I pay for Gemini, ChatGPT, and Claude, and there’s a clear winner

There are plenty of things I genuinely don’t enjoy doing — repetitive tasks, setting reminders, organizing endless to-do lists, and all those chores that eat into the day. That’s where AI tools like ChatGPT, Gemini, and Claude have slowly slipped into my routine. They handle the grunt work for me, which means I can spend more time on things I actually care about.

Paying for all three premium versions of these AI apps was great for unlocking their full capabilities, and for a while, I was actively trying to use each of them properly. But somewhere along the way, I noticed a pattern. Despite having access to all three, I rarely stuck with ChatGPT or Gemini for long stretches. Instead, I’d open them, start a task, maybe experiment for a bit, and then somehow drift back to Claude.

After months of noticing this habit and trying to understand why it kept happening, I think I finally figured it out.

When I use an AI tool, I don’t want to constantly babysit it. I don’t want to rewrite prompts five different ways or explain myself repeatedly. I just want the tool to understand what I mean and deliver something useful the first, or maybe the second time around. That’s exactly where Claude has worked best for me.

Whenever I give Claude a prompt, there’s this sense of reliability to the response — the replies are usually clear, easy to understand, and close to what I actually want without forcing me into an endless cycle of corrections. Over time, it has also begun to understand how I think and what I am really asking for, which makes the whole experience feel much more human.

But prompts are just one part of it. The bigger reason I keep going back to Claude is Cowork. In my use case, it feels like having a digital employee running in the background. I’ve set Cowork to handle all the repetitive tasks I constantly forget or simply don’t want to waste mental energy on. For instance, it sends daily reminders in a work group chat twice a day, removes duplicate files from my MacBook, renames messy, unnamed files, and manages automation tasks that would otherwise keep piling up throughout the week.

The reminder feature alone has made a ridiculous difference in my daily routine. Imagine having to remember the exact same reminder twice a day, every single day of the week. With my forgetful personality, I probably would have set reminders for the reminders themselves. Now, I don’t even think about it. Claude Cowork takes care of everything in the background, and the task is usually done before it even crosses my mind.

What makes it even more interesting is the Dispatch feature. If I’m away from home and need something done on my computer, I can simply use my phone to communicate with Claude. The whole thing feels like carrying a walkie-talkie connected to my desktop. I just sent a message from my phone, and Claude started handling tasks on my computer remotely. So if I suddenly need Excel opened while I’m out, I can dispatch the task from my phone, and it’s already waiting for me by the time I get back.

That’s the kind of experience I actually want from AI. I don’t want a tool that adds extra steps; I want something that removes friction from my day and takes repetitive tasks off my plate. I’ve always believed more in smart work than endlessly grinding through repetitive tasks, and that’s exactly why Claude fits so naturally into my workflow.

Claude Code is very interesting, even though coding is not a major part of my workflow. The entire idea behind it follows the same philosophy: reducing the amount of manual work you need to do yourself. Code can read an entire codebase, make changes across files, run tests, and even deliver committed code. For people who are not engineers but still want to build things, that shift feels pretty massive. And honestly, the idea of simply describing what you want and watching AI handle large parts of the technical work still feels slightly surreal to me.

I can’t fully speak about the real-world Claude Code experience yet because I genuinely haven’t spent enough time using it myself. But I have experimented quite a bit with Claude’s Artifacts feature, and that alone gave me a glimpse of the larger idea.

What I immediately liked about Artifacts is that it doesn’t throw you into some complicated technical setup right away. Instead, the experience starts with a conversation. I pick the category I want to create — whether that’s an app, a website, a productivity tool, a quiz, or even some random experimental idea — and Claude begins figuring things out alongside me instead of expecting me to arrive with a perfectly structured brief.

This is probably the biggest difference I noticed while using Claude. Most AI tools tend to run in the wrong direction after misunderstanding something halfway through a task. And instead of stopping to clarify, they just keep going and hand you a result that technically exists, but has absolutely nothing to do with what you actually wanted.

If something is unclear or an issue arises during the build, Claude pauses and asks follow-up questions. At first, I didn’t think much of it, but over time, I realized how much smoother that makes the overall experience. Instead of forcing me to restart everything from scratch later, it tries to understand the task properly as it works on it. And that’s really why I keep enjoying the experience. It’s like discussing an idea with a human, clarifying things along the way, and eventually getting back something that actually resembles what I had in mind.

Now, to be clear, I’m not saying ChatGPT and Gemini are bad AI tools. They’re genuinely good at a lot of things, like creating images, streamlining workflows, and sorting everything out. But personally, I’ve often found myself getting frustrated while using them, especially Gemini, because it sometimes feels like the tool isn’t fully understanding what I’m trying to say. I end up explaining the same thing repeatedly, rewording prompts, and adding extra context, and by the end, I’m somehow more mentally exhausted than when I started.

For the longest time, I assumed the problem was me. I thought maybe my prompts just weren’t detailed enough. But then I started using the exact same prompts across different AI tools, and that’s where the difference became obvious to me. ChatGPT handled them better, but Claude consistently came the closest to understanding what I actually meant.

That’s exactly why I keep finding myself going back to Claude, even while paying for ChatGPT and Gemini alongside it. I want an AI tool that genuinely helps me get things done, not one that leaves me mentally drained. And in my workflow, Claude has consistently felt like the tool that requires the least effort to use.

What makes it even more interesting is that it’s also a cheaper option for me right now. I pay around $17 per month for Claude Pro, while Google AI Pro and ChatGPT Plus are around $20 per month. The price difference itself isn’t massive, but the experience has been. When a tool saves me time, reduces friction, understands what I mean, and quietly handles repetitive work in the background, even a small pricing advantage starts feeling much bigger over time. I wasn’t after the cheapest option; I was after the one that made my life easier. But getting both in one subscription? I’m not complaining.