Why more tools cost you more time

Harvard researchers tracking AI tool use found that productivity peaks at around three tools and drops with every one added after that. The reason comes down to something called "attention residue" — when you switch from one tool or task to another, part of your focus stays behind on the one you just left, and it slows down how well you perform the new one until the earlier task feels properly finished.

Add the subscription cost, the mental overhead of remembering which app does what, and the time spent comparing tools instead of using them, and the time savings AI is supposed to deliver quietly disappear into the switching itself.

The rule

Pick exactly three roles and assign one tool to each:

One for writing and editing — emails, reports, documents, first drafts.

One for research and quick lookups — answering questions, summarising articles, checking facts.

One for spreadsheets and numbers — building trackers, cleaning data, writing formulas.

The goal isn't finding the objectively "best" tool for each job — it's cutting the number of tools you have to remember, compare, and pay for down to a set small enough to actually master.

How to choose your three

Look honestly at your last two weeks. Whichever tool you opened without thinking about it for a given task is your real default for that role — keep it. For the other two roles, favour whatever you already have through work rather than adding something new: Copilot if your workplace runs Microsoft 365, Gemini if it runs Google Workspace. Paying twice for overlapping capability rarely earns its keep.

Once you've picked your three, stop opening everything else for 30 days. If you genuinely miss one, bring it back deliberately. Most people don't.

What to do with the tools you cut

Don't just let a subscription lapse quietly and forget why you signed up in the first place — that's how the same tool sneaks back in six months later. Write one line for yourself on why each one didn't make the cut: "overlapped entirely with [tool]," "never actually opened it after the first week," "good at [X] but I don't do enough of that to justify it." Next time a shiny new AI app catches your attention, that note is a faster gut-check than trying the whole thing from scratch again.

The same logic applies at a team level if you're the one signing off on software spend. A quick audit of who's actually using what, versus what's just sitting on the company card, tends to turn up more waste than people expect — subscription fatigue isn't just a personal productivity problem, it's a real, recoverable cost.

List every AI tool or app I currently have a login for or pay a subscription to: [list them] For each one, tell me: - What I actually use it for (be specific, not "general AI stuff") - Whether that task overlaps with something another tool on this list already does - Which three tools I should keep as my defaults for writing, research, and spreadsheets — and which ones I could cancel or stop opening this month
💡 This week's action

Pick one AI subscription or app you haven't opened in the last fortnight and cancel it, or set a calendar reminder to cancel it before the next billing date. You'll likely feel nothing except a little relief — and one less thing competing for your attention every time you sit down to work.

A note on trial periods

Free trials are the easiest way this creeps up on you — a tool gets added "just to try it," the trial ends, the card gets charged automatically, and three months later it's still sitting there unused. When you do decide to try something new going forward, set the cancellation reminder on the day you sign up, not the day you remember you should. It costs nothing and saves you from exactly the situation this guide is helping you get out of.