WEBVTT 00:00:00.000 --> 00:00:03.180 So here's a question. Why are you still paying 00:00:03.180 --> 00:00:07.419 $20 a month for advanced high -end AI tools? 00:00:07.459 --> 00:00:11.900 Right. Especially when Google just quietly released 00:00:11.900 --> 00:00:15.439 this entire suite of enterprise -grade features. 00:00:15.720 --> 00:00:18.219 And 99 % of it is completely free to use right 00:00:18.219 --> 00:00:20.339 now. Exactly. That really is the essential question. 00:00:21.120 --> 00:00:24.100 We are absolutely drowning in AI updates. But 00:00:24.100 --> 00:00:27.579 this one, this Google AI Studio update. It's 00:00:27.579 --> 00:00:31.980 a big, quiet shift that actually matters. I think 00:00:31.980 --> 00:00:34.320 most people still think AI is just a chat bot, 00:00:34.320 --> 00:00:36.740 you know, for writing emails or something. Right, 00:00:36.759 --> 00:00:39.229 just a text generator. But what the source material 00:00:39.229 --> 00:00:41.729 for today shows us is that this suite lets you 00:00:41.729 --> 00:00:44.810 build real, functional apps, create professional 00:00:44.810 --> 00:00:47.409 video, generate custom stock photos. Well, without 00:00:47.409 --> 00:00:50.090 hiring a developer or a designer. Exactly. So 00:00:50.090 --> 00:00:52.070 our mission today is to cut through all that 00:00:52.070 --> 00:00:54.070 noise and give you the actionable shortcuts. 00:00:54.450 --> 00:00:56.250 We've done the deep dive, and we're going to 00:00:56.250 --> 00:00:59.409 pull out five really powerful no -cost applications 00:00:59.409 --> 00:01:02.070 you can start using today. For any business or 00:01:02.070 --> 00:01:03.850 creative project, we're going to map it all out 00:01:03.850 --> 00:01:06.030 for you. We'll cover building apps with what 00:01:06.030 --> 00:01:08.450 they call vibe coding. Getting studio quality 00:01:08.450 --> 00:01:11.569 audio from text, creating b -roll video, live 00:01:11.569 --> 00:01:14.150 feedback on your work. And making custom images 00:01:14.150 --> 00:01:16.689 that actually match your brand. Let's do it. 00:01:16.870 --> 00:01:18.989 OK, let's unpack this. I think we should start 00:01:18.989 --> 00:01:20.689 with the one that feels like the biggest leak 00:01:20.689 --> 00:01:24.090 forward. Building your own custom apps without 00:01:24.090 --> 00:01:26.450 touching a single line of code. This is their 00:01:26.450 --> 00:01:30.049 build feature. It's built on this idea of vibe 00:01:30.049 --> 00:01:32.349 coding. Vibe coding. Yeah, you literally just 00:01:32.349 --> 00:01:35.219 create a functional app. by describing the tool 00:01:35.219 --> 00:01:39.200 you want in simple, plain English. And the real 00:01:39.200 --> 00:01:41.780 value here, the shortcut for you, is this two 00:01:41.780 --> 00:01:43.900 -step prompting process. Don't go straight into 00:01:43.900 --> 00:01:46.319 the build tool. Right, that's the key. Start 00:01:46.319 --> 00:01:49.079 in the normal Google Gemini chat, describe your 00:01:49.079 --> 00:01:52.140 app idea just, you know, loosely, and then ask 00:01:52.140 --> 00:01:54.680 Gemini to write the clear, structured prompt 00:01:54.680 --> 00:01:56.859 for the build tool. And that works so well because 00:01:56.859 --> 00:02:00.140 it takes your kind of messy thoughts and organizes 00:02:00.140 --> 00:02:03.109 them perfectly for the machine. It turns ambiguity 00:02:03.109 --> 00:02:05.450 into clear instructions. It does. I mean, if 00:02:05.450 --> 00:02:07.769 we look at the recipe idea tool example from 00:02:07.769 --> 00:02:09.650 the source. Yeah, you just type in the ingredients 00:02:09.650 --> 00:02:11.650 you have in your fridge. And the app it builds 00:02:11.650 --> 00:02:14.789 spits out three recipes with steps and even a 00:02:14.789 --> 00:02:16.849 shopping list for what you're missing. What's 00:02:16.849 --> 00:02:20.009 so wild is the speed of iteration. Oh, it's instant. 