WEBVTT 00:00:00.000 --> 00:00:02.580 AI was supposed to give us our time back. Yeah, 00:00:02.759 --> 00:00:04.980 that was the big promise. But somehow we're just 00:00:04.980 --> 00:00:08.140 spending more hours wrangling prompts. We're 00:00:08.140 --> 00:00:12.000 drowning in endless browser tabs all day. And 00:00:12.000 --> 00:00:14.439 we do manual data transfers every single day. 00:00:14.640 --> 00:00:16.940 It really feels like a modern technological trap. 00:00:17.100 --> 00:00:20.280 We definitely got faster at tiny individual tasks, 00:00:20.679 --> 00:00:24.179 but the overall workflow remains completely manual. 00:00:24.420 --> 00:00:26.420 Right. We are still the ones connecting the dots 00:00:26.420 --> 00:00:28.839 manually. We are the exhausted middle managers 00:00:28.839 --> 00:00:31.879 of our own tools. Exactly. Welcome to our deep 00:00:31.879 --> 00:00:34.700 dive for today. We are exploring a completely 00:00:34.700 --> 00:00:37.299 different approach to work. That is exactly our 00:00:37.299 --> 00:00:40.020 mission for today. We are moving past the exhausted 00:00:40.020 --> 00:00:42.520 consumer pattern completely. That pattern is 00:00:42.520 --> 00:00:44.500 where you manually switch between different tools. 00:00:44.619 --> 00:00:47.259 Right. Instead, we are learning to build automated 00:00:47.259 --> 00:00:49.259 background systems. We're going to do this using 00:00:49.259 --> 00:00:51.719 Claude and Notebook LM. I have to admit something 00:00:51.719 --> 00:00:54.340 vulnerable to you here. I still wrestle with 00:00:54.340 --> 00:00:58.259 prompt drift myself. B. You think you have a 00:00:58.259 --> 00:01:00.920 perfect instruction set. Oh, yeah. But then the 00:01:00.920 --> 00:01:05.060 AI slowly loses the plot over time. It is a constant 00:01:05.060 --> 00:01:09.359 frustrating battle every day. You are definitely 00:01:09.359 --> 00:01:12.060 not alone in that struggle. Most founders are 00:01:12.060 --> 00:01:15.609 feeling that exact same operational pain. AI 00:01:15.609 --> 00:01:18.530 speeds up our daily task execution wonderfully. 00:01:18.930 --> 00:01:21.150 But it doesn't reduce our operational complexity 00:01:21.150 --> 00:01:24.370 at all. It often just creates faster, messier 00:01:24.370 --> 00:01:27.469 information silos. Before we start building these 00:01:27.469 --> 00:01:30.670 new automated systems today, let's unpack the 00:01:30.670 --> 00:01:32.870 fundamental problem we are facing. Yeah, we need 00:01:32.870 --> 00:01:35.769 to understand this invisible drain on our energy. 00:01:35.969 --> 00:01:38.069 It comes directly from our fragmented working 00:01:38.069 --> 00:01:41.030 habits. The data on this fragmentation is actually 00:01:41.030 --> 00:01:43.670 staggering. It really is. Look at the recent 00:01:43.670 --> 00:01:46.269 McKinsey research on this topic. Knowledge workers 00:01:46.269 --> 00:01:49.349 lose 20 % of their week. Wow. That time is spent 00:01:49.349 --> 00:01:51.689 just searching for internal information. That 00:01:51.689 --> 00:01:54.409 is nearly a full workday every single week. You 00:01:54.409 --> 00:01:56.170 are just hunting for things across different 00:01:56.170 --> 00:01:58.969 apps. Yeah. And the APQC research backs this 00:01:58.969 --> 00:02:02.129 up entirely. They found workers lose nearly three 00:02:02.129 --> 00:02:05.150 hours weekly. Just trying to find necessary operational 00:02:05.150 --> 00:02:07.329 files. Exactly. They are searching through messy 00:02:07.329 --> 00:02:09.900 documents and messaging colleagues. It is an 00:02:09.900 --> 00:02:12.379 incredibly frustrating way to work. It drains 00:02:12.379 --> 00:02:14.639 your executive function before you even start. 