My AI Productivity Stack —
the 5 tools I use every single day
Tested 60+ AI tools over two years. These are the only five still open when I sit down to work.
There’s a certain kind of productivity content that lists every AI tool ever made and calls it a “stack.” This isn’t that. What follows is genuinely what I reach for first — tools that have quietly reshaped how I write, think, plan, and build.
The AI tool landscape has matured fast, especially heading into 2026. Models are sharper, interfaces are smoother, and the real question isn’t “what can AI do?” anymore — it’s “which of these things is actually worth my time?” That’s the filter this list was built through.
No affiliate links. No sponsored mentions. Just an honest accounting of what’s earned a permanent spot in my day.
I use Claude as my primary reasoning companion — not just for writing, but for genuine thinking. When I’m stuck on a strategy decision or trying to untangle a complex argument, I treat it less like a search engine and more like a well-read colleague who happens to be available at 11 pm.
What I’ve found different about Claude compared to other frontier models is how it handles nuance. It pushes back when my reasoning is shaky, offers alternative framings without being sycophantic, and gives me its honest read on a problem. The extended context window means I can paste an entire document and work through it without losing the thread.
The Projects feature has become central to how I work — I keep persistent context for each client and long-running project, so every conversation picks up right where the last one left off. It’s cut a lot of the “re-explain everything” overhead that used to eat my mornings alive.
I’ve been a Notion user for years, but the AI layer transformed it from a filing cabinet into a system that works with me. The updated AI can now reason across your entire workspace — not just the page you happen to have open.
My daily ritual: open my inbox, ask AI to surface the three most relevant items based on today’s plans, then let it update my meeting notes while I’m still in the meeting. The AI-generated summaries are reliable enough that I’ve nearly stopped taking manual notes altogether.
Where Notion AI really earns its place is embedded intelligence. You’re not switching apps to get a smart response — it’s available inline, inside the same space where your knowledge already lives. That friction reduction is more valuable than it sounds.
I’m not a full-time engineer, but I write enough code — scripts, automations, data pipelines, the occasional web build — that my editor setup matters. Cursor changed my relationship with code in a way that traditional AI autocomplete plugins never managed.
The difference is project-wide understanding. Ask it to “add error handling to every API call in this codebase” and it actually does it — across files, consistently, without you wiring it together. The agent mode handles vague instructions like “make this faster” by profiling, identifying bottlenecks, and implementing fixes in one coherent pass.
If you write any amount of code and haven’t switched yet, this is your sign. The free tier is genuinely useful; the paid plan is worth it once you hit the daily limits.
My Google usage dropped by roughly 70% since I started using Perplexity seriously. It synthesizes live web sources and gives you cited, up-to-date answers — not a chatbot layered over a search index, but a genuine research engine.
The Spaces feature lets me create persistent research environments — I keep dedicated Spaces for competitive analysis, industry trends, and ongoing client research. Each one remembers context, preferred sources, and my working hypotheses across sessions.
Where Perplexity beats general-purpose AI is recency and sourcing. When I need to know what happened in a market last week, or verify a statistic before it goes into a client deck, Perplexity is my fast-twitch research reflex. It handles that job better than anything else I’ve found.
This one feels unglamorous but its impact on my week is enormous. Otter.ai joins every meeting, transcribes in real time, and its AI layer extracts action items, key decisions, and follow-ups automatically — without me lifting a finger after the call ends.
The chat interface built on top of transcripts is genuinely useful. I can ask “what did Sarah say she’d send me by Friday?” or “was there agreement on the pricing structure?” and get an accurate, sourced answer. No re-listening to hour-long recordings.
What I didn’t expect was how it changed me as a participant. When you know transcription is handled, you stop furiously typing notes and actually listen. The conversations themselves have gotten more productive — a second-order benefit I didn’t anticipate when I first signed up.
The full stack at a glance
What makes this stack work isn’t any one tool — it’s how the five of them cover different parts of the day without overlapping awkwardly. Claude handles deep thinking. Notion handles knowledge. Cursor handles building. Perplexity handles research. Otter handles the relentless calendar. Together they form something that genuinely feels like leverage.
The honest truth about AI productivity tools is that most of the real gains come from changing how you work — not just bolting a tool onto the same old process. Each of these five pushed me to rethink a workflow from scratch. That’s why they stuck. Everything else I tested was a feature, not a foundation.
Your stack will probably look different. If you’re a designer, Figma AI might be where you live. If you’re all-in on one super-app, that’s valid too. But if you’re building from zero, or questioning what you’ve already got, this is a solid place to start. Pick one. Go deep on it. Only add the next tool when you’ve genuinely hit a ceiling the first one can’t raise.