00:02:20.360 --> 00:02:22.159 You can just give it feedback right there in 00:02:22.159 --> 00:02:24.280 the chat. You can type, OK, great, now add a 00:02:24.280 --> 00:02:27.400 calorie estimate, and poof. The app just updates. 00:02:27.719 --> 00:02:30.580 You've refined a working tool in seconds. And 00:02:30.580 --> 00:02:32.780 when you're happy, you get options to save it, 00:02:33.080 --> 00:02:36.860 download the code, or, the big one, deploy. That's 00:02:36.860 --> 00:02:39.060 how you make it live and shareable. Now that 00:02:39.060 --> 00:02:42.180 does require setting up a simple Google Cloud 00:02:42.180 --> 00:02:45.000 project to finalize. Right. And for listeners 00:02:45.000 --> 00:02:47.240 who aren't technical, that part can sound a little 00:02:47.240 --> 00:02:49.460 intimidating. Is it really simple, or is that 00:02:49.460 --> 00:02:51.900 just marketing speak? No, it's fair. The cloud 00:02:51.900 --> 00:02:54.139 environment sounds scary. But for this, it's 00:02:54.139 --> 00:02:55.819 really just setting up an account and giving 00:02:55.819 --> 00:02:58.759 your app a virtual address to live at. It's free 00:02:58.759 --> 00:03:00.860 for these basic uses, and you don't touch any 00:03:00.860 --> 00:03:03.159 complex stuff. So the barrier is actually pretty 00:03:03.159 --> 00:03:07.169 low. Very low. Whoa. to imagine building and 00:03:07.169 --> 00:03:10.409 deploying a custom functional app in under five 00:03:10.409 --> 00:03:13.610 minutes. It just shows if the prompt is the key, 00:03:14.210 --> 00:03:16.169 you don't need to be a programmer anymore. You 00:03:16.169 --> 00:03:18.110 just need to be a better communicator with the 00:03:18.110 --> 00:03:20.849 AI. Exactly. If a business wants to build an 00:03:20.849 --> 00:03:24.110 idea app, that structured prompt is everything. 00:03:24.409 --> 00:03:27.689 Defining the input. like a skill or budget, ensures 00:03:27.689 --> 00:03:31.530 you get a specific structured output, like five 00:03:31.530 --> 00:03:34.370 ideas, a description, and the first three steps 00:03:34.370 --> 00:03:37.509 for each. That structure is vital. Now let's 00:03:37.509 --> 00:03:39.810 shift over to the audio space. If you're creating 00:03:39.810 --> 00:03:42.550 any kind of content, newsletters, blog posts, 00:03:43.270 --> 00:03:45.469 you know a huge chunk of your audience would 00:03:45.469 --> 00:03:48.259 rather listen than read. And this generate native 00:03:48.259 --> 00:03:50.879 speech with Gemini feature is surprisingly high 00:03:50.879 --> 00:03:54.139 quality. It avoids that old robotic sound. It's 00:03:54.139 --> 00:03:56.000 very natural. It's perfect for turning written 00:03:56.000 --> 00:03:58.000 content into something engaging really, really 00:03:58.000 --> 00:03:59.719 quickly. They give you two main modes, right? 00:04:00.219 --> 00:04:01.639 Right. Single speaker is great for, you know, 00:04:01.759 --> 00:04:04.240 reading a blog post. But multi -speaker is where 00:04:04.240 --> 00:04:06.379 it gets interesting. For Q &As, interviews. Or 00:04:06.379 --> 00:04:09.439 for training content where you need to have distinct 00:04:09.439 --> 00:04:11.580 voices to make things clear. And this is where 00:04:11.580 --> 00:04:13.939 the control gets really granular with the technical 00:04:13.939 --> 00:04:16.519 nuance of temperature. Yeah, temperature is basically 00:04:16.519 --> 00:04:18.800 the creativity dial for the voice. It goes from 00:04:18.800 --> 00:04:21.939 zero to two. Zero being totally flat and robotic. 