00:02:14.860 --> 00:02:17.280 And it gets exponentially worse with team coordination. 00:02:17.620 --> 00:02:19.800 A three -person company loses several working 00:02:19.800 --> 00:02:22.979 days every week. Most businesses just treat this 00:02:22.979 --> 00:02:27.240 as normal operational overhead. But it absolutely 00:02:27.240 --> 00:02:29.659 shouldn't be normal for us anymore. Right. Asana 00:02:29.659 --> 00:02:32.780 actually found that 53 % of work time is wasted. 00:02:33.580 --> 00:02:35.879 They call this frustrating phenomenon work about 00:02:35.879 --> 00:02:38.830 work. That means endless searching, app switching, 00:02:39.330 --> 00:02:42.490 and internal coordination. We are managing the 00:02:42.490 --> 00:02:46.110 work instead of actually doing it. I have a fundamental 00:02:46.110 --> 00:02:49.310 question about our current approach. Does adding 00:02:49.310 --> 00:02:52.669 more advanced AI tools actually solve this, or 00:02:52.669 --> 00:02:55.090 does it just create more disjointed tabs to manage? 00:02:55.349 --> 00:02:57.990 Well, that is the crucial architectural question. 00:02:58.129 --> 00:03:00.289 Usually, adding more standalone tools makes the 00:03:00.289 --> 00:03:02.250 fragmentation worse. The biggest problem isn't 00:03:02.250 --> 00:03:05.319 any individual AI tool failing. No, it is the 00:03:05.319 --> 00:03:07.340 mechanical gap between the different tools. You 00:03:07.340 --> 00:03:09.819 do deep research in one browser tab today. Then 00:03:09.819 --> 00:03:11.759 you copy and save notes somewhere else entirely. 00:03:12.300 --> 00:03:15.080 Exactly. Then you write client summaries in yet 00:03:15.080 --> 00:03:17.599 another distinct place. Every time you hand off 00:03:17.599 --> 00:03:20.520 data manually, friction happens. That friction 00:03:20.520 --> 00:03:23.300 slowly drains your team's collective mental energy. 00:03:23.550 --> 00:03:27.009 So the friction of moving data manually drains 00:03:27.009 --> 00:03:29.789 our focus completely? Yes. We desperately need 00:03:29.789 --> 00:03:32.810 to build an automated bridge. Since manual handoffs 00:03:32.810 --> 00:03:35.550 are the clear enemy here, how do we actually 00:03:35.550 --> 00:03:38.330 bridge that digital gap mechanically? We combine 00:03:38.330 --> 00:03:41.030 two specific tools with opposite complementary 00:03:41.030 --> 00:03:44.289 strengths. We use Claude and Notebook LM together 00:03:44.289 --> 00:03:46.729 as a unified system. They are fundamentally different 00:03:46.729 --> 00:03:49.830 by their core architectural design. Right. Let's 00:03:49.830 --> 00:03:52.009 break down those distinct architectural differences 00:03:52.009 --> 00:03:55.169 carefully. Claude is essentially the active dynamic 00:03:55.169 --> 00:03:58.030 execution engine. It searches the live web for 00:03:58.030 --> 00:04:00.569 recent external context. Yeah, it monitors daily 00:04:00.569 --> 00:04:03.349 updates and moves quickly, but its outputs can 00:04:03.349 --> 00:04:05.650 lack strict factual grounding sometimes. It relies 00:04:05.650 --> 00:04:07.909 heavily on its vast but generalized training 00:04:07.909 --> 00:04:10.460 data. Notebook LM works in the exact opposite 00:04:10.460 --> 00:04:12.599 way conceptually. It isn't built to find new 00:04:12.599 --> 00:04:15.139 outside information independently. But it is 00:04:15.139 --> 00:04:19.079 a highly secure, grounded knowledge base. Let's 00:04:19.079 --> 00:04:21.220 pause and define that specific jargon for a second. 00:04:21.519 --> 00:04:24.300 Good idea. Grounded knowledge means AI answers 00:04:24.300 --> 00:04:27.360 restricted only to your approved business documents. 