970 --> 00:13:59.129 are specifically designed to browse the web and 00:13:59.129 --> 00:14:01.309 analyze content effectively. So your prompt becomes 00:14:01.309 --> 00:14:03.730 really simple. Something like, act as a world 00:14:03.730 --> 00:14:05.289 -class business analyst, visit the following 00:14:05.289 --> 00:14:07.470 website URL, and provide a concise summary of 00:14:07.470 --> 00:14:09.350 their main value proposition and target audience. 00:14:09.909 --> 00:14:11.830 Perplexity handles all the browsing and extraction 00:14:11.830 --> 00:14:13.850 heavy lifting behind the scenes super clean. 00:14:13.990 --> 00:14:15.929 And the ultimate enhancement they suggest here 00:14:15.929 --> 00:14:18.379 is pretty cool, the competitor watchtower. You 00:14:18.379 --> 00:14:20.179 can build an NEN workflow that runs on a schedule, 00:14:20.240 --> 00:14:22.500 say, daily. It scrapes the homepages of your 00:14:22.500 --> 00:14:25.039 main competitors. It compares the text it finds 00:14:25.039 --> 00:14:27.559 today to the version it saved yesterday. And 00:14:27.559 --> 00:14:29.799 if it detects significant changes, it sends both 00:14:29.799 --> 00:14:32.500 the old text and the new text to ChatGPT with 00:14:32.500 --> 00:14:35.600 a prompt like, analyze the before and after versions 00:14:35.600 --> 00:14:38.299 of this competitor's homepage copy, summarize 00:14:38.299 --> 00:14:40.879 the strategic importance of the changes. And 00:14:40.879 --> 00:14:43.100 then, boom, it sends you a Slack or Discord alert 00:14:43.100 --> 00:14:45.419 with that analysis. Critical strategic insights 00:14:45.419 --> 00:14:48.820 delivered automatically. Whoa. Hang on. Imagine 00:14:48.820 --> 00:14:51.419 scaling that, processing hundreds of competitor 00:14:51.419 --> 00:14:53.899 websites every day, getting instant strategic 00:14:53.899 --> 00:14:57.019 analysis pushed to you. That's kind of like having 00:14:57.019 --> 00:14:59.019 a palantir from Lord of the Rings, isn't it? 00:14:59.019 --> 00:15:01.340 That seeing stone that's always vigilant, always 00:15:01.340 --> 00:15:04.120 watching your rivals. So what's the biggest shift 00:15:04.120 --> 00:15:07.820 this brings to how companies do competitive intelligence? 00:15:08.320 --> 00:15:10.539 It completely changes the game from being reactive 00:15:10.539 --> 00:15:13.000 to proactive. Right. You gain this continuous. 00:15:13.799 --> 00:15:15.940 automated surveillance of your competitive landscape. 00:15:16.179 --> 00:15:18.720 You spot competitor strategic shifts, maybe new 00:15:18.720 --> 00:15:21.100 product launches, major messaging changes, pricing 00:15:21.100 --> 00:15:23.460 adjustments almost immediately, not weeks later. 00:15:23.620 --> 00:15:26.059 That gives you a crucial timing advantage. Okay, 00:15:26.100 --> 00:15:28.759 so we've seen AI working inside these NNN workflows, 00:15:29.019 --> 00:15:31.519 handling images, text, voice stuff, research. 00:15:31.879 --> 00:15:34.039 But what if the interface we're all familiar 00:15:34.039 --> 00:15:36.320 with now, that conversational chat GPT window 00:15:36.320 --> 00:15:38.639 itself, what if that could actually command your 00:15:38.639 --> 00:15:41.620 entire NNN automation system? Now you're talking 00:15:41.620 --> 00:15:44.399 about what the guide calls the nuke button, connecting 00:15:44.399 --> 00:15:48.279 the chat GPT interface directly to NAG. And they're 00:15:48.279 --> 00:15:50.200 not wrong to call it maybe the holy grail of 00:15:50.200 --> 00:15:53.419 integration right now. because it combines ChatGPT's 00:15:53.419 --> 00:15:56.919 brilliant conversational brain with NAD's complete 00:15:56.919 --> 00:15:58.960 set of hands and feet, its ability to connect 00:15:58.960 --> 00:16:00.960 to and control thousands of other applications. 