00:04:22.180 --> 00:04:24.980 The least expressive. And two is the most expressive, 00:04:25.019 --> 00:04:27.680 most natural. It varies its pace and tone. It's 00:04:27.680 --> 00:04:29.680 the difference between a teleprompter read and 00:04:29.680 --> 00:04:31.920 a real conversation. Exactly. And the sources 00:04:31.920 --> 00:04:34.560 all say you have to experiment here. Don't start 00:04:34.560 --> 00:04:38.959 at zero. We'd say start around 1 .2. And test 00:04:38.959 --> 00:04:40.939 a few voices, because the pace and tone change 00:04:40.939 --> 00:04:43.589 a lot. Depending on your settings. So practically 00:04:43.589 --> 00:04:46.069 this means you can instantly make audio versions 00:04:46.069 --> 00:04:49.769 of newsletters or Turn dry training guides into 00:04:49.769 --> 00:04:52.089 like little dialogues or even turn written drafts 00:04:52.089 --> 00:04:55.290 into actual multi -voice podcast style episodes 00:04:55.290 --> 00:04:58.629 save so much time How does using that multi -speaker 00:04:58.629 --> 00:05:01.230 mode really elevate training content compared 00:05:01.230 --> 00:05:03.490 to just plain text? Well, you can assign different 00:05:03.490 --> 00:05:05.750 voices, different personalities. You can set 00:05:05.750 --> 00:05:08.529 one voice to a low measured temperature as the 00:05:08.529 --> 00:05:11.170 narrator, and another voice to a higher temperature 00:05:11.170 --> 00:05:14.829 as the enthusiastic questioner. So it makes complex 00:05:14.829 --> 00:05:16.930 ideas feel more like a real conversation. It 00:05:16.930 --> 00:05:19.889 makes them way more engaging. OK, so moving from 00:05:19.889 --> 00:05:23.569 sound to sight, let's talk video. Video 3 .1. 00:05:23.660 --> 00:05:26.819 It's a massive step up. It produces professional 00:05:26.819 --> 00:05:29.100 looking footage. I mean, we're talking smooth 00:05:29.100 --> 00:05:32.579 camera movement, great lighting, that nice background 00:05:32.579 --> 00:05:35.160 blur. And it even adds sound effects, which is 00:05:35.160 --> 00:05:38.259 new. It's a huge leap past those old kind of 00:05:38.259 --> 00:05:41.699 jittery models. For sure. But the real smart 00:05:41.699 --> 00:05:45.100 way to start a hack for any listener is the VO3 00:05:45.100 --> 00:05:47.139 gallery. Absolutely. The best way to learn is 00:05:47.139 --> 00:05:49.519 to imitate, not just trial and error. And in 00:05:49.519 --> 00:05:52.100 the gallery, you can see the exact prompt they 00:05:52.100 --> 00:05:54.779 use for every single one of those amazing example 00:05:54.779 --> 00:05:56.480 videos. It's like your personal training ground. 00:05:56.560 --> 00:05:58.699 You can literally just copy a prompt, edit it, 00:05:58.759 --> 00:06:00.970 and tailor it for what you need. So say you need 00:06:00.970 --> 00:06:03.769 some b -roll for a YouTube video, you could type 00:06:03.769 --> 00:06:06.470 a prompt like medium shot of a person at a wooden 00:06:06.470 --> 00:06:08.990 desk typing a report, soft sunlight streaming 00:06:08.990 --> 00:06:11.629 from a window behind them. And in about 60 seconds 00:06:11.629 --> 00:06:14.730 you get realistic usable b -roll footage. And 00:06:14.730 --> 00:06:16.810 if you need it to be longer, the extend function 00:06:16.810 --> 00:06:18.810 is super simple. You just tell it extend this 00:06:18.810 --> 00:06:20.949 for five more seconds and pan the camera to the 00:06:20.949 --> 00:06:23.110 right. And it just continues the shot perfectly. 00:06:23.329 --> 00:06:25.730 Now, you have an expert tip on the cost strategy 00:06:25.730 --> 00:06:27.410 here, because there are two versions. Right. 00:06:27.550 --> 00:06:29.970 This is important. VO2 is completely free. It 00:06:29.970 --> 00:06:33.209 works great. VO3 .1 is the latest, highest quality 00:06:33.209 --> 00:06:36.410 version, but it needs a very cheap Gemini API 00:06:36.410 --> 00:06:39.470 key. So this is a classic efficiency play. Exactly. 