WEBVTT 00:00:00.000 --> 00:00:03.940 Imagine AI not just giving you advice, but actually 00:00:03.940 --> 00:00:06.059 doing tasks for you. Yeah, like really doing 00:00:06.059 --> 00:00:08.160 it. Clicking buttons, filling forms, adapting 00:00:08.160 --> 00:00:11.800 when some weird pop -up shows up. This isn't 00:00:11.800 --> 00:00:14.599 like science fiction anymore. It's here. It really 00:00:14.599 --> 00:00:17.039 is. And it's changing everything. Welcome to 00:00:17.039 --> 00:00:19.760 the deep dive. Today, we're taking a deep dive 00:00:19.760 --> 00:00:23.160 into what I see as a truly monumental leap in 00:00:23.160 --> 00:00:26.280 artificial intelligence, Chad GPT's agent mode. 00:00:26.440 --> 00:00:28.699 This isn't just another upgrade. It feels like 00:00:28.699 --> 00:00:31.539 a fundamental shift in how we can partner with 00:00:31.539 --> 00:00:33.880 AI. Exactly. It's gone from being this incredibly 00:00:33.880 --> 00:00:37.039 smart chatbot, I mean, a brilliant advisor, really 00:00:37.039 --> 00:00:40.079 to becoming a capable doer. We're going to explore 00:00:40.079 --> 00:00:42.520 this evolution from a virtual assistant giving 00:00:42.520 --> 00:00:44.719 great advice to an autonomous partner that can 00:00:44.719 --> 00:00:46.659 actually get stuff done for you. Our mission 00:00:46.659 --> 00:00:49.840 today is to cut through, let's say, the hype 00:00:49.840 --> 00:00:52.280 and clarify exactly what makes Agent Mode so 00:00:52.280 --> 00:00:54.429 special. We'll get into the crucial mindset shift 00:00:54.429 --> 00:00:56.909 you'll probably need to adopt to use it effectively. 00:00:57.270 --> 00:00:59.570 And then we'll walk you through seven practical 00:00:59.570 --> 00:01:02.829 methods to turn this AI into a kind of tireless 00:01:02.829 --> 00:01:06.109 24 -7 business partner. Think of it as your shortcut 00:01:06.109 --> 00:01:08.750 to freeing yourself up. You know, for those big 00:01:08.750 --> 00:01:11.290 picture strategic things that only you can really 00:01:11.290 --> 00:01:14.069 handle. We'll also touch on some important limits 00:01:14.069 --> 00:01:16.510 and yeah, ethical stuff too, so you're fully 00:01:16.510 --> 00:01:19.030 informed and ready to delegate wisely. Okay, 00:01:19.030 --> 00:01:22.799 so most of us use chat GPT in the Let's call 00:01:22.799 --> 00:01:26.060 it the conventional way, drafting emails, brainstorming 00:01:26.060 --> 00:01:28.859 content, maybe summarizing long articles. It's 00:01:28.859 --> 00:01:32.200 incredibly powerful there. But in that mode, 00:01:32.560 --> 00:01:34.659 we're really only tapping into a tiny fraction 00:01:34.659 --> 00:01:36.959 of its potential. It's like having a super wise 00:01:36.959 --> 00:01:40.719 friend who gives excellent advice, but stop short 00:01:40.719 --> 00:01:43.000 of actually helping you do it. Right, until now. 00:01:43.150 --> 00:01:45.549 Agent Mode flips that script completely. It's 00:01:45.549 --> 00:01:47.569 like your wise advisor has suddenly grew hands 00:01:47.569 --> 00:01:50.349 and, you know, initiative. It can complete tasks 00:01:50.349 --> 00:01:52.469 on autopilot. It goes beyond just telling you 00:01:52.469 --> 00:01:54.469 what to do. It's a difference between someone 00:01:54.469 --> 00:01:56.329 outlining the steps and someone actually walking 00:01:56.329 --> 00:01:58.650 the path for you. That's a profound difference. 00:01:58.890 --> 00:02:01.730 What's truly unique, and honestly still a little 00:02:01.730 --> 00:02:04.049 mind -bending for me, is its ability to take 00:02:04.049 --> 00:02:06.569 action within a virtual space. It runs its own 00:02:06.569 --> 00:02:09.469 virtual browser, which means it can engage with 00:02:09.469 --> 00:02:12.659 the web almost like a human would. It's not just 00:02:12.659 --> 00:02:15.500 pulling info, it's actively interacting. Yeah, 00:02:15.680 --> 00:02:17.979 that means clicking buttons, filling out forms, 00:02:18.240 --> 00:02:20.759 navigating between pages, all that stuff. And 00:02:20.759 --> 00:02:22.979 what's really impressive is how it adapts. It 00:02:22.979 --> 00:02:25.159 can handle unexpected pop -ups, change its plan 00:02:25.159 --> 00:02:27.680 if a website looks different than expected. This 00:02:27.680 --> 00:02:30.960 puts it way beyond simple automation tools, right? 