00:16:01.419 --> 00:16:04.559 The setup, they call it the JARVIS Blueprint, 00:16:04.740 --> 00:16:08.299 involves creating a custom GPT pursuit. You need 00:16:08.299 --> 00:16:10.120 ChatGPT Plus for this, and you configure that 00:16:10.120 --> 00:16:13.200 custom GPT to talk to your N8n instance via something 00:16:13.200 --> 00:16:15.779 called a webhook. You can think of a webhook 00:16:15.779 --> 00:16:18.080 as like a secret direct phone number that lets 00:16:18.080 --> 00:16:21.440 ChatGPT call N8n and tell it to start a specific 00:16:21.440 --> 00:16:24.990 workflow. The power this unlocks is, well, it's 00:16:24.990 --> 00:16:27.149 almost unlimited. Suddenly, you can do full calendar 00:16:27.149 --> 00:16:29.409 management, deep operations within your CRM, 00:16:29.450 --> 00:16:31.750 complex database queries, social media publishing. 00:16:32.049 --> 00:16:34.429 Basically, anything an 8N can do with its 1 ,000 00:16:34.429 --> 00:16:36.389 -plus integrations, you can potentially trigger 00:16:36.389 --> 00:16:38.129 from a natural language chat with your custom 00:16:38.129 --> 00:16:40.710 GPT. They give a great real -world example for 00:16:40.710 --> 00:16:42.669 this. Imagine typing a single conversational 00:16:42.669 --> 00:16:45.350 command into your custom GPT. Okay, John Smith 00:16:45.350 --> 00:16:48.750 from Acme Inc. Just agreed to a demo. Can you 00:16:48.750 --> 00:16:50.669 schedule a 30 -minute call for next Tuesday at 00:16:50.669 --> 00:16:52.960 2 p .m. Eastern? Also create a new deal for him 00:16:52.960 --> 00:16:55.860 in our HubSpot CRM. Set the value at $5 ,000. 00:16:56.240 --> 00:16:58.240 In the draft, a personalized confirmation email. 00:16:58.480 --> 00:17:02.259 Make sure to include a calendar invite. PM. That 00:17:02.259 --> 00:17:04.279 single sentence isn't just a simple query anymore. 00:17:04.480 --> 00:17:06.240 It triggers this whole symphony of automation 00:17:06.240 --> 00:17:08.900 behind the scenes. NA then checks your calendar 00:17:08.900 --> 00:17:11.079 availability, creates the event, updates HubSpot, 00:17:11.140 --> 00:17:13.640 drafts the email using maybe info from the CRM, 00:17:13.640 --> 00:17:15.859 generates the invite, sends it all off, all handled 00:17:15.859 --> 00:17:18.079 automatically from one simple instruction. Okay, 00:17:18.160 --> 00:17:20.259 that's pretty mind -blowing. What's the fundamental 00:17:20.259 --> 00:17:23.180 shift then that this kind of direct AI control 00:17:23.180 --> 00:17:26.319 enables for, say, a business owner or a manager? 00:17:26.660 --> 00:17:28.980 How does it change their daily interaction with 00:17:28.980 --> 00:17:31.619 all their operational tools? It allows natural 00:17:31.619 --> 00:17:34.140 language commands, just... talking or typing 00:17:34.140 --> 00:17:36.700 like you normally would to orchestrate these 00:17:36.700 --> 00:17:40.240 incredibly complex multi -tool business operations 00:17:40.240 --> 00:17:43.220 you basically move from clicking through endless 00:17:43.220 --> 00:17:46.319 menus and dashboards in 10 different apps to 00:17:46.319 --> 00:17:48.559 simply telling your ai assistant what you need 00:17:48.559 --> 00:17:50.700 done it's like having a super competent executive 00:17:50.700 --> 00:17:53.259 assistant who's also plugged into every system 00:17:53.259 --> 00:17:56.839 you use