00:04:27.639 --> 00:04:31.060 That is a perfect, concise definition. Notebook 00:04:31.060 --> 00:04:34.800 LM uses your actual PDF files securely and strictly. 00:04:35.000 --> 00:04:37.420 It relies on your standard operating procedures 00:04:37.420 --> 00:04:40.079 and meeting notes. Separately, both of these 00:04:40.079 --> 00:04:42.540 tools feel a little incomplete. Claude pulls 00:04:42.540 --> 00:04:45.300 fresh context. Notebook LM keeps things strictly 00:04:45.300 --> 00:04:48.279 organized. Yeah, but once connected, it stops 00:04:48.279 --> 00:04:50.860 feeling like a simple AI chat. It starts working 00:04:50.860 --> 00:04:53.120 like an autonomous background system. But how 00:04:53.120 --> 00:04:55.639 do these two actually communicate without human 00:04:55.639 --> 00:04:58.259 intervention? Standard Cloud cannot natively 00:04:58.259 --> 00:05:00.620 pilot a web browser. Right. It cannot just organically 00:05:00.620 --> 00:05:02.819 log into your Notebook LM account. So how does 00:05:02.819 --> 00:05:05.339 the bridge actually work? That is where the technical 00:05:05.339 --> 00:05:08.420 mechanism becomes truly fascinating. We use a 00:05:08.420 --> 00:05:10.699 developer tool called Cloud Code Routines. It 00:05:10.699 --> 00:05:12.759 is a command line interface that runs locally. 00:05:13.139 --> 00:05:15.540 Exactly. It can execute Python scripts and handle 00:05:15.540 --> 00:05:18.220 API handoffs automatically. So it isn't just 00:05:18.220 --> 00:05:20.740 magic happening in the browser. Claude Code is 00:05:20.740 --> 00:05:22.620 actually running small scripts in the background. 00:05:22.839 --> 00:05:25.300 Right. You can instruct Claude Code to format 00:05:25.300 --> 00:05:27.639 its research findings. It saves those findings 00:05:27.639 --> 00:05:31.060 directly to a synced Google Drive folder. And 00:05:31.060 --> 00:05:33.180 Notebook LM is directly integrated with that 00:05:33.180 --> 00:05:36.180 specific Google Drive folder. Yes. It automatically 00:05:36.180 --> 00:05:39.259 ingests the new text files without manual uploads. 00:05:39.779 --> 00:05:42.550 Claude accesses the Scout. Notebook LM is the 00:05:42.550 --> 00:05:46.050 secure home base. Exactly. The Scout drops the 00:05:46.050 --> 00:05:49.050 Intel in a secure lockbox. The home base reads 00:05:49.050 --> 00:05:51.769 it and updates its internal maps. Now let's apply 00:05:51.769 --> 00:05:54.269 that setup to a realistic business scenario. 00:05:54.910 --> 00:05:57.069 Let's construct a hypothetical company to trace 00:05:57.069 --> 00:05:59.769 this data lifecycle. Let's do it. Imagine we 00:05:59.769 --> 00:06:02.089 are a software company selling logistics tools. 00:06:02.810 --> 00:06:04.350 Let's look at the front lines of our business 00:06:04.350 --> 00:06:07.339 first. We are talking about preparing for a high 00:06:07.339 --> 00:06:11.000 -stakes sales call. This is a classic, frustrating 00:06:11.000 --> 00:06:13.680 time sink for many founders. A meeting with a 00:06:13.680 --> 00:06:15.560 major shipping firm gets booked for tomorrow. 00:06:16.259 --> 00:06:18.519 Most founders spend 15 to 20 minutes researching 00:06:18.519 --> 00:06:20.920 manually. They skim through LinkedIn profiles 00:06:20.920 --> 00:06:23.300 and look at company websites. And they usually 00:06:23.300 --> 00:06:25.420 end up sounding totally generic on the call. 00:06:25.600 --> 00:06:27.600 Yeah, they sound like everyone else calling that 00:06:27.600 --> 00:06:30.639 exact same leak. The absolute best salespeople 00:06:30.639 --> 00:06:34.069 make prospects feel understood immediately. They 00:06:34.069 --> 00:06:36.310 do this before the conversation even formally 00:06:36.310 --> 00:06:39.389 starts. We can achieve that leverage using our 00:06:39.389 --> 00:06:42.509 Scout and Homebase. Yes. You instruct your Claude 00:06:42.509 --> 00:06:45.009 code routine to research the prospect deeply. 