The Experienced Professional's AI Stack

What this is for

You've watched a hundred "must-have AI tools" lists and none of them tell you what to actually cut. This is the opposite of a list. It's a stack: one primary tool per job, one alternate, and a clear ranking of what to pay for first. Use it to stop evaluating tools and start using three of them well.

Before you start

One question changes every recommendation below: does your employer already pay for AI? If you're in a Microsoft shop, Microsoft 365 Copilot may already be sitting in Outlook, Word, and Teams. If you're on Google Workspace, Gemini is already in Gmail and Docs. Check before you spend a dollar. The licensed tool your IT department approved beats a better tool you have to smuggle in, for one reason that outranks features: it's covered by your company's data agreement, and your pasted client material is not training someone else's model.

The stack, one tool per job

Job	Primary	Alternate
First-draft writing	ChatGPT (Plus)	Claude
Research with sources	Perplexity	ChatGPT (web search)
Spreadsheets and data	ChatGPT (data analysis)	Claude with file upload
Slides	Gamma	Microsoft 365 Copilot in PowerPoint
Meeting notes	Fathom	Otter.ai or Fireflies
Images for decks and docs	ChatGPT or Gemini (built-in image generation)	Midjourney, if you're doing this at volume
Building small tools/sites	Claude (Artifacts)	Replit
Learning something new	ChatGPT (voice mode)	Claude
Email	Gemini in Gmail or Copilot in Outlook	Superhuman
Voice/dictation	Wispr Flow	Apple/Windows built-in dictation

Why these and not others: Perplexity wins research because it cites sources by default. You can check its work, which matters more than raw fluency. Claude wins building and long-document work because it handles bigger context and writes cleaner code without hand-holding. ChatGPT wins general drafting because its voice mode is still the most natural daily interface. For email and images, the honest 2026 answer is that the AI built into what you already pay for (Gmail, Outlook, your chat subscription) has caught up to the standalone specialists for normal professional use. Gamma beats fighting with PowerPoint templates if you don't have Copilot; if you do, try Copilot first, it's already licensed. Fathom, Otter, and Fireflies are close enough that whichever your team already uses is the right answer; Fathom is the pick here because its free tier covers a solo professional well.

The free-tier-only stack

If you're not paying for anything yet: ChatGPT free tier for drafting and simple questions, Perplexity free tier for anything you need sourced, Gemini (built into Docs and Gmail if you have Workspace) for email and doc help, and your phone's built-in dictation for voice. That covers 80% of daily use with zero subscriptions. The gap you'll feel: slower responses, tighter usage caps, and weaker reasoning on complex multi-step tasks.

If you have $60/month, pay for these first

1. ChatGPT Plus ($20): the widest daily utility, first dollar spent.
2. Claude Pro ($20): pay for this once you're doing real writing or building work; the difference in quality on long documents is the whole reason.
3. Perplexity Pro ($20): pay for this once you're making decisions based on research, not just curiosity. The sourcing is worth it the first time it saves you from repeating a bad stat in a client meeting.

Two notes on price. First, every one of these has a tier above it (ChatGPT Pro, Claude Max, Perplexity Max) at 5 to 10 times the cost; those are for heavy daily builders, and you'll know if that's you. You will not accidentally need one. Second, annual billing knocks roughly 15 to 20% off each of these; switch once you've kept a tool for two consecutive quarters.

Everything else (Gamma, Otter, Midjourney) is worth paying for only once the free tier of that specific job is actively costing you time. Not before.

Overhyped, not worth a subscription yet

• AI "agents" that promise to run your whole workflow unattended. They're genuinely better than a year ago at bounded tasks (research runs, multi-step document work), but "unattended" still means "checked afterward," and you're paying premium prices to be a supervisor.
• Personal AI avatars and video clones. Novel once, rarely used twice, and the quality-to-cost ratio is bad for anyone who isn't doing this at scale.
• Note-taking apps that layer AI on top of AI-generated meeting notes. You already have a transcript tool; a second app summarizing the summary adds cost, not clarity.

Run this on your own stack

Paste this into any of the chat tools and answer its questions honestly:

ROLE: You are a pragmatic technology advisor for an experienced professional.
CONTEXT: I'm going to list every AI tool I currently pay for or use weekly,
with the rough monthly cost and what I actually use it for.
MY LIST: [TOOL, COST, WHAT YOU USE IT FOR, e.g. ChatGPT Plus, $20, drafting
and research]
TASK: For each tool, tell me: keep, cancel, or downgrade to free tier. Flag
any two tools doing the same job. Flag any job I described that has no tool.
Then give me the three-tool version of my stack and the total monthly cost.
CONSTRAINTS: Be blunt. Optimize for what I use weekly, not what sounds
impressive. Do not recommend any new tool unless a job on my list is
uncovered.

Where this goes wrong

• Subscribing on launch week. People subscribe to a new tool the week it launches, then never cancel the old one. Run the quarterly check below instead of reacting to launch hype.
• Picking by feature list instead of by job. A tool with more features that you don't use is worse than a simpler one you actually open daily.
• Assuming today's leader stays the leader. It won't. This category moves fast enough that the best research tool eighteen months ago is not the best one now, which is also why this page carries a verified date.

The quarterly re-evaluation rule

Once a quarter, pick your three heaviest-use jobs (probably drafting, research, and one more specific to your role) and ask: is my current primary tool still clearly ahead, or just familiar? If you can't name one concrete thing it does better than the alternate, you're paying for habit. Fifteen minutes, four times a year. That's the whole system. Do not re-evaluate every time a new tool launches; launches are marketing events, not evidence.

The fastest version of that quarterly check is the prompt above, re-run with your current list. And if you want the modifiers that make every one of these tools sharper, the 12 Prompt Shortcuts list is the companion to this stack.

The 2-minute version

Check what your employer already licenses before buying anything. Pick one tool per job from the table above. If money's tight, run the free-tier stack. If you have $60, get ChatGPT Plus, then Claude Pro, then Perplexity Pro, in that order. Recheck your three main tools once a quarter, not every launch week.

---

Want to know where you actually stand with AI? The free AI Readiness Assessment takes 10 minutes and gives you a personalized roadmap instead of another tool list.

Weekly, on Tuesdays. The Gen Xcelerator: practical AI intelligence for people who already know how to do the job. Subscribe free.

Experience Is the API. GenXcelerate