7 AI tools that's so good it feels like cheating.

I try a lot of AI tools, at least 10+ weekly, so if you do the maths, you know how much I've tried this year alone. The thing is, most of them don't stick, except for these 7 AI tools that I keep coming back to, not just for product design or UX design, but for work, life, and everything in between. These tools have literally changed how I operate day to day, and I think they might do the same for you too. My name is Samuel Fabayo, and in the last 9 years, I've designed software products for startups in Lagos, Nigeria, and a few high-growth companies in the UK, including Deliveroo and Monzo Bank. Let's jump right in. The first tool I use is Granola, and this has been like a really, really game-changing tool for me because basically it's a tool that just sits in any meetings or any conversations with me and helps me transcribe those meetings. But here's the thing that makes Granola quite special. Nobody can really know if I'm using Granola to transcribe a meeting, so it's not going to appear on the meeting like some kind of creepy bot, right? It just shows up, it's just there in the corner. I can turn it on, I can turn it off. It gives me a signal to know when it's actually recording a meeting, and then once I'm done, it gives me this really detailed summary of the meeting that always focuses on the most important things, including the action items and everything else. Granula is one of those tools that just feels truly magical. And one of the really great things about Granula is that I can just easily have a chat with that meeting afterwards to understand anything that wasn't clear in the meeting. A really good use case is if you're interviewing for like a new role, right? It is perfect for that because you can use it to sit in your meetings and transcribe all your interviews. You can put all those interviews into a specific folder so that you can eventually understand the kind of points that you make during interviews. You can spot the patterns based on how you answer questions, and that really helps you to understand how you perform in interviews. You can even ask questions on how you can improve and how you can answer questions better next time. Then you can move whatever discussions and transcripts you've had into ChatGPT to have like a deeper dive. So yeah, that's Granular. Second tool I use regularly is called Recraft. Some of the pull bars for the thumbnails of this video were generated in Recraft. The top 3 ones, you can see that it has some kind of shine. Those ones were generated in Recraft while the other ones were generated with ChatGPT-4o. Recraft is a really good tool for image generation. And unlike every other image generation tool like ChatGPT-4o where you can generate images in a chat, you're kind of limited to that chat interface in ChatGPT. But Recraft is specifically designed for generating images. It has like an infinite canvas like Figma where you can keep iterating and they have their own image generation model called the Recraft V3 as well. And the reason I find Recraft really helpful is that you get to choose from different templates or different styles for the kind of images that you want to generate. So if you're thinking about like an illustration, a particular style, or you're thinking about an image that's inspired by some vector illustration, cartoon character, pencil sketch, you can find hundreds of those different styles right there in Recraft. I think there are too many options and it could get overwhelming at times, but I guess people like a lot of options. You can also combine styles so that whenever you want to generate an image, you can get your own inspiration from another kind of image that you find online, and then you can blend those styles across different images in Recraft to get something unique. I think Recraft has the potential to be the Photoshop of the AI era. Tools number 3 and 4 is ChatGPT and Claude. Now, 3rd and 4th tools are usually suspect because I believe, I mean, everyone uses either of these tools in some way, shape, or form already, but I think I also use them in like some different ways, so I'd like to share that. ChatGPT and Claude are like brothers in arms for me, even though OpenAI and Anthropic, their parent companies, are fierce competitors. I basically say that ChatGPT is for life while Claude is for work. So I use ChatGPT for— I mean, I could almost say I use it for anything because it basically runs my life. Because I've realized that the more context that ChatGPT has about me, the better it gets and the more value that it has to me, which is why having a model that runs locally is very important. I think the way I use it is quite unique because whenever I know that there's something that I'm working on that I'm going to run continuously, I immediately understand that this is going to be a project. So I run a project for that. And that helps me because whenever I run a project through ChatGPT like that, I immediately start having conversations with that project from the beginning. I'm not dumping any prompts in just yet. I just want it to have a lot of context around the project I'm working on. And then I've realized that the more conversations I have with it, the easier it is for it to get context and the better results I have with the conversations that I have later. So for example, let's say I'm trying to get an apartment or get a new place. I've already set up a project around that where I'm going to say, you're going to be the one that will help me get a new place. You're going to help me research places, help me understand things like cost of rent, crime rates, and transportation schedules around the location that I'm going to get a place in. And then I continue to have conversations with the project whenever I want to do anything around searching for an apartment. I also have another way that I use ChatGPT whenever I'm reading something, especially books that feel complex to understand, for example, books about philosophy or just some complex books, ChatGPT is like my explainer. So when I find something confusing, I take screenshots or snap the page on the book and then have ChatGPT discuss that book with me. Even when I'm done, I still have conversations with ChatGPT about the book, like having a buddy that you just chat to about something you've read so you can remember what you've read. Claude, on the other hand, I use for everyday work. I think Claude is really good at writing. ChatGPT's writing is so bad. So I don't even bother with it. But I use Claude for writing. It really helps me with analysis. It's really good at coding. So I use it for more work-related stuff. Whenever I want to write something that's really sharp, maybe I'm working on a design and I want to write a copy for the design, Claude really helps with this. And I just like that because just like ChatGPT, I use the Claude projects feature in a similar manner. When I have a project that I'm working on, I make sure that I maintain context within that project and I have conversations that are