00:06:39.680 --> 00:06:42.779 You use the free V2 to test and perfect your 00:06:42.779 --> 00:06:45.639 prompts. Get the lighting, the style, the movement 00:06:45.639 --> 00:06:48.600 exactly right. And only when you know precisely 00:06:48.600 --> 00:06:50.660 what you want. You take that perfected prompt 00:06:50.660 --> 00:06:54.160 over to V3 .1 for the final high quality sound 00:06:54.160 --> 00:06:56.240 enhanced export. If I'm filming something like 00:06:56.240 --> 00:06:59.279 a product demo, should I use V2 or V3 .1 for 00:06:59.279 --> 00:07:02.199 that final export? You test with V2 to get the 00:07:02.199 --> 00:07:04.459 prompt perfect, but for the final version you're 00:07:04.459 --> 00:07:08.290 going to share. Always use V3 .1. The quality 00:07:08.290 --> 00:07:10.329 and the sound enhancement make it worth it. That 00:07:10.329 --> 00:07:12.810 makes sense. Okay, next up, let's talk about 00:07:12.810 --> 00:07:15.550 instant utility. Getting real -time objective 00:07:15.550 --> 00:07:18.069 feedback. This solves that problem we all have, 00:07:18.209 --> 00:07:19.790 that frustration when you're working on a sales 00:07:19.790 --> 00:07:21.990 page or a spreadsheet, and you just wish you 00:07:21.990 --> 00:07:23.730 had an expert looking over your shoulder. This 00:07:23.730 --> 00:07:27.250 is Gemini Live with screen sharing. Yep. In the 00:07:27.250 --> 00:07:30.209 standard chat mode, you just hit Live, then the 00:07:30.209 --> 00:07:33.310 screen sharing icon. The AI is now seeing your 00:07:33.310 --> 00:07:35.529 screen in real time. And the powerful use case 00:07:35.529 --> 00:07:38.189 here is defining a role for the AI. Absolutely. 00:07:38.449 --> 00:07:40.569 You share your sales page and you tell the AI 00:07:40.569 --> 00:07:43.910 your role is a conversion rate optimization specialist. 00:07:44.269 --> 00:07:47.610 A CRO specialist. So it's focused on maximizing 00:07:47.610 --> 00:07:50.189 sales. and then you just ask for improvements. 00:07:50.430 --> 00:07:52.269 And this is so much better than just uploading 00:07:52.269 --> 00:07:54.790 a screenshot. Oh, way better. Right. Because 00:07:54.790 --> 00:07:56.550 it's live, you can click through different pages, 00:07:56.709 --> 00:07:59.649 you can navigate your site, and the AI is analyzing 00:07:59.649 --> 00:08:01.550 everything as you do it. It's a real back and 00:08:01.550 --> 00:08:03.769 forth conversation. You get actionable suggestions 00:08:03.769 --> 00:08:06.269 based on that expert persona you gave it. You 00:08:06.269 --> 00:08:08.550 get that expert eye instantly. Is giving the 00:08:08.550 --> 00:08:11.069 AI that specific role, like a CRO specialist, 00:08:11.290 --> 00:08:13.970 really necessary to get valuable feedback? Oh, 00:08:14.089 --> 00:08:17.180 yes. Defining a role sharpens the AI's focus. 00:08:17.660 --> 00:08:20.259 It makes sure the analysis is targeted, actionable, 00:08:20.759 --> 00:08:23.019 and not just generic advice. Okay, finally, let's 00:08:23.019 --> 00:08:24.860 cover custom images. This is where it all comes 00:08:24.860 --> 00:08:27.360 together. AI Studio gives you access to the full 00:08:27.360 --> 00:08:29.529 suite of their models in one place. You've got 00:08:29.529 --> 00:08:31.769 Nano Banana for editing, Image in 4 for high 00:08:31.769 --> 00:08:35.190 quality creation. And Image in 4 Ultra for that 00:08:35.190 --> 00:08:37.350 professional commercial grade output. So this 00:08:37.350 --> 00:08:40.490 means you can create your own custom stock photos 00:08:40.490 --> 00:08:44.450 that are 100 % unique to you. You control everything. 00:08:44.730 --> 00:08:47.730 The aspect ratio, the resolution, the specific 00:08:47.730 --> 00:08:50.289 details, like a diverse group of three colleagues 00:08:50.289 --> 00:08:53.710 laughing in a modern, brightly lit office. But 00:08:53.710 --> 00:08:55.929 here's the ultimate clever strategy we found. 