00:02:31.240 --> 00:02:33.199 Those usually just break when something changes. 00:02:33.419 --> 00:02:35.919 That adaptability is the key. Yeah, I have to 00:02:35.919 --> 00:02:38.139 admit, I still sometimes wrestle with crafting 00:02:38.139 --> 00:02:41.389 the perfect initial prompt myself. getting the 00:02:41.389 --> 00:02:43.969 instructions just right. But agent mode seems 00:02:43.969 --> 00:02:46.750 to help bridge that gap. It reasons through the 00:02:46.750 --> 00:02:49.330 problem, adjusts. Even if my first request wasn't 00:02:49.330 --> 00:02:51.909 crystal clear, kind of works with you, like an 00:02:51.909 --> 00:02:54.770 apprentice. So how fundamentally does this change 00:02:54.770 --> 00:02:58.069 how we interact with AI? It shifts us from giving 00:02:58.069 --> 00:03:01.090 step -by -step instructions to delegating outcomes 00:03:01.090 --> 00:03:03.979 directly. That's the core change. OK, so this 00:03:03.979 --> 00:03:06.400 needs a real shift in thinking then, a completely 00:03:06.400 --> 00:03:08.479 different approach. Instead of those detailed 00:03:08.479 --> 00:03:10.659 commands, you just delegate the final goal. You 00:03:10.659 --> 00:03:13.340 say, find me 10 potential suppliers, not go to 00:03:13.340 --> 00:03:15.180 this page, click that button, search for this 00:03:15.180 --> 00:03:19.159 term. The AI handles the how. Precisely. And 00:03:19.159 --> 00:03:20.979 it figures out the necessary steps all on its 00:03:20.979 --> 00:03:23.900 own, which is, well, it's pretty incredible. 00:03:24.479 --> 00:03:27.639 However, there are some practical limitations 00:03:27.639 --> 00:03:30.069 to be aware of right now. especially around how 00:03:30.069 --> 00:03:32.270 much you can actually use it. Right. So for folks 00:03:32.270 --> 00:03:33.990 on the Plus Plan, you're currently looking at 00:03:33.990 --> 00:03:37.370 40 agent tasks per month. Now, that's definitely 00:03:37.370 --> 00:03:39.569 enough for some solid experimentation. You can 00:03:39.569 --> 00:03:41.550 really get a feel for it, automate some smaller 00:03:41.550 --> 00:03:43.849 things. But if you're a higher tier user, say 00:03:43.849 --> 00:03:46.849 on the Pro or Team Plans, that jumps significantly. 00:03:46.990 --> 00:03:50.169 You get 400 tasks per month. That's a much bigger 00:03:50.169 --> 00:03:52.789 capacity cost around $200 a month, I believe. 00:03:52.830 --> 00:03:55.370 It really enables deeper integration for businesses 00:03:55.370 --> 00:03:57.909 looking to scale its use. That's a serious commitment. 00:03:58.139 --> 00:04:00.919 And, you know, while these agents are getting 00:04:00.919 --> 00:04:03.520 smarter, some websites do still have ways to 00:04:03.520 --> 00:04:06.620 block AI bots. It's an ongoing thing. But the 00:04:06.620 --> 00:04:08.939 tech is evolving super fast. Agents are getting 00:04:08.939 --> 00:04:11.379 better at acting human -like, so those barriers 00:04:11.379 --> 00:04:14.159 are gradually lowering. It's kind of like a cat 00:04:14.159 --> 00:04:16.259 -and -mouse game. Okay. Considering all that, 00:04:16.600 --> 00:04:18.220 what's the biggest takeaway for someone just 00:04:18.220 --> 00:04:20.800 starting out with agent mode? Focus on delegating 00:04:20.800 --> 00:04:23.800 the end goal, not micromanaging the steps. Trust 00:04:23.800 --> 00:04:26.160 the AI. Right. And once you make that mental 00:04:26.160 --> 00:04:29.139 shift, the possibilities really open up. Let's 00:04:29.139 --> 00:04:31.420 dive into the first big method, transforming 00:04:31.420 --> 00:04:35.060 your market research. In any competitive market, 00:04:35.300 --> 00:04:38.410 knowledge is power, right? Agent mode can dig 00:04:38.410 --> 00:04:41.129 incredibly deep here, uncovering insights that 00:04:41.129 --> 00:04:43.750 might take a person weeks. Yeah, you start by 00:04:43.750 --> 00:04:45.449 giving it clear context about your business, 00:04:45.670 --> 00:04:48.230 who you are, what you do, who your ideal customer 00:04:48.230 --> 00:04:50.550 is. Make sure it gets your specific situation. 