00:06:45.529 --> 00:06:48.550 You give Claude a highly structured, automated 00:06:48.550 --> 00:06:51.110 prompt. You tell it to research the prospect's 00:06:51.110 --> 00:06:53.689 exact company name. Right. You ask for their 00:06:53.689 --> 00:06:56.670 main logistical value proposition explicitly. 00:06:57.069 --> 00:06:59.589 You want recent supply chain blog posts from 00:06:59.589 --> 00:07:02.189 the last 60 days. You also want any recent news 00:07:02.189 --> 00:07:04.829 or industry press mentions. You even have it 00:07:04.829 --> 00:07:06.689 check the shipping founder's recent LinkedIn 00:07:06.689 --> 00:07:08.490 activity. Then it structures everything into 00:07:08.490 --> 00:07:11.189 a very specific, readable document. Exactly. 00:07:11.370 --> 00:07:14.230 It builds a company overview and recent market 00:07:14.230 --> 00:07:17.029 activity section. It outlines main logistical 00:07:17.029 --> 00:07:19.970 challenges and a suggested call angle. It saves 00:07:19.970 --> 00:07:22.110 the structured text directly into your connected 00:07:22.110 --> 00:07:24.230 Google Drive. Which immediately sinks the fresh 00:07:24.230 --> 00:07:27.279 external sources into your notebook LM. And this 00:07:27.279 --> 00:07:29.959 is where the magic synergy finally happens. Your 00:07:29.959 --> 00:07:32.399 notebook LM already holds your internal company 00:07:32.399 --> 00:07:35.540 case studies. It holds your complex pricing documents 00:07:35.540 --> 00:07:38.100 and detailed service overviews. Now the system 00:07:38.100 --> 00:07:40.259 sees both sides of the equation simultaneously. 00:07:40.420 --> 00:07:42.980 Right. It sees what the shipping prospect actually 00:07:42.980 --> 00:07:45.860 cares about today. And it sees what your logistic 00:07:45.860 --> 00:07:48.560 software can uniquely solve. The resulting synthesis 00:07:48.560 --> 00:07:51.189 is immediate. and highly customized for you. 00:07:51.370 --> 00:07:53.569 You get instant audio overviews of the tailed 00:07:53.569 --> 00:07:56.389 sales strategy. You get detailed strategic mind 00:07:56.389 --> 00:07:59.290 maps and specific call summaries. It takes mere 00:07:59.290 --> 00:08:02.310 minutes instead of nearly half an hour. It completely 00:08:02.310 --> 00:08:04.569 changes how you approach the initial discovery 00:08:04.569 --> 00:08:06.730 conversation. It feels much more like working 00:08:06.730 --> 00:08:09.509 with a senior strategy partner. I have a fundamental 00:08:09.509 --> 00:08:12.009 issue with this setup, though. My immediate fear 00:08:12.009 --> 00:08:14.290 is hallucinated sales pitches happening here. 00:08:14.670 --> 00:08:17.750 That is a very valid concern. we cannot promise 00:08:17.750 --> 00:08:19.810 routing features our engineering team hasn't 00:08:19.810 --> 00:08:23.310 built yet. If quad pulls wild claims, how do 00:08:23.310 --> 00:08:26.350 we stop it? That is exactly why Notebook LM is 00:08:26.350 --> 00:08:29.810 so architecturally crucial here. It actively 00:08:29.810 --> 00:08:32.330 cross -references the new external web research 00:08:32.330 --> 00:08:35.309 findings. It anchors everything strictly against 00:08:35.309 --> 00:08:38.250 your verified internal company materials. Exactly. 