okay so we've laid out 10 really powerful 00:17:56.839 --> 00:17:58.960 automation blueprints here incredible potential 00:17:58.960 --> 00:18:01.700 but here's where it gets really interesting maybe 00:18:01.700 --> 00:18:04.299 a bit daunting How do you even begin? How do 00:18:04.299 --> 00:18:06.240 you start implementing this stuff without trying 00:18:06.240 --> 00:18:08.460 to, you know, boil the entire ocean at once? 00:18:08.500 --> 00:18:10.539 For someone just starting out, where's the smart 00:18:10.539 --> 00:18:12.660 place to dive in first? Yeah, that's crucial. 00:18:12.859 --> 00:18:15.539 The key is definitely to be strategic. Focus 00:18:15.539 --> 00:18:17.339 on what the guide calls the 80 -20 approach, 00:18:17.579 --> 00:18:20.240 your high -impact starter pack. They strongly 00:18:20.240 --> 00:18:22.299 recommend starting with the four blueprints that 00:18:22.299 --> 00:18:24.180 tend to provide the biggest, most immediate ROI. 00:18:24.730 --> 00:18:27.069 Those are information extraction, that was hashtag 00:18:27.069 --> 00:18:30.410 three, lead scoring, hashtag four, text classification 00:18:30.410 --> 00:18:32.970 for email, hashtag eight, and setting up your 00:18:32.970 --> 00:18:35.369 first ROC system for internal knowledge, hashtag 00:18:35.369 --> 00:18:37.769 five. Those four often tackle the biggest time 00:18:37.769 --> 00:18:40.650 sinks and error sources right away. As for realistic 00:18:40.650 --> 00:18:42.950 implementation timeline, you can maybe aim for 00:18:42.950 --> 00:18:45.690 weeks one to two to nail down information extraction. 00:18:46.190 --> 00:18:48.759 Weeks three, four for lead scoring. Week five, 00:18:48.940 --> 00:18:51.140 six, getting email classification running smoothly. 00:18:51.400 --> 00:18:53.359 And week seven, eight, building out your first 00:18:53.359 --> 00:18:55.720 basic ROG system. Then, you know, months two, 00:18:55.839 --> 00:18:57.400 six, you can start layering in the more advanced 00:18:57.400 --> 00:19:00.099 stuff like the AI art director or the competitor 00:19:00.099 --> 00:19:02.579 watchtower. It's definitely a build, not a sprint. 00:19:02.740 --> 00:19:04.380 Right. Build, don't sprint. And once you are 00:19:04.380 --> 00:19:05.940 building, how do you know if it's actually working? 00:19:06.099 --> 00:19:08.400 I mean, beyond just feeling less busy, how do 00:19:08.400 --> 00:19:10.819 you measure what truly matters? Good point. You 00:19:10.819 --> 00:19:12.480 need to measure both the quantitative metrics 00:19:12.480 --> 00:19:16.150 and the qualitative metrics. So, quantitatively, 00:19:16.150 --> 00:19:18.869 you track things like hours saved per week, that's 00:19:18.869 --> 00:19:21.390 a big one, reductions in specific error rates, 00:19:21.549 --> 00:19:24.269 say, in data entry, improvements in customer 00:19:24.269 --> 00:19:26.470 response times, and, of course, increases in 00:19:26.470 --> 00:19:28.230 lead conversion rates if you've automated scoring. 00:19:28.630 --> 00:19:30.990 But don't forget the qualitative side. Things 00:19:30.990 --> 00:19:33.970 like noticeable drops in your team's stress levels, 00:19:34.109 --> 00:19:36.690 getting positive customer feedback, specifically 00:19:36.690 --> 00:19:38.910 mentioning faster or more accurate responses. 