00:04:50.970 --> 00:04:53.529 Then you point it towards two key areas, your 00:04:53.529 --> 00:04:56.509 competitors and your target customers. For competitors, 00:04:56.870 --> 00:04:59.620 you can ask it to find Well, everything. Their 00:04:59.620 --> 00:05:01.500 product services, their unique selling points, 00:05:01.759 --> 00:05:05.360 pricing, marketing campaigns, even the feel of 00:05:05.360 --> 00:05:07.540 their website copy. It paints a really detailed 00:05:07.540 --> 00:05:11.100 picture. But the customer research. Oh, man, 00:05:11.100 --> 00:05:13.060 this is where it gets really good. The agent 00:05:13.060 --> 00:05:15.699 can visit forums, you know, like Quora, relevant 00:05:15.699 --> 00:05:18.899 Facebook groups, product review sites. And the 00:05:18.899 --> 00:05:21.379 goal is to grab word for word quotes. What are 00:05:21.379 --> 00:05:23.180 people's problems? What do they really want? 00:05:23.279 --> 00:05:24.699 What are their biggest complaints? What makes 00:05:24.699 --> 00:05:27.300 them happy? You want that raw voice. And then 00:05:27.300 --> 00:05:29.920 you can have it synthesize all that data. Maybe 00:05:29.920 --> 00:05:31.399 make a pie chart showing the top complaints. 00:05:31.480 --> 00:05:34.060 You see it visually instantly. Or generate a 00:05:34.060 --> 00:05:36.240 downloadable spreadsheet, neatly categorized. 00:05:36.519 --> 00:05:39.100 Problems, desires, complaints, praise, super 00:05:39.100 --> 00:05:42.199 organized. Imagine this prompt. Act as a market 00:05:42.199 --> 00:05:44.819 research analyst. My business is a web design 00:05:44.819 --> 00:05:47.290 company for dental clinics in Vietnam. Pretty 00:05:47.290 --> 00:05:50.430 specific, right? Task one, competitor analysis 00:05:50.430 --> 00:05:52.970 for the top three companies, put it in a markdown 00:05:52.970 --> 00:05:56.769 table with specific details. Task two, customer 00:05:56.769 --> 00:05:59.649 insights from forums, pull 20 verbatim comments 00:05:59.649 --> 00:06:02.329 on pain points, desires, barriers into a spreadsheet. 00:06:02.769 --> 00:06:05.629 Task three, synthesize it, create a pie chart 00:06:05.629 --> 00:06:09.009 of the five most common complaints. Boom, actionable 00:06:09.009 --> 00:06:11.350 intelligence. That's powerful using the customer's 00:06:11.350 --> 00:06:13.970 own words. How does that specific approach ensure 00:06:13.970 --> 00:06:16.410 better conversion? By reflecting the customer's 00:06:16.410 --> 00:06:18.910 own words, it builds instant trust and connection. 00:06:19.009 --> 00:06:21.129 They feel understood. Okay, moving from understanding 00:06:21.129 --> 00:06:24.110 the market to shaping your own strategy. Let's 00:06:24.110 --> 00:06:25.709 talk about something a lot of solo founders feel. 00:06:26.250 --> 00:06:28.209 Isolation. When you run a business alone, it's 00:06:28.209 --> 00:06:29.850 so easy to get stuck in your own head, right? 00:06:29.889 --> 00:06:31.649 There's no one really challenging your assumptions. 00:06:32.649 --> 00:06:34.810 And that's where agent mode can help you build, 00:06:35.129 --> 00:06:37.550 well, I like to call it a virtual advisory board. 00:06:38.069 --> 00:06:41.410 Picture this. You've got access to the thinking 00:06:41.410 --> 00:06:44.610 patterns of great business minds like Ray Dalio, 00:06:44.750 --> 00:06:48.110 Simon Sinek, Brene Brown, Adam Grant, Warren 00:06:48.110 --> 00:06:51.509 Buffett, all ready to give you feedback. A virtual 00:06:51.509 --> 00:06:53.730 brain trust. The setup is pretty wild. You pick 00:06:53.730 --> 00:06:55.709 your mentors, then you give the agent this massive 00:06:55.709 --> 00:06:57.970 research task. Go deep, find everything they've 00:06:57.970 --> 00:07:00.689 said, written, taught, capture their core philosophies, 00:07:00.709 --> 00:07:02.589 how they make decisions. Yeah, it grabs their 00:07:02.589 --> 00:07:04.750 decision -making frameworks, their go -to strategies, 00:07:04.930 --> 00:07:07.209 how they approach problems, their typical advice 00:07:07.209 --> 00:07:09.949 patterns. The agent compiles all this into these 00:07:09.949 --> 00:07:12.829 really comprehensive documents, maybe 20 to 50 00:07:12.829 --> 00:07:15.360 pages each. detailed profiles for each mentor. 00:07:15.540 --> 00:07:18.040 Then you load these profiles into a custom GPT, 00:07:18.379 --> 00:07:19.899 which, if you haven't used them, is basically 00:07:19.899 --> 00:07:22.660 your own personalized chat GPT trained on specific 00:07:22.660 --> 00:07:24.879 knowledge, in this case, your mentor's thinking. 