New AI tools come out every day. I test a lot of them, but very few make the cut into my permanent stack. The ones that do tend to be tools that save the most time.
Saving three hours a day saves you over a thousand hours per year, and for a lot of people, AI is already able to do that. An exercise I recommend everyone do is over the course of like a week, write down all your most time-consuming tasks, then set aside a day to find the best AI workflow to assist or even fully automate those tasks. Many people will find that they're able to save multiple hours a day.
I'll go through a- big list of tools across different use cases to help us out. So a lot will apply to you, some won't. Let's get into it.
Google felt like they were behind in AI for a while, but have been dropping some amazing stuff recently. A lot of it is completely free. I'll cover a couple of the biggest time savers for me.
I'll start with Google AI Studio's real-time AI assistant. It can be used for a lot of things, and it's awesome in particular for learning new software. This is at aistudio.google.com. There's a lot of other tools in here, mostly for developers, but anyone can use this one right here that says stream real time. Your first time doing this, there will be some approvals to allow it to access your microphone and screen.
Now I'll click start recording, then select video source and pick the window I want, which will be this one with Adobe Premiere. I have Premiere Pro opened up. How can I remove this background?
Sure. You want to remove the blue background in the video. You can do this by using the Ultra Key effect in Premiere Pro. Would you like to know how to do that?
Yes, where's that? Okay, the Ultra Key effect can be found in the Effects panel. Do you know how to get to the Effects panel? Is that down here where it says Effects?
Yes, that's the one. Now, if you type Ultra Key in the search bar, you can find the effect. Do you know where the search bar is?
Yep, I see Ultra Key here. Great, now drag that effect onto the video clip in the timeline. Would you like me to guide you through the rest of the steps? Yeah, what's next? The background is still there.
Okay, now go to the Effect Controls panel at the top left of the screen. There, you should see the options for the Ultra Key effect. Click the eyedropper icon next to the Key Color option, and then click on the blue background in the program monitor. Okay, yeah, that removed it pretty well, but...
It looks a bit hazy still. Got it. It looks like it's a bit hazy. You can adjust the matte generation settings in the Ultra Key effect to refine the key.
Do you want me to help you with that? I just like drag them around. Looks like transparency is doing something.
What should I do here? Okay, it looks like the transparency setting is affecting the haze. Try adjusting the transparency, highlight, and shadow sliders. to refine the edges of the key. Do you want to keep experimenting with the sliders yourself, or would you like some specific guidance on how to adjust them?
They're looking alright, is there any other settings I should do? Alright, you can also try adjusting the matte cleanup settings like choke and soften to further refine the edges. Would you like me to explain what those do?
No, I think I've got it looking pretty good here. So now can I, have you teach me how to clean up and enhance my audio? Sure, I can help with that. In Premiere Pro, You can use the essential sound panel to clean up. You get the idea.
That's incredible. It answers almost instantly and you can interrupt and clarify when needed. An expert real-time tutor to help you learn anything has been one of the dreams in AI.
It is not all the way there yet. It will still make plenty of mistakes, but it will be right and save time quite a bit. And YouTube won't always have like the specific scenario you need, or sometimes you don't even know what you need to look up to achieve something. And on your phone, you can do the same thing, but using the camera. Maybe you have some homework you need a tutor for and are totally not just cheating.
You could have it walk you through a science experiment, help you set up furniture, understand documents, all sorts of stuff. It's like the vision feature in ChatGPT, but real time instead of uploading photos. Another one from them is Google Gemini Deep Research.
It's an agentic model that will take multiple steps to deeply research a topic. Here's an example. I actually just got a PRP injection in my shoulder a few days ago.
I did a ton of research into it beforehand, but here's how I could have done that in here. So I'll make sure deep research is selected from the dropdown, then add a prompt. Provide a detailed summary of platelet-rich plasma injections for shoulder-related conditions with a focus on tendinopathy.