running within that project. Also, because these AI tools hallucinate a lot and they can comfortably or confidently mislead you, I use them to check or verify each other. So when I get a response from ChatGPT, I put the same prompt into Claude and I see what Claude generates, and I sort of compare the 2. Of course, you need the paid plan for this to be able to work well. So the next tool I'm going to be talking about is V0, which is like my first draft tool. I'm a frontend engineer. I usually use V0 to get off the blank screen state whenever I have a blank screen. For example, if I want to work on a design or a piece of UI and I've been thinking or brainstorming about it, I can go back to Claude or go into a project where I've described the context of the thing that I'm trying to work on. And then once it has enough context, I can ask it to help me write the prompt that I would then put into v0. And then v0 helps to generate iterations of what the frontend might look like. But the thing is, I don't really use v0 to create the entire thing or an entire project. Sometimes it helps with MVPs, but when it comes to a finished project, I don't use it for that. I just use it to get beyond that first draft phase, like that blank screen phase. Because once I do that, V0 sometimes surprises me with something I wasn't even thinking about. It usually does something that I wasn't expecting for the kind of design I'm trying to come up with. So it sort of puts that together. Once I have that, it helps me get beyond the blank screen problem. And the thing I've realized with AI tools in general is they are good in-betweens. They're not good at the edges of creation. And when I say that, they are not good in the input and the output stage. They're not good at understanding input properly if you don't provide the proper input to them. And they're also not good at providing like 100% output. You need to do a lot of verification and it's so hard to get something to like from like an 80% to like 100%. But as these AI tools improve, they will get better at the input and at the output stage. And right now I believe they're really good in between. So that is what I use them for. And when you think about like ideas or iterations, that's what really happens in the middle of the creative process. And that's why I use V0 for things like this. So once I get that iteration out, then I get a lot of ideas I can then move back into Figma or Framer to start to refine further. I think the future of UI design tooling is going to be a tool that allows you to do everything in the same place. Like you can write prompts to generate something. You can also iterate further on that thing with the traditional design tools, like making small changes because I find it very difficult to start writing prompts to make changes to really tiny details. So I think the tool of the future that would really excel is the one that really allows you to move between writing prompts and traditional tooling like changing things and making changes to components. Another vibe coding tool that I use regularly is Cursor. I really like Cursor because it gives me more control than V0 does. The limitation of vibe coding tools like V0 or Lovable is that they use specific libraries to generate their UI, so there's a cap on what you can achieve if you want to do something that's really different. Cursor doesn't have that problem, but you need to have a good idea of what you want to achieve when vibecoding with Cursor, or else it's just going to give you barebones UI. A tool that I built with Cursor that I currently use every day is a Chrome extension that helps me manage my tabs. I currently have about 500 tabs open on my iPhone browser. That's just me. I usually have a lot of tabs open. So I built this tool because I want to manage my tabs because sometimes I could have like 20 tabs open on my Chrome browser and then usually gets confusing for me. This extension helps me set a limit to the number of tabs that I want open and then I can set a maximum number of tabs and if I try to open 11 tabs for example, let's say I've set 10 tabs as the maximum, the extension blocks it and asks me to close an existing tab or increase my limit. I have zero coding knowledge, I don't know how to code but I was able to build this with quick Figma design on Cursor, and it took me about 12 hours in 2 weekends. I've also been using Cursor to learn how to code, so if it generates some snippets of code, I ask it to explain that code for me line by line, and I can even go into the code to make small changes that I want to make. I think every designer should learn how to use live coding tools like Cursor, V0, Lovable, or Polymer. Soon these tools won't be optional for designers, they'll be strongly required, and I think this is happening already. I saved the best for last because tool number 7 is an amazing AI browser called Dia. Dia was released about 2 months ago, so it's still very new, and it's just one of those tools, just like Granola, that once you start using, you really can never stop using because you're just like, wow, this is so amazing. So what does Dia do? Dia is a browser that helps me to open up my tabs, and then on the side, I have like a side interface where I can chat with the open tabs. But the really good thing about it is that every web page that I have open becomes context for Dia. So that means I can do things where I no longer have to be searching through websites details. Remember I told you that input and context is the problem with AI tools, and DIA solves this. I can literally just ask DIA Assistant on the side about everything that's going on on a website or whatever information that I'm looking for. One of the great things about DIA is that it's so beautifully designed, and as a designer, that's something that I value so much. Having a tool that hits such a high bar for the quality of the experience at the beta release is so amazing. Another amazing thing about DIA is being able to add skills, which makes my usage so much faster. So for example, my skills are going to be like the way I like information to be delivered to me, the way I like to write, or the way I like DIA to understand context around what I want to do. So if I have a prompt around how I want DIA to be able to summarize a web page for me or summarize a YouTube video, I don't have to write a new prompt every time. I can create a skill for that and then just type something like /summarize, and DIA gets that you and just summarizes everything for me. I'm sure there are so many other use cases for the browser that I haven't even gotten into yet, but it's been something that I've been using consistently for the last couple of weeks now, and the design is just so amazing, like I already said. Plus, they are still in beta, so it's only going to get better from here. Aside from V0 and Cursor, which I mentioned earlier, I've tried 6 other AI design and coding tools to find out which one delivers the best results. Watch this video if you want to see more AI tools in action, and I'll see you there.