00:19:39.309 --> 00:19:42.230 Those matter too. Okay, so given all these possibilities, 00:19:42.369 --> 00:19:45.609 this accelerating future, What's the single most 00:19:45.609 --> 00:19:47.690 important mindset shift you think people need 00:19:47.690 --> 00:19:49.869 to make to really prepare for what's coming with 00:19:49.869 --> 00:19:52.750 automation? I think it really boils down to moving 00:19:52.750 --> 00:19:55.910 from being primarily a doer of tasks to becoming 00:19:55.910 --> 00:19:59.369 primarily a designer of systems. Your core role 00:19:59.369 --> 00:20:02.789 shifts from manually executing processes to thoughtfully 00:20:02.789 --> 00:20:05.470 designing, building, and overseeing these automated 00:20:05.470 --> 00:20:07.730 systems that execute the processes for you. It's 00:20:07.730 --> 00:20:09.789 about orchestrating, not just doing. So wrapping 00:20:09.789 --> 00:20:13.410 this all up. What does it all really mean? The 00:20:13.410 --> 00:20:16.809 big idea here seems to be this. We're moving 00:20:16.809 --> 00:20:21.089 beyond just interacting with AI as like a static 00:20:21.089 --> 00:20:23.809 tool in a window. We're moving towards building 00:20:23.809 --> 00:20:27.150 these integrated AI powered systems that operate 00:20:27.150 --> 00:20:30.289 autonomously, really transforming entire workflows 00:20:30.289 --> 00:20:32.329 from the ground up. That's exactly. It's about 00:20:32.329 --> 00:20:35.410 fundamentally changing how you work. shifting 00:20:35.410 --> 00:20:38.650 from being that doer of individual tasks to being 00:20:38.650 --> 00:20:41.009 the designer of intelligent systems. It's about 00:20:41.009 --> 00:20:43.289 building serious leverage for yourself and your 00:20:43.289 --> 00:20:46.309 business, multiplying your output capacity, and 00:20:46.309 --> 00:20:48.670 maybe most importantly, reclaiming your most 00:20:48.670 --> 00:20:51.900 valuable non -renewable asset, your time. You 00:20:51.900 --> 00:20:53.259 know, the guide we've walked through today, it 00:20:53.259 --> 00:20:55.420 really feels like it presents that classic red 00:20:55.420 --> 00:20:57.200 pill, blue pill moment for every entrepreneur, 00:20:57.339 --> 00:20:59.420 every business owner out there. Yeah, it really 00:20:59.420 --> 00:21:01.000 does. You can choose the blue pill, right? Keep 00:21:01.000 --> 00:21:02.660 doing everything manually. Keep trading your 00:21:02.660 --> 00:21:05.240 limited time for linear output. Constantly feel 00:21:05.240 --> 00:21:06.759 like you're struggling to keep up with the demands. 00:21:07.059 --> 00:21:10.420 Or you can take the red pill. Start implementing 00:21:10.420 --> 00:21:12.920 these kinds of automations. Reclaim that time. 00:21:13.299 --> 00:21:15.680 Multiply your output potential. And actually 00:21:15.680 --> 00:21:17.799 build a business that truly works for you instead 00:21:17.799 --> 00:21:20.299 of you constantly working for it. And think about 00:21:20.299 --> 00:21:22.430 this. While your competitors are maybe offline 00:21:22.430 --> 00:21:25.329 for the weekend, recharging. Your AI agents, 00:21:25.549 --> 00:21:29.150 they don't take weekends. They can be there 247, 00:21:29.349 --> 00:21:32.289 generating leads, qualifying prospects, answering 00:21:32.289 --> 00:21:34.069 customer questions, supporting your business 00:21:34.069 --> 00:21:36.029 around the clock. The tools are here. They're 00:21:36.029 --> 00:21:38.230 available. Many are low cost or even free to 00:21:38.230 --> 00:21:40.569 start. The blueprints, like the ones we discussed, 00:21:40.750 --> 00:21:42.670 have been laid out. So the only question left 00:21:42.670 --> 00:21:45.329 really isn't what is possible anymore. We know 00:21:45.329 --> 00:21:47.809 incredible things are possible. The real question 00:21:47.809 --> 00:21:50.619 now is, what will you build? Out to your music.