Include potential risks and current clinical evidence. Highlight any recent advancements or controversies specific to shoulder treatments. Also compare the outcomes of PRP injections with corticosteroid treatments for musculoskeletal conditions.
So I'll send that. It thinks for a second, then it shows me its plan. It has eight steps it will walk through.
I can edit that if I'd like to, but I'll start it as is. Now it starts its research. It's researching 27 websites.
That seems like a lot, but it's actually on the low end for what it usually uses. This will be over a hundred websites a lot of the time, but I chose a pretty simple example. It's got some PubMed, what looks like some different clinics, a bunch of other journals. Looks like a solid list. I'll come back when it's done.
All right. That took around three minutes and here's what it came back with. This is an incredibly comprehensive report based on these findings with what it is, how it works, side effects, treatment protocols, risks, effectiveness, comparisons. There's a lot in here, which basically confirms the reasons why I got the treatment already.
But Deep Research is well-named, and I was able to tailor it specific to my situation with my shoulder rather than all PRP uses. This isn't for just random one-off questions that you might pop into ChatGPT or Google, but it is incredible for deep research. The use cases are endless.
This is part of Gemini Advanced, which comes with a lot of other features, but does cost $20 a month after a one-month free trial. There are, of course, endless ways to use ChatGPT to save time, so in the description will be a free ChatGPT resource bundle provided by HubSpot. There's a total of five PDFs that go in-depth on how you can utilize ChatGPT in your career, and it was just updated for 2025. There's one that walks through a process similar to how I mentioned to find your most time-consuming tasks AI can help with, but it's a much deeper and more step-by-step guide to do that. But my favorite in here is called Supercharge Your Workday with ChatGPT. It covers specific examples of how ChatGPT can be used in various industries.
Sales and marketing, customer support, time management and task organization, personal development and learning. It walks through step by step with different tips. And I really like the section titled 100 ways to try ChatGPT today.
It has 100 sample prompts you can use and modify. No matter what career you have, there's sure to be a bunch in there that apply. And keeping in line with this video, prompt templates like that are another way to save a ton of time. Again, that resource is completely free. Just use the link in the description to go download that.
I'm always happy to partner with HubSpot to provide free resources to the people that watch this channel. But back on the free end from Google is Notebook LM, which I've covered extensively before, but they have some new features since then. It's another personalized AI-powered research assistant. DeepSeek R1 came out the other day and blew everyone away.
So I uploaded a PDF of the research paper, a link to the GitHub, and a couple websites about it in here. You can also upload YouTube videos, or I'll even use PDFs of an entire book I'm reading sometimes. It can take in a lot of information.
But now I can have it summarize all the information, or even more helpful is I can ask a question, then it will give me an answer and cite exactly where that information came from in my uploads. I can also create a study guide that will give me a quiz with an answer key, essay questions, and a glossary of key terms. I could also create a briefing doc, FAQs, timeline, or the most unique option is the podcast feature. You can add prompts to customize and guide what they talk about if you'd like.
Then this will generate a full podcast all about the topic. With the new interactive mode, you can also jump into the podcast and ask a question. They'll answer that, or you can steer the conversation in any direction you want.
Yeah, I'm excited to dive in. And what's really interesting is how they're achieving this reasoning in AI. It's quite a departure from how things are normally done. Yeah, usually we think about AI being trained on, you know, these massive data sets, like feeding it, you know, the entire library of...
Congress or something, and that's how it learns. But that can be really expensive and time-consuming. And what happens when it encounters something outside of that data set, you know, something totally new, it's not going to be able to adapt very well.
Exactly. And that's one of the limitations that DeepSeek's team has been really focused on addressing. They've created this model called DeepSeek R10. And the really fascinating thing is it was trained using only reinforcement learning, no label data at all. Whoa, did someone just break the fourth wall?
Let's hear it. And they open sourced the whole thing, right? Can you speak about that for a minute?
Absolutely. And it's a really key part of what makes DeepSeek stand out. Yeah, it's not just the technology itself, but the way they're sharing it with everyone.