00:07:25.220 --> 00:07:27.560 You're building a custom knowledge base. So whenever 00:07:27.560 --> 00:07:29.779 you hit a wall, you pitch your problem to this 00:07:29.779 --> 00:07:32.709 virtual board. And you get back genuine critique. 00:07:32.850 --> 00:07:35.449 It pokes holes in your logic. It's exactly that 00:07:35.449 --> 00:07:38.629 kind of tough love that a solo founder often 00:07:38.629 --> 00:07:41.810 really needs, but rarely gets. It's incredibly 00:07:41.810 --> 00:07:44.069 valuable for refining ideas. Right. For instance, 00:07:44.129 --> 00:07:47.029 you could prompt it. Act as a biographical researcher. 00:07:47.310 --> 00:07:49.829 Create a detailed profile of Ray Dalio's business 00:07:49.829 --> 00:07:52.730 philosophy. Access summaries of his book principles, 00:07:52.970 --> 00:07:55.670 YouTube interviews, articles. Extract his core 00:07:55.670 --> 00:07:58.089 principles, his five -step decision process, 00:07:58.370 --> 00:08:01.410 problem -solving approach. education style, write 00:08:01.410 --> 00:08:05.350 a 20 page doc for AI role play, include direct 00:08:05.350 --> 00:08:08.089 quotes. This makes the AI embody their perspective 00:08:08.089 --> 00:08:10.709 really well. What's the core benefit of this 00:08:10.709 --> 00:08:13.290 virtual advisory board? It provides critical, 00:08:13.649 --> 00:08:15.990 unbiased, tough love to challenge your assumptions, 00:08:16.230 --> 00:08:18.410 gets you out of that echo chamber. OK, shifting 00:08:18.410 --> 00:08:20.790 gears again, let's talk podcast guesting. If 00:08:20.790 --> 00:08:22.490 you're an expert wanting to share your knowledge, 00:08:22.689 --> 00:08:24.990 build your brand, guesting on podcasts is super 00:08:24.990 --> 00:08:27.579 effective. An agent mode can handle the research 00:08:27.579 --> 00:08:30.100 grunt work here, potentially saving you, well, 00:08:30.199 --> 00:08:33.000 quite a bit compared to hiring a PR agency. Totally. 00:08:33.759 --> 00:08:35.899 The setup starts with defining your expertise 00:08:35.899 --> 00:08:38.899 very clearly. What unique value do you offer? 00:08:39.340 --> 00:08:42.820 Then, state your goal cleanly. appear on as many 00:08:42.820 --> 00:08:45.200 relevant podcasts as possible. Then you give 00:08:45.200 --> 00:08:48.240 the agent specific criteria. The podcast must 00:08:48.240 --> 00:08:50.539 cover your topic area. It has to be active, published 00:08:50.539 --> 00:08:53.360 recently, say in the last 30 days. It needs to 00:08:53.360 --> 00:08:56.220 feature guests regularly, obviously. And ideally, 00:08:56.700 --> 00:08:59.379 have a decent listenership. Maybe aim for at 00:08:59.379 --> 00:09:01.500 least 1 ,000 regular listeners as a starting 00:09:01.500 --> 00:09:03.480 point. Makes it worth your time. And the agent 00:09:03.480 --> 00:09:05.830 comes back with a spreadsheet. all filled out. 00:09:05.950 --> 00:09:08.070 Podcast name, estimated listeners, host name, 00:09:08.429 --> 00:09:10.769 contact info, email, or form link. It can even 00:09:10.769 --> 00:09:12.929 suggest potential topic gaps you could pitch. 00:09:13.289 --> 00:09:14.929 Stuff they haven't covered recently that fits 00:09:14.929 --> 00:09:17.169 your expertise. It's incredibly helpful for pitching. 00:09:17.690 --> 00:09:19.610 Plus there's a cool bonus, automated scheduling. 00:09:19.950 --> 00:09:22.190 You can set agent mode to rerun the search daily, 00:09:22.389 --> 00:09:24.309 weekly, monthly, whatever you need. So you always 00:09:24.309 --> 00:09:26.210 have a fresh stream of potential opportunities 00:09:26.210 --> 00:09:29.009 flowing in without manually searching each time. 00:09:29.210 --> 00:09:32.330 Yeah, a prompt might look like. Act as a PR specialist. 00:09:32.570 --> 00:09:34.970 My expertise is productivity optimization for 00:09:34.970 --> 00:09:38.070 content creators. Search Google, Spotify, Apple 00:09:38.070 --> 00:09:41.470 podcasts for 15 podcasts, fitting these criteria 00:09:41.470 --> 00:09:44.289 topic, active guests, audience size, compile 00:09:44.289 --> 00:09:46.490 results into a markdown table, podcast name, 00:09:46.710 --> 00:09:49.149 host, link, contact notes, potential topics, 00:09:49.909 --> 00:09:53.269 super clear instructions. How does this specific 00:09:53.269 --> 00:09:56.029 approach ensure you find relevant opportunities, 00:09:56.090 --> 00:09:59.039 not just any podcast? Specific criteria filter 00:09:59.039 --> 00:10:01.740 for active podcasts that feature guests and actually 00:10:01.740 --> 00:10:04.100 match your expertise. It cuts out the irrelevant 00:10:04.100 --> 00:10:06.519 stuff. OK, next up, YouTube content research. 