AI Tools I Actually Use Every Day

There’s a new AI tool launched every 15 minutes, or at least it feels that way. After months of testing everything I could get my hands on, I’ve landed on a small set of tools that I actually reach for daily — not because they’re the flashiest, but because they save me real time on real work.

The Ones That Stuck

Claude is my go-to for anything that requires reasoning or writing. Whether I’m drafting strategy docs at work, debugging code for this site, or just thinking through a complex problem, Claude handles nuance better than anything else I’ve tried. The coding ability in particular is a step above — it understands context across files, catches edge cases, and actually explains its reasoning.

Cursor turned me from someone who dabbled in code to someone who builds things. The inline AI suggestions feel less like autocomplete and more like pair programming. When I’m stuck, I can describe what I want in plain English and get working code back. It’s the tool that made this portfolio site possible.

ChatGPT still has its place for me, mainly for quick lookups and brainstorming. The mobile app is great for voice conversations when I’m walking the dog and want to think through an idea out loud. The plugin ecosystem is also handy for one-off tasks.

What Surprised Me

The biggest unlock wasn’t any single tool — it was learning to chain them together. I’ll use Claude to architect a feature, Cursor to implement it, and then go back to Claude to review and refactor. Each tool has a sweet spot, and knowing when to switch between them matters more than which one is “best.”

I also didn’t expect how much AI would change my approach to learning. Instead of spending hours reading documentation front to back, I can ask pointed questions and get explanations tailored to my level of understanding. It’s like having a patient tutor available 24/7.

What Didn’t Work

Not everything landed. I tried a handful of “AI-powered” project management tools that added more friction than they removed. Auto-generated meeting summaries sound great in theory, but I found myself spending as much time correcting them as I would’ve spent writing notes. And most AI writing assistants produce text that sounds like… well, an AI writing assistant.

The pattern I’ve noticed: AI works best when it augments something you already do, not when it tries to replace an entire workflow.

Where This Is All Going

What excites me most is that we’re still early. The tools I use today will probably feel primitive in a year. Agents that can actually execute multi-step tasks, models that understand your full work context, AI that gets better the more you use it — all of this is coming fast.

For now, I’m focused on building the muscle memory to work with these tools effectively. The people who learn to collaborate with AI — not just prompt it, but actually think alongside it — are going to have a massive advantage. And honestly, it’s just fun to build things that would have felt impossible a year ago.