The fact that they've made their models and research open source is a big deal. It really sets them apart. It's kind of like they're saying, here's what we've done.
Now let's all build on it together. Those are the best options from Google that I wanted to cover. Now I'll jump through some other tools a lot faster. They're across.
different use cases, so some may or may not be related to what you need. With Gamma, just write a text prompt and it will generate an entire beautiful presentation with engaging text and relevant images. There are other tools that do this, but Gamma has been my favorite one.
You can also create documents and websites with Gamma, but I've primarily used it for presentations. Now, if you need a coding assistant, Cursor is the leading tool right now. You can build and deploy fully functioning apps without any coding knowledge. It's all just from prompting.
Then you go back and forth for any changes you need. I don't have time to go into that much here, and I'm definitely not an expert at this anyways. Riley Brown has a bunch of videos on building cool stuff with it. There's lots of channels with deeper dives into Cursor to go check out.
Complimentary to that is Wizard, which is ideal for creating UI mockups, wireframes, and designs quickly using AI. You can generate from text, turn a screenshot into an editable mockup, or a hand-drawn sketch into a digital design. Then through prompting, you can iterate quickly, add or modify components, and for a powerful combination, you can export the code snippets and pass them into Cursor. So Wizard can create the entire visual design quickly, then Cursor can handle the backend logic and fine-tune it all.
You can do this all without any deep technical skills. The ability to easily build out apps that you can customize exactly to your liking is incredible. Whether that's like... personalized app just for your own use, or to build for deploying publicly.
The entire SaaS world is changing rapidly with all these new tools. Napkin AI is amazing for creating graphics and visuals from text. They just came out with a big update that's faster and has a lot more designs. It's super easy to use and is still completely free as of the filming of this video. I'll do an example I used before so I can see how the new update compares.
You can upload your own information or I'll use AI to help with this one. Five things most likely to cause the AI apocalypse in order of likelihood. All right, there's a solid list. I'll start with just the first one. I'll click the lightning button.
And here's a bunch of visuals I can cycle through. A lot of these look awesome. Then if you find one you like, there's additional styles for each. So there's tons and tons to choose from.
And the colors are all fully customizable as well. I could also highlight all the text and do it for all five. I'll give that a try. Now there's a bunch more really solid visuals. And I think these are awesome to use in videos.
I can download one, then blur out parts and add some zooms. And now it's an amazing looking, really engaging graphic that I made in just a few seconds. Using an AI meeting assistant can be a huge time saver. There's a lot of options out there. I've tested a few.
My favorite was Otter. especially if you're looking for something user-friendly with a shorter learning curve. On the more advanced side, with a lot more features and advanced functionalities, I liked Fireflies. But these all work similarly, integrating into Zoom, Google Meet, and Microsoft Teams. It will transcribe your full meeting, then give you a summary.
Then you can search the text to play back the audio at adjustable speeds, then tap the words to jump to different spots. It can highlight key points to review later. It has automated slide capture it will insert into the notes. It's super helpful if you have a lot of meetings to be able to review them, but also to be able to stay focused and attentive during the meeting.
And there's further use cases depending on what you do. Like if you or your team do a lot of sales calls, you could ask to identify weak points in the process or to find common objections. If you need more advanced uses like that, I'd recommend Fireflies.
This next one is a big one if you have a YouTube channel. I have a Flux Laura trained on my face, and anytime I want an image of myself in any setting or pose, I can just prompt for it. I use this on Replicate, and I've taught this multiple times on this channel, so I'll link to a longer video if you need it, but it's really easy to do. Basically, you just need at least 10 images of yourself from different angles and expressions.
You upload those, then name it, and select a few dropdowns. It costs around $5 to train, which you only need to do once, then it's just a few cents after that any time you need to generate an image. You can easily make amazing, unique thumbnails without setting up your camera and lighting every time.
Awesome. Another research assistant that probably everyone watching this channel knows by now is Perplexity. I mainly use Perplexity because of its speed at finding things like stats, research, or quotes I may need for a video. Although I don't have the pro plan anymore, all I need is the free.