00:08:38.429 --> 00:08:41.590 It uses vector search to strictly map allowable 00:08:41.590 --> 00:08:44.450 claims. It simply will not invent new random 00:08:44.450 --> 00:08:46.830 software services for you. Right. It cross -checks 00:08:46.830 --> 00:08:49.730 web data against our actual company documents. 00:08:49.970 --> 00:08:53.029 Exactly. The Scout finds the opportunity, but 00:08:53.029 --> 00:08:55.870 the home base validates it. You still review 00:08:55.870 --> 00:08:59.000 the final output before the actual meeting. But 00:08:59.000 --> 00:09:01.399 the quality control is structurally embedded 00:09:01.399 --> 00:09:04.340 in the workflow. Yes. So you successfully use 00:09:04.340 --> 00:09:07.460 the automated system to close the deal. The shipping 00:09:07.460 --> 00:09:10.179 firm is now a paying logistics software client. 00:09:10.340 --> 00:09:12.960 Awesome. But what happens six months down the 00:09:12.960 --> 00:09:15.600 long operational line? Let's talk about long 00:09:15.600 --> 00:09:18.389 -term dynamic knowledge management now. This 00:09:18.389 --> 00:09:20.330 is where things usually break down badly for 00:09:20.330 --> 00:09:23.149 scaling companies. Your customer support AI starts 00:09:23.149 --> 00:09:26.009 giving clients outdated information. Your logistics 00:09:26.009 --> 00:09:28.850 product changes, but the AI remains totally stuck. 00:09:29.009 --> 00:09:32.169 Yeah. A stale AI is honestly kind of like a bad 00:09:32.169 --> 00:09:34.509 employee. It's like a senior salesperson who 00:09:34.509 --> 00:09:37.190 stopped reading critical company updates. They 00:09:37.190 --> 00:09:39.850 sound incredibly confident, but they are factually 00:09:39.850 --> 00:09:43.269 entirely wrong. That is honestly much worse than 00:09:43.269 --> 00:09:46.110 having no AI at all. You lose hard -earned client 00:09:46.110 --> 00:09:49.669 trust very quickly that way. How do we fix this 00:09:49.669 --> 00:09:52.230 inevitable frustrating knowledge drift? It's 00:09:52.230 --> 00:09:54.429 a process. Do we have to manually delete old 00:09:54.429 --> 00:09:57.090 vectors from the database? No, you put your core 00:09:57.090 --> 00:09:59.769 foundational documents into Notebook LM first. 00:10:00.110 --> 00:10:02.809 This includes your internal software FAQs and 00:10:02.809 --> 00:10:05.269 your client onboarding documents. This becomes 00:10:05.269 --> 00:10:07.929 the reliable main brain behind your customer 00:10:07.929 --> 00:10:11.009 support agent. Exactly. So the central organizational 00:10:11.009 --> 00:10:13.750 brain is set up securely. When major software 00:10:13.750 --> 00:10:16.289 features change, you only update those source 00:10:16.289 --> 00:10:19.000 documents. Yes. But to catch the subtle shifts, 00:10:19.399 --> 00:10:22.059 you set up a weekly edge case sweep. You let 00:10:22.059 --> 00:10:24.440 Claude routinely check places where novel problems 00:10:24.440 --> 00:10:26.700 usually appear. How does this sweep actually 00:10:26.700 --> 00:10:28.539 function mechanically in the background? You 00:10:28.539 --> 00:10:31.360 set up a scheduled Claude code routine via API. 00:10:31.740 --> 00:10:33.799 Claude connects to your support inbox and internal 00:10:33.799 --> 00:10:36.360 Slack channels securely. It looks closely at 00:10:36.360 --> 00:10:39.090 the last seven days of raw data. Right. It uses 00:10:39.090 --> 00:10:42.070 semantic search to find very specific types of 00:10:42.070 --> 00:10:45.090 anomalies. It looks for complex routing questions 00:10:45.090 --> 00:10:47.429 the support AI could not answer. Or intricate 00:10:47.429 --> 00:10:49.409 questions it answered incorrectly during the 00:10:49.409 --> 00:10:52.070 busy week. Yeah. It also finds client complaints 00:10:52.070 --> 00:10:55.110 suggesting your technical documentation is outdated. 