00:10:06.720 --> 00:10:08.600 If you're building a channel, finding topics 00:10:08.600 --> 00:10:11.120 and titles that actually get views, that's the 00:10:11.120 --> 00:10:12.779 perpetual challenge, right? You used to mean 00:10:12.779 --> 00:10:15.240 hours of manual digging or paying for expensive 00:10:15.240 --> 00:10:18.629 tools, a real time sink. Agent Mode totally streamlines 00:10:18.629 --> 00:10:20.990 this. You tell it, go to YouTube, research videos 00:10:20.990 --> 00:10:23.429 in my niche. It looks for what's doing well in 00:10:23.429 --> 00:10:25.789 search results, what has reasonable competition, 00:10:25.929 --> 00:10:27.690 not impossible to rank for, and maybe covers 00:10:27.690 --> 00:10:29.929 different video formats you need. And it gives 00:10:29.929 --> 00:10:32.450 you back a simple table. Here are some overall 00:10:32.450 --> 00:10:34.370 topics based on what people actually search for. 00:10:34.490 --> 00:10:36.730 And here's some compelling clickable titles for 00:10:36.730 --> 00:10:39.029 those topics. And for the power users, you can 00:10:39.029 --> 00:10:41.830 get more advanced. Ask it to find recent viral 00:10:41.830 --> 00:10:44.610 videos in your niche. Analyze the title structures 00:10:44.610 --> 00:10:47.250 that are working. The just ideas based on patterns 00:10:47.250 --> 00:10:49.549 from similar successful channels gives you a 00:10:49.549 --> 00:10:52.419 real edge. Here's a potent example prompt. Act 00:10:52.419 --> 00:10:55.480 as a YouTube content strategist. My channel is 00:10:55.480 --> 00:10:58.279 about indoor houseplants for beginners. Go to 00:10:58.279 --> 00:11:01.899 YouTube, search keywords like zplantcare, find 00:11:01.899 --> 00:11:04.360 10 topics with high search volume but moderate 00:11:04.360 --> 00:11:07.820 competition. For each topic, propose three title 00:11:07.820 --> 00:11:10.940 variations using common formulas, listicle, question, 00:11:11.159 --> 00:11:13.639 and how to. Whoa, I'm ending that running daily. 00:11:13.799 --> 00:11:16.440 Just getting a fresh list of validated content 00:11:16.440 --> 00:11:18.519 ideas handed to you without lifting a finger 00:11:18.519 --> 00:11:20.470 for the research part, that's... Yeah. That's 00:11:20.470 --> 00:11:23.149 huge for creators. Exactly. The power is getting 00:11:23.149 --> 00:11:26.389 topics driven by actual data, matching your expertise 00:11:26.389 --> 00:11:29.009 with real audience demand. And it doesn't maybe 00:11:29.009 --> 00:11:30.690 20 minutes while you're doing something else. 00:11:30.789 --> 00:11:32.269 That's a game changer. I mean, think of the hour 00:11:32.269 --> 00:11:35.009 saved. So what makes these generated titles so 00:11:35.009 --> 00:11:37.389 effective? Why are they better? They're data 00:11:37.389 --> 00:11:40.750 driven, keyword rich, and specifically optimized 00:11:40.750 --> 00:11:43.929 for clickability based on proven formulas. All 00:11:43.929 --> 00:11:46.649 right, let's talk websites. Every single moment 00:11:46.649 --> 00:11:49.620 of confusion a visitor has on your site. It chips 00:11:49.620 --> 00:11:51.639 away at your conversion rate, simple as that. 00:11:52.159 --> 00:11:54.200 And most business owners, while you're just too 00:11:54.200 --> 00:11:56.179 close to your own stuff to spot those friction 00:11:56.179 --> 00:11:59.279 points easily, you know your site too well. Agent 00:11:59.279 --> 00:12:01.820 Mode offers a really clever solution here, a 00:12:01.820 --> 00:12:04.740 website conversion audit. It basically looks 00:12:04.740 --> 00:12:06.679 at your site through the eyes of a potential 00:12:06.679 --> 00:12:09.399 customer focused on conversion rate optimization 00:12:09.399 --> 00:12:13.019 or CRO. Okay, so first, you tell the agent to 00:12:13.019 --> 00:12:15.419 find the top three, five businesses similar to 00:12:15.419 --> 00:12:18.360 yours, maybe nationwide leaders. These are usually 00:12:18.360 --> 00:12:20.720 companies that have likely spent a lot on optimizing 00:12:20.720 --> 00:12:22.960 their sites for conversion. They're great benchmarks. 