And I have actually started to move back to ChatGPT for a lot of this since their native search feature has gotten so much better, but now displays images and sources. It does a lot of what Perplexity has been really good at. I debated on still including it in this video, but for some use cases, you may find it to be the best tool. and they are starting to implement actions with Perplexity Assistant. You can ask it to play a specific song.
It can help book a restaurant. You can search with Vision, which is a nice upgrade. ChatGPT came out with operators as well. Overall, agents like this, even though there's a lot of hype around them, they're not there yet for anything practical.
I do think it's a good idea to mess around with them to get a feel for how AI interacts with websites on its own, just for learning's sake and to stay ahead of the curve. One day we may have a dashboard of agents we're overseeing and just Kind of watching for mistakes and places to intervene and nudge. But I think we're a ways out from that.
But sticking with LLMs, these are the type of AI tool people are generally the most familiar with. But it's also one of the biggest time savers for me personally, and I assume could be for most people too. This is an area where, like I mentioned at the beginning, if you write down your most time consuming tasks, there is very often a way that AI can automate or assist with those tasks.
And LLMs are the most common solution. The top two I use are Claude and ChatGPT. Claude is my go-to writing assistant. Although ChatGPT has definitely gotten better at creative writing, but I don't think it's as good as Claude yet.
There's other writing tools like Jasper, Lex, WriteSonic, Copy, and tons of others. If there's a specific type of writing you do a lot, you may want to find one that's tailored for that. They typically all use Claude or ChatGPT under the hood, but they just have solid prompt engineering and a nice UI on top of it. For me, I just do my own prompting and have never felt the need for a specialized writing tool.
Typically, I just get outlines and do the bulk of the actual tone and writing myself anyways. With ChatGPT, using custom GPTs or... projects to streamline a repetitive task is something just about everyone will be able to do.
The projects allow you to keep chats organized and contextually connected. It's great for long-term multi-step workflows. Here's a pretty easy one that I use for shorts.
So I have a big list of types of hooks and example hooks that I uploaded to the project files. Then I have step-by-step instructions as well. To start, all I have to do is input my topic and it will give me a list of hooks in each style.
So I can refine and pick one from there. Then it will move on to the next step of the workflow. That's a whole...
collaborative process that speeds up short form creation a lot. I do also have this set up in Claude. This isn't particularly creative writing, so both work really well.
Custom GPTs are great for very specific, fixed behaviors, and they also have the benefit of exploring GPTs others have made public. If you have something you're thinking of creating, see if someone else already has. There's tons of options in here. One example I use that I really like is I have GPTs for crafting prompts.
That's a huge time saver. This one's for videos. I just enter a simple prompt and it will elaborate into a much longer prompt that works better for AI video tools. And it gives multiple options.
That's really nice to have. I have another one for LLMs, basically training ChatGPT to prompt itself, and another for text-to-image prompts. And an often overlooked feature is you can summon your GPTs in any chat.
Just click the at sign and choose your GPT from the list, then ask for what you need. Then it will use all the context and instructions you've built into that GPT to tailor a response. But this is a coloring book page generator I made for my four-year-old. I'll send that.
And even though this is in an existing chat, without any additional context, it knows what to do and generates me a coloring book page. Again, I do recommend you actually create a list of your most time-consuming tasks and try to find ways AI can help. Hopefully this video will give you some ideas and this huge list of tools will be a good place to start.
I've got a huge update from Futurepedia. We partnered with Skill Leap AI. So we looked around and they had built the best AI learning platform I could find.
So we... partnered with them and have now built that platform into Futurepedia. There's over 20 amazing, extremely in-depth courses covering all different aspects of AI and how to use it in your own life and business.
It's the perfect place to learn a lot of the stuff I talked about in this video. It is also where I will be launching all of the courses I'm working on right now. So I'll have a link down in the description where you can go to try that out for free.
And thank you so much for watching. I'll see you in the next one.