00:10:55.629 --> 00:10:57.409 And it finds new software features mentioned 00:10:57.409 --> 00:10:59.870 internally by engineers but left undocumented. 00:11:00.429 --> 00:11:02.889 It finds the dangerous knowledge gaps automatically 00:11:02.889 --> 00:11:07.250 for you. Exactly. Claude synthesizes these messy 00:11:07.250 --> 00:11:11.230 new edge cases into crisp, short sentences. It 00:11:11.230 --> 00:11:13.230 formats them into a simple text file update. 00:11:13.370 --> 00:11:16.350 then it drops that file into the sync drive automatically 00:11:16.350 --> 00:11:18.750 every week. It's basically like stacking Lego 00:11:18.750 --> 00:11:21.350 blocks of data. You just keep adding new foundational 00:11:21.350 --> 00:11:24.250 pieces to the solid base. That is a brilliant 00:11:24.250 --> 00:11:26.970 visual way to picture the continuous process. 00:11:27.330 --> 00:11:29.529 The knowledge base systematically updates itself 00:11:29.529 --> 00:11:32.049 around the latest operational reality. Your support 00:11:32.049 --> 00:11:34.549 AI agent stays razor sharp instead of slowly 00:11:34.549 --> 00:11:37.129 drifting away. Right. But wait, is it fundamentally 00:11:37.129 --> 00:11:40.169 dangerous to let AI rewrite its own brain? That 00:11:40.169 --> 00:11:42.809 is a great question. What if it learns the entirely 00:11:42.809 --> 00:11:46.289 wrong procedural lesson from Slack? What if it 00:11:46.289 --> 00:11:49.730 misinterprets a sarcastic joke as a new company 00:11:49.730 --> 00:11:52.429 policy? That is a very profound and important 00:11:52.429 --> 00:11:55.090 safeguard to discuss now. The background workflow 00:11:55.090 --> 00:11:57.230 does all the heavy analytical lifting for you. 00:11:57.570 --> 00:12:00.269 It formats the proposed new rules cleanly for 00:12:00.269 --> 00:12:03.710 review. But a mandatory human review is always 00:12:03.710 --> 00:12:06.710 required before going live. Final contextual 00:12:06.710 --> 00:12:09.210 judgment should still belong entirely to your 00:12:09.210 --> 00:12:12.820 human team. The AI drafts the new rules, but 00:12:12.820 --> 00:12:16.340 a human clicks approve. Exactly. It saves massive 00:12:16.340 --> 00:12:18.779 analytical time without sacrificing your final 00:12:18.779 --> 00:12:22.019 executive control. This builds long -term operational 00:12:22.019 --> 00:12:25.059 trust. Which improves complex client retention 00:12:25.059 --> 00:12:28.379 immensely. Absolutely. Scaling your business 00:12:28.379 --> 00:12:30.519 operations requires technical systems you can 00:12:30.519 --> 00:12:32.960 actually trust. Building these resilient workflows 00:12:32.960 --> 00:12:35.080 takes the right kind of foundational infrastructure. 00:12:35.440 --> 00:12:38.139 Check our detailed show notes for resources on 00:12:38.139 --> 00:12:41.080 building these reliable background systems. Now, 00:12:41.100 --> 00:12:42.980 let's get back to our deep dive conversation. 00:12:43.299 --> 00:12:45.399 Let's do it. Okay, we are back to our deep dive 00:12:45.399 --> 00:12:48.159 exploration today. We have successfully secured 00:12:48.159 --> 00:12:50.679 our automated sales preparation workflow. We 00:12:50.679 --> 00:12:53.639 have stabilized our internal customer support 00:12:53.639 --> 00:12:57.779 knowledge base beautifully. Now, how do we monitor 00:12:57.779 --> 00:13:01.000 the outside world automatically? We do it systematically 00:13:01.000 --> 00:13:03.820 without lifting a single manual finger. Running 00:13:03.820 --> 00:13:06.039 a tech business without continuous competitive 00:13:06.039 --> 00:13:09.200 intelligence is highly dangerous. It is quite 