00:12:23.360 --> 00:12:26.100 Then the agent meticulously studies their sales 00:12:26.100 --> 00:12:28.519 process. How do they structure information? What 00:12:28.519 --> 00:12:30.200 are their checkout steps? How do their forms 00:12:30.200 --> 00:12:32.960 work? Where's the pricing? Do they use live chat, 00:12:33.360 --> 00:12:36.720 payment plans, video demos? It analyzes everything. 00:12:36.899 --> 00:12:39.200 After studying the competition, it audits your 00:12:39.200 --> 00:12:42.259 funnel. a user going all the way through, from 00:12:42.259 --> 00:12:44.820 landing on your home page to checking out, experiencing 00:12:44.820 --> 00:12:47.220 it just like a real visitor would. And the output 00:12:47.220 --> 00:12:50.179 is this really clear comparison report. It identifies 00:12:50.179 --> 00:12:52.960 the specific gaps, like, hey, your top competitors 00:12:52.960 --> 00:12:55.179 show pricing with one click, but yours is hidden 00:12:55.179 --> 00:12:58.480 three pages deep. Actionable stuff. And finally, 00:12:58.519 --> 00:13:00.759 the agent prioritizes the suggested changes. 00:13:01.159 --> 00:13:03.419 It separates the quick wins things you could 00:13:03.419 --> 00:13:05.940 maybe fix this afternoon from the bigger strategic 00:13:05.940 --> 00:13:08.899 shifts, gives you a clear roadmap. A prompt could 00:13:08.899 --> 00:13:11.919 be something like, Act as a CRO specialist. My 00:13:11.919 --> 00:13:15.100 website sells online yoga courses. Task one, 00:13:15.519 --> 00:13:18.379 analyze three successful competitor sites. Document 00:13:18.379 --> 00:13:21.720 key elements, CTAs, pricing visibility, social 00:13:21.720 --> 00:13:25.000 proof, checkout steps. Task two, audit my website 00:13:25.000 --> 00:13:27.340 the same way. Task three, create a comparison 00:13:27.340 --> 00:13:29.340 report. Identify the five biggest differences. 00:13:29.539 --> 00:13:31.720 Recommend five immediate prioritized actions. 00:13:31.960 --> 00:13:34.340 And the real magic here is that the agent is 00:13:34.340 --> 00:13:36.720 actually clicking through the funnels like a 00:13:36.720 --> 00:13:40.740 person would. It feels the friction points. Every 00:13:40.740 --> 00:13:42.960 difference it flags is a potential improvement 00:13:42.960 --> 00:13:45.899 opportunity, often based on elements your top 00:13:45.899 --> 00:13:48.340 competitors have already tested and proven to 00:13:48.340 --> 00:13:50.120 work. How does this differ from just looking 00:13:50.120 --> 00:13:52.980 at standard website analytics, like bounce rate 00:13:52.980 --> 00:13:56.649 or time on page? It simulates a real user's journey, 00:13:56.950 --> 00:13:59.389 identifying specific friction points and usability 00:13:59.389 --> 00:14:02.389 issues directly, which analytics data can only 00:14:02.389 --> 00:14:04.669 hint at indirectly. OK, let's say your business 00:14:04.669 --> 00:14:06.330 is doing pretty well, but you're kind of hitting 00:14:06.330 --> 00:14:08.710 a plateau. You're wondering, what's next? How 00:14:08.710 --> 00:14:11.539 can I make more money? Agent Mode can actually 00:14:11.539 --> 00:14:13.720 help you brainstorm and discover new revenue 00:14:13.720 --> 00:14:16.600 streams. Find those new horizons. Yeah, it's 00:14:16.600 --> 00:14:18.940 great for unlocking new income sources you might 00:14:18.940 --> 00:14:21.360 not have thought of. You start by briefing the 00:14:21.360 --> 00:14:23.440 agent on your current business. What do you do? 00:14:23.620 --> 00:14:25.799 Who do you help? How do you make money now? And 00:14:25.799 --> 00:14:28.120 importantly, what are your limitations? Maybe 00:14:28.120 --> 00:14:30.139 you're limited by geography or you're at full 00:14:30.139 --> 00:14:32.899 capacity. Then you ask it for creative money 00:14:32.899 --> 00:14:35.600 -making ideas. Maybe categorize to make it easier 00:14:35.600 --> 00:14:39.200 to digest. Think like affiliate sales. What related 00:14:39.200 --> 00:14:41.340 products Could you recommend remote services? 00:14:41.419 --> 00:14:43.580 Could you offer consulting calls, planning sessions, 00:14:44.039 --> 00:14:46.320 or one -time creation ongoing sales templates, 00:14:46.740 --> 00:14:49.519 courses, guides people can buy anytime? You can 00:14:49.519 --> 00:14:51.940 even get specific about price points. Ask for 00:14:51.940 --> 00:14:55.299 ideas ranging from, say, a $27 download up to 00:14:55.299 --> 00:14:58.679 a $5 ,000 high -touch coaching package. It gives 00:14:58.679 --> 00:15:01.299 you a spectrum of possibilities. Right. Consider 00:15:01.299 --> 00:15:03.940 this prompt. Act as a business development consultant. 