00:08:56.190 --> 00:08:58.929 the smart YouTube banner trick. You don't start 00:08:58.929 --> 00:09:01.450 in the images tab. You go back to the regular 00:09:01.450 --> 00:09:04.750 Gemini chat, paste in your YouTube channel's 00:09:04.750 --> 00:09:07.570 URL, and just ask Gemini to analyze it for its 00:09:07.570 --> 00:09:10.269 style and brand colors. That is brilliant. You 00:09:10.269 --> 00:09:12.690 know, I still wrestle with prompt drift myself, 00:09:12.850 --> 00:09:15.289 especially trying to balance realism and artistic 00:09:15.289 --> 00:09:18.220 vision in image prompts. Right. letting the AI 00:09:18.220 --> 00:09:20.159 write the perfect prompt for you based on your 00:09:20.159 --> 00:09:22.840 own brand. That's a great hack. Gemini will spit 00:09:22.840 --> 00:09:24.860 out this hyper -specific prompt that's already 00:09:24.860 --> 00:09:27.000 optimized for Image in 4 Ultra using your brand 00:09:27.000 --> 00:09:29.019 colors and everything. You just take that prompt, 00:09:29.179 --> 00:09:31.759 go to the studio, set the aspect ratio to 16 00:09:31.759 --> 00:09:34.840 .9 for a banner. And you get a perfectly tailored 00:09:34.840 --> 00:09:37.889 image. And if it needs a little tweak... That's 00:09:37.889 --> 00:09:39.850 where NanoBanana comes in. Right, their post 00:09:39.850 --> 00:09:41.789 -editing tool. It's perfect for those little 00:09:41.789 --> 00:09:44.149 surgical edits. You download the image, upload 00:09:44.149 --> 00:09:46.309 it there, and you can do detailed adjustments, 00:09:46.490 --> 00:09:48.909 remove an object, or get rid of the background. 00:09:49.389 --> 00:09:52.330 Besides just removing a background, what is NanoBanana's 00:09:52.330 --> 00:09:55.210 core strength for business use? It's specifically 00:09:55.210 --> 00:09:58.129 for that detailed editing. For fixing the small 00:09:58.129 --> 00:10:00.669 parts of an AI -generated image after it's been 00:10:00.669 --> 00:10:04.070 created, Imogen makes the block. NanoBanana does 00:10:04.070 --> 00:10:06.240 the fine carving. This has been a really deep 00:10:06.240 --> 00:10:08.419 dive. Let's do a rapid -fire summary for everyone 00:10:08.419 --> 00:10:11.000 listening. Okay, let's do it. First, use that 00:10:11.000 --> 00:10:14.179 two -step prompt process GeminiChat first. Then 00:10:14.179 --> 00:10:17.480 Vibe Code to build custom apps in minutes. Then 00:10:17.480 --> 00:10:19.919 master the temperature setting for audio. Start 00:10:19.919 --> 00:10:23.000 around 1 .2 to get that natural, expressive sound 00:10:23.000 --> 00:10:25.460 for your content. For video, leverage the free 00:10:25.460 --> 00:10:27.879 VO2 for all your testing and prompt perfection. 00:10:28.220 --> 00:10:30.720 Then move to the higher quality V3 .1 for your 00:10:30.720 --> 00:10:34.080 final exports. Get instant, expert CRO feedback 00:10:34.080 --> 00:10:37.039 on your live website by assigning the AI a specific 00:10:37.039 --> 00:10:40.000 role during screen sharing. And finally, generate 00:10:40.000 --> 00:10:43.200 custom stock images using that Smart Gemini then 00:10:43.200 --> 00:10:45.740 image and process. Let it write the prompt for 00:10:45.740 --> 00:10:48.240 you, then use NanoBanana for any final edits. 00:10:48.559 --> 00:10:51.460 All of these features save hours, they save money, 00:10:51.679 --> 00:10:53.860 and they don't require any advanced technical 00:10:53.860 --> 00:10:56.559 knowledge. It is truly remarkable what's being 00:10:56.559 --> 00:10:59.080 offered for free right now. So what's the big 00:10:59.080 --> 00:11:01.850 idea here? What does this all mean? The key to 00:11:01.850 --> 00:11:05.129 mastering this entire free suite, it really all 00:11:05.129 --> 00:11:07.230 comes down to the prompt. Right. The biggest 00:11:07.230 --> 00:11:09.470 limitation you're going to face isn't the technology. 00:11:09.909 --> 00:11:12.230 It's your ability to clearly articulate what 00:11:12.230 --> 00:11:14.850 you want to the machine. The tools are there. 00:11:15.090 --> 00:11:17.350 You just have to focus on being a great prompt 00:11:17.350 --> 00:11:21.230 engineer. Go to istudio .google .com right now. 00:11:21.309 --> 00:11:23.610 Just pick one of these use cases and focus on 00:11:23.610 --> 00:11:25.809 mastering it today. We think you'll be surprised 00:11:25.809 --> 00:11:27.590 at what you can create. Thanks for joining us 00:11:27.590 --> 00:11:29.330 for the deep dive. We'll catch on the next one.