00:13:09.200 --> 00:13:10.980 literally like driving down the highway with 00:13:10.980 --> 00:13:13.480 your mirrors covered. Exactly. You can theoretically 00:13:13.480 --> 00:13:15.740 move forward, but you are totally strategically 00:13:15.740 --> 00:13:18.879 blind. You won't see the sudden market shift 00:13:18.879 --> 00:13:21.879 until it crashes into you. So we intentionally 00:13:21.879 --> 00:13:25.139 build a dedicated competitive radar system. You 00:13:25.139 --> 00:13:27.220 set this up directly inside your notebook LM 00:13:27.220 --> 00:13:29.720 environment first. You add your top competitors 00:13:29.720 --> 00:13:32.820 main marketing websites. And their detailed pricing 00:13:32.820 --> 00:13:35.799 pages. You also add your own strategic positioning 00:13:35.799 --> 00:13:39.519 document or current pitch deck. Right. That establishes 00:13:39.519 --> 00:13:41.879 your baseline strategic context for the entire 00:13:41.879 --> 00:13:45.019 AI system. That way, Notebook LM automatically 00:13:45.019 --> 00:13:47.919 notices what actually changed recently. It understands 00:13:47.919 --> 00:13:50.519 the context of why a competitor's change actually 00:13:50.519 --> 00:13:52.879 matters. Yeah. It takes about five minutes to 00:13:52.879 --> 00:13:55.679 do this initial foundational setup. Then you 00:13:55.679 --> 00:13:57.980 use Claude code routines to give the system legs. 00:13:58.259 --> 00:14:00.799 This lets you schedule a recurring automated 00:14:00.799 --> 00:14:03.960 weekly research task. You literally write scheduled 00:14:03.960 --> 00:14:06.080 background instructions for the AI to follow 00:14:06.080 --> 00:14:09.399 blindly. Every single Sunday night at 8 p .m., 00:14:09.399 --> 00:14:11.860 Claude quietly goes to work. Let's say we are 00:14:11.860 --> 00:14:14.899 tracking competitors like Linear, Notion, or 00:14:14.899 --> 00:14:17.419 ClickUp. It scans their external sites for subtle 00:14:17.419 --> 00:14:20.080 pricing tier changes. And new product launches. 00:14:20.600 --> 00:14:23.120 It diligently checks for new engineering job 00:14:23.120 --> 00:14:25.500 postings and subtle press mentions. It looks 00:14:25.500 --> 00:14:27.840 strictly at fresh data from the last seven days. 00:14:28.059 --> 00:14:30.820 Right. It meticulously organizes all of those 00:14:30.820 --> 00:14:32.779 strategic findings by a specific competitor. 00:14:33.419 --> 00:14:35.940 Then it formats a clean update file for the synced 00:14:35.940 --> 00:14:37.919 drive. It adds the most important competitive 00:14:37.919 --> 00:14:40.779 findings as new structured sources. Then it updates 00:14:40.779 --> 00:14:43.940 the internal notebook LM strategic mind map automatically. 00:14:44.100 --> 00:14:46.019 Yeah. It generates a short executive summary 00:14:46.019 --> 00:14:49.690 and a synthesized audio overview. Whoa. Imagine 00:14:49.690 --> 00:14:52.590 waking up Monday to a fully generated audio briefing 00:14:52.590 --> 00:14:55.429 of your competitors every move. Two -sex silence. 00:14:56.070 --> 00:14:58.809 That is just incredible strategic power for any 00:14:58.809 --> 00:15:01.250 modern software founder. But how does this actually 00:15:01.250 --> 00:15:03.289 run completely without the user clicking anything? 00:15:03.610 --> 00:15:06.320 Good question. I thought Claude code always required 00:15:06.320 --> 00:15:09.120 a manual terminal command to execute. How do 00:15:09.120 --> 00:15:11.000 we make it fully autonomous in the background? 