00:15:04.059 --> 00:15:06.419 My business is a portrait photography studio 00:15:06.419 --> 00:15:08.879 for entrepreneurs in Hanoi. We're at full capacity 00:15:08.879 --> 00:15:12.179 and geographically limited. Clear context. Suggest 00:15:12.179 --> 00:15:14.860 10 new revenue streams. Categorize them into 00:15:14.860 --> 00:15:16.899 affiliate marketing, reload services, and digital 00:15:16.899 --> 00:15:19.500 products. For each, give a description, suggest 00:15:19.500 --> 00:15:21.940 a price, and estimate the monthly revenue potential. 00:15:22.340 --> 00:15:24.620 Put it all in a PowerPoint. Very structured request. 00:15:24.879 --> 00:15:27.320 How does this approach ensure the ideas are actually 00:15:27.320 --> 00:15:30.240 new and relevant, not just generic stuff? It 00:15:30.240 --> 00:15:32.740 leverages your existing skills, client base, 00:15:33.039 --> 00:15:35.740 and specific limitations to generate creative, 00:15:36.139 --> 00:15:38.679 categorized suggestions tailored to you. It's 00:15:38.679 --> 00:15:41.759 grounded in your reality. Okay, last method before 00:15:41.759 --> 00:15:45.639 we talk ethics. Hiring. It's such a time drain, 00:15:45.960 --> 00:15:48.779 right? Finding good people takes forever. Agent 00:15:48.779 --> 00:15:50.919 mode can do that initial heavy lifting, sifting 00:15:50.919 --> 00:15:52.820 through candidates so you're reviewing the best 00:15:52.820 --> 00:15:55.279 fits, the real cream of the crop. Oh, tell me 00:15:55.279 --> 00:15:59.460 about it. I remember spending Days days looking 00:15:59.460 --> 00:16:01.980 through portfolios for a designer once by the 00:16:01.980 --> 00:16:04.440 end. My eyes were crossing Delegating that first 00:16:04.440 --> 00:16:07.399 pass sounds amazing. Imagine meeting a YouTube 00:16:07.399 --> 00:16:09.620 thumbnail designer You could tell the agent find 00:16:09.620 --> 00:16:11.879 ten designers on upwork their portfolios need 00:16:11.879 --> 00:16:14.539 to look like alia doll style And my budget is 00:16:14.539 --> 00:16:17.509 max fifty dollars per thumbnail Super specific. 00:16:18.009 --> 00:16:20.669 Exactly. You specify the exact skills, the experience 00:16:20.669 --> 00:16:22.870 level needed, budget constraints, maybe even 00:16:22.870 --> 00:16:24.669 personality traits if they're important. It cuts 00:16:24.669 --> 00:16:26.649 through all the noise and narrows the field dramatically. 00:16:26.809 --> 00:16:29.669 And it brings back this need spreadsheet, candidate 00:16:29.669 --> 00:16:32.870 name, their rate, a direct link to their profile 00:16:32.870 --> 00:16:36.470 on, say, Upwork, and critically, a link to an 00:16:36.470 --> 00:16:39.129 actual work sample that closely matches what 00:16:39.129 --> 00:16:41.509 you asked for. Make shortlisting so much easier. 00:16:41.629 --> 00:16:44.700 Yeah, a detailed prompt could be. Act as a recruiter. 00:16:44.899 --> 00:16:47.240 I need a YouTube thumbnail designer on Upwork. 00:16:47.820 --> 00:16:50.179 Requirements. Needs the Mr. Beast style. More 00:16:50.179 --> 00:16:53.100 than two years experience. Under $30 per thumbnail. 00:16:53.419 --> 00:16:57.340 Must have a 95 % plus job success score. Task, 00:16:57.679 --> 00:17:00.840 search Upwork, review portfolios, create a spreadsheet 00:17:00.840 --> 00:17:03.860 of the top five candidates, name Upwork profile 00:17:03.860 --> 00:17:06.420 link, rate, and a link to a thumbnail sample 00:17:06.420 --> 00:17:08.640 that matches the style. And remember, the agent 00:17:08.640 --> 00:17:10.819 runs in its own virtual browser, so it's not 00:17:10.819 --> 00:17:12.980 logged into your personal accounts. If it hits 00:17:12.980 --> 00:17:14.759 a site that needs a login, it'll just pause and 00:17:14.759 --> 00:17:16.720 ask you. You quickly log in for it, then let 00:17:16.720 --> 00:17:18.640 it continue its work. It keeps your credentials 00:17:18.640 --> 00:17:21.099 safe. So what could literally take you days of 00:17:21.099 --> 00:17:23.539 painful searching? The agent can knock out in 00:17:23.539 --> 00:17:25.740 maybe 20, 30 minutes. You get a qualified short 00:17:25.710 --> 00:17:28.130 lists with relevant work samples. That's just 00:17:28.130 --> 00:17:30.509 pure efficiency. So how does the agent handle 00:17:30.509 --> 00:17:32.930 sensitive login information if it needs access 00:17:32.930 --> 00:17:35.349 to a job board or something? It operates in its 00:17:35.349 --> 00:17:37.670 own secure virtual browser and will simply prompt 00:17:37.670 --> 00:17:40.490 you to log in manually if required. Your details 00:17:40.490 --> 00:17:44.869 aren't stored or exposed. OK, so obviously, with 00:17:44.869 --> 00:17:47.809 a tool this powerful, this versatile, Agent Mode, 00:17:48.970 --> 00:17:51.529 it's really crucial we talk about using it responsibly. 00:17:51.769 --> 00:17:54.170 You need to be thoughtful. Absolutely. First 00:17:54.170 --> 00:17:56.849 point has to be data privacy. You should never 00:17:56.849 --> 00:17:59.069 instruct the agent to collect sensitive personal 00:17:59.069 --> 00:18:01.569 information about people. Respecting privacy 00:18:01.569 --> 00:18:03.230 is paramount. It's just like if you were doing 00:18:03.230 --> 00:18:06.349 it yourself. Second, avoid anything spammy. Yeah, 00:18:06.430 --> 00:18:08.730 it can find contact info, but make sure your 00:18:08.730 --> 00:18:11.829 outreach is personalized, genuine, no mass, generic, 00:18:12.210 --> 00:18:14.549 emails. That's just bad practice anyway, AI or 00:18:14.549 --> 00:18:17.250 not. Quality over quantity. And maybe the most 00:18:17.250 --> 00:18:20.180 important point, human oversight. Yeah. Always, 00:18:20.200 --> 00:18:22.859 always review the agent's work. It's an incredibly 00:18:22.859 --> 00:18:25.339 powerful assistant, but it's still a tool. It 00:18:25.339 --> 00:18:27.460 doesn't replace your judgment. Double check anything 00:18:27.460 --> 00:18:29.720 critical before you act on it. The final decision, 00:18:29.980 --> 00:18:32.640 the responsibility is still yours. Why is that 00:18:32.640 --> 00:18:34.720 human oversight still so critical with agent 00:18:34.720 --> 00:18:37.079 mode, even though it's so capable? Because AI 00:18:37.079 --> 00:18:40.480 is fundamentally a tool. Human judgment, context, 00:18:40.680 --> 00:18:42.500 and ethical responsibility remain absolutely 00:18:42.500 --> 00:18:45.059 essential. It empowers you, but it doesn't absolve 00:18:45.059 --> 00:18:48.710 you. So wrapping up. We've taken a pretty deep 00:18:48.710 --> 00:18:52.289 dive into ChatGPT's agent mode today. The big 00:18:52.289 --> 00:18:55.089 idea, the real takeaway, is that AI has genuinely 00:18:55.089 --> 00:18:57.349 evolved beyond just thinking and writing for 00:18:57.349 --> 00:19:00.009 you. It's moved past being just an advisor. It's 00:19:00.009 --> 00:19:02.529 now an autonomous partner. It's capable of actually 00:19:02.529 --> 00:19:04.849 acting on your behalf across the web. This really 00:19:04.849 --> 00:19:06.970 changes the game, especially for solo entrepreneurs, 00:19:07.109 --> 00:19:09.869 small teams. It lets you delegate research, analysis, 00:19:09.970 --> 00:19:12.809 even complex support tasks, frees up your time 00:19:12.809 --> 00:19:15.289 for strategic thinking. That's invaluable. And 00:19:15.289 --> 00:19:17.839 a final thought to leave you with. As agent mode 00:19:17.839 --> 00:19:20.720 keeps getting better, keeps evolving, what seemingly 00:19:20.720 --> 00:19:23.160 human -only tasks might it tackle next? How else 00:19:23.160 --> 00:19:25.619 might it transform our daily work? Definitely 00:19:25.619 --> 00:19:27.980 something to mull over. We really encourage you 00:19:27.980 --> 00:19:30.319 to just start with one of these methods. Pick 00:19:30.319 --> 00:19:32.299 the one that addresses your biggest business 00:19:32.299 --> 00:19:34.839 bottleneck right now. Once you see the results, 00:19:35.119 --> 00:19:37.279 I guarantee you will quickly find ways to adapt 00:19:37.279 --> 00:19:39.480 these techniques to other parts of your business. 00:19:39.819 --> 00:19:42.640 The future of this kind of automation. It's not 00:19:42.640 --> 00:19:45.440 coming. It's here now. Thank you for joining 00:19:45.440 --> 00:19:48.400 us on this deep dive. Keep learning, keep exploring 00:19:48.400 --> 00:19:50.480 the possibilities, and we'll catch you next time.