00:15:11.139 --> 00:15:13.440 That is the absolute beauty of the operating 00:15:13.440 --> 00:15:15.960 system's native scheduling tools. You can use 00:15:15.960 --> 00:15:19.179 standard cron jobs on Mac or task scheduler on 00:15:19.179 --> 00:15:21.919 Windows. You wrap the Claude code command in 00:15:21.919 --> 00:15:24.220 a simple repairing background script. Exactly. 00:15:24.580 --> 00:15:26.480 Once you set that routine, it triggers itself 00:15:26.480 --> 00:15:28.759 completely automatically. You basically program 00:15:28.759 --> 00:15:32.259 a weekly alarm clock for an AI spy. That is exactly 00:15:32.259 --> 00:15:34.600 what it is in practical business reality. And 00:15:34.600 --> 00:15:36.799 it only takes about 20 minutes to set up initially. 00:15:37.000 --> 00:15:40.039 After that brief setup, it mostly runs completely 00:15:40.039 --> 00:15:43.320 on its own forever. We really need to slow down 00:15:43.320 --> 00:15:46.039 and synthesize this whole journey. Yeah, let's 00:15:46.039 --> 00:15:48.879 recap. The massive takeaway here is not just 00:15:48.879 --> 00:15:52.100 about saving mundane time. Yes, saving 40 hours 00:15:52.100 --> 00:15:54.580 a month is a genuinely fantastic operational 00:15:54.580 --> 00:15:57.809 result. But it is really a fundamental profound 00:15:57.809 --> 00:16:00.509 shift in your psychological mindset. Most casual 00:16:00.509 --> 00:16:03.590 consumers use individual AI apps to get a single 00:16:03.590 --> 00:16:06.830 isolated answer. They start from absolute ground 00:16:06.830 --> 00:16:09.970 zero every single time they type. A tool gives 00:16:09.970 --> 00:16:12.690 you one static answer when you finally ask it. 00:16:12.730 --> 00:16:15.629 Right. Smart founders build interconnected background 00:16:15.629 --> 00:16:17.929 systems that work tirelessly in the shadows. 00:16:18.250 --> 00:16:21.149 These integrated systems keep generating better, 00:16:21.409 --> 00:16:24.789 richer answers over time naturally. They aggressively 00:16:24.789 --> 00:16:27.490 compound your operational market advantage while 00:16:27.490 --> 00:16:29.909 you sleep peacefully. If you only build one single 00:16:29.909 --> 00:16:32.909 system from today, listen closely now. Make it 00:16:32.909 --> 00:16:35.250 the automated competitive radar we just discussed 00:16:35.250 --> 00:16:37.850 deeply. It takes just 20 short minutes to set 00:16:37.850 --> 00:16:40.169 up once today. It runs automatically every single 00:16:40.169 --> 00:16:42.990 week for you without fail. It will completely 00:16:42.990 --> 00:16:44.950 change how you confidently start your Monday 00:16:44.950 --> 00:16:47.909 mornings. Here's a final slightly unsettling 00:16:47.909 --> 00:16:49.909 thought for you to ponder today. Oh, I am ready. 00:16:50.350 --> 00:16:53.490 Our background AI agents are constantly researching 00:16:53.490 --> 00:16:55.909 our industry competitors now. They're constantly 00:16:55.909 --> 00:16:58.929 adapting to the subtle daily market changes automatically. 00:16:59.090 --> 00:17:01.490 Right. But our smart competitors are undoubtedly 00:17:01.490 --> 00:17:04.750 building background AI agents too. Yeah. Are 00:17:04.750 --> 00:17:06.869 we rapidly approaching a future where businesses 00:17:06.869 --> 00:17:10.819 are just massive AI ecosystems? ecosystems quietly 00:17:10.819 --> 00:17:13.180 negotiating and competing with each other constantly 00:17:13.180 --> 00:17:15.980 in the background, while we humans just sit back, 00:17:16.200 --> 00:17:18.380 sip coffee, and read the Monday morning summaries. 00:17:18.619 --> 00:17:21.380 Mm -hmm. That is a wild, fascinating thought 00:17:21.380 --> 00:17:23.359 to leave on today. Keep your strategic mirrors 00:17:23.359 --> 00:17:26.079 uncovered out there. We will see you next time. 00:17:26.460 --> 00:17:27.299 OTR music.