Not a subscriber?
Join 10,000+ getting a unique perspective every Saturday on growing their internet business with actionable AI workflows, systems, and insights.
You're in! Check your email
Oops! Something went wrong while submitting the form đŸ€”
September 6, 2025

The end of AI copy-paste hell

I've worked with over 100 professionals trying to get AI integrated into their daily work. You know what I learned?

The AI you see hyped up on LinkedIn has almost nothing to do with the AI people actually use.

The hype is all about magical, one-click solutions. The reality is a soul-crushing loop of copy-pasting between browser tabs just to get a half-decent answer.

Let's be honest, AI adoption is a mess. It's hard for a reason:

  • The workflows are clunky.
  • People are (rightfully) scared of losing their jobs.
  • Nobody trusts the hallucinations.
  • And the big one: "If it takes me more time to verify what the AI did than to just do it myself, I'm not using it."

I was stuck in that same loop a couple of months ago.

I just wanted to solve one "simple"—yet incredibly painful—problem I had every day: minimising the time between what I wanted AI to do and the output.

But it was never simple. It was a whole dance: highlight, copy, open a new tab, paste, wait for the page to load, fix the weird formatting, realize you forgot to add context, tab back, re-copy, re-paste... you know the routine.

Then I hit one keyboard shortcut, watched a new chat open with my text and prompt already in place, and the reply started instantly.

Forty-five minutes → forty-five seconds.

That “oh come on” moment is why I wrote this. AI can save you hours, but the tiny bits of friction (new tabs, copy/paste, hunting for prompts) make us default to old habits. So let’s remove the friction.

Below are 3 low-effort AI habits I use daily that save me 20+ hours/week. They’re simple. They stack. And they’ll make you feel like you’ve enabled a cheat code.

Quick map of what’s inside:

  • Habit #1: The Everywhere AI Shortcut
  • Habit #2: Text Expanders (prompts on tap)
  • Habit #3: The Prompt Multiplier Method (use when stakes are high)
  • Bonus: Embed AI “triggers” where you already work
  • TL;DR checklist + plug-and-play examples

Habit #1: The Everywhere AI Shortcut

Goal: Open the exact AI action you need from anywhere in 1–2 keystrokes. No tabs. No hunting. Just go.

How it works in real life

I press Option + Space → type gbt␣ and paste a prompt → Enter.

A brand-new ChatGPT chat opens and starts running.

If I type claude p → Enter, I land on my Claude Projects page.

gem → Enter, I’m in a new Gemini chat.

No context switching. No “let me just open one more tab.”

What you need

  • Mac: A launcher like Alfred (paid) or Raycast (free)
  • Windows: PowerToys Run (free) for launcher behavior

Setup (conceptual, step-by-step)

  1. Install a launcher and set a global hotkey (I use Option + Space).
  2. Create custom web searches for your AI tools:
    • ChatGPT new chat: keyword gbt + a dynamic {query} parameter so what you type after gbt becomes the prompt.
    • Claude new chat / Projects page: same idea, but with Claude’s URLs.
    • Gemini: Gemini doesn’t expose a dynamic URL for prefilled prompts, so use a static new-chat URL and just open an empty chat with keyword gem.

Pro Tips

  • Web search: create a second shortcut like gbt search␣{query} to force ChatGPT to use browsing when you need it.
  • Force deeper thinking (paid): add “Think hard about this.” to the end of your prompt to trigger extended thinking. In my testing (and what I’ve seen elsewhere), defaulting to Thinking mode (not Auto) often yields more rigorous answers on complex tasks.
  • Use static jumps: Quick-open places you visit daily:
    • claude p → Claude Projects
    • gbt p → your ChatGPT Projects page

Why this saves hours

If you're using AI extensively (like me):

  • You eliminate 5–8 micro-steps per ask.
  • You stop forgetting to use AI because the entry point is your keyboard, not your browser.

This might sound like just a 5% increase in output. It's not. When added up across a day, it becomes truly impactful.

Habit #2: Text Expanders (Prompts on Tap)

Goal: Type a 3–6 letter trigger and watch your full, tuned prompt appear—anywhere you’re typing.

What it looks like

I type colgpt (short for “concise/clear”) and—boom—my full editing prompt drops in. I use it 20+ times/day to tighten writing. I don’t have the patience to retype it every time, and you shouldn’t either.

Tools that work

  • Mac (Alfred): Snippets
  • Mac (Raycast): Text Expander extension
  • Windows: Beeftext (free, open-source)
  • Keyword: colgpt (or whatever you want)
  • Content: your full prompt with bracketed options (see below)

I use these, to refine emails or writing with a specific writing style, respond to prospects, check grammar and spelling, among many other use cases.

My 3 most-used prompts (steal these)

1) Clear & Concise Editor — colgpt

You are an expert editor. Rewrite the text I paste next to be:
- {trim level: "trim the fat" | "halve the word count" | "make it ultra-concise"}
- Plain, specific, active voice
- Keep the meaning intact
Return: improved text only.

2) Email Polisher — polish‍

Revise the email below to be clear, brief, and friendly-professional.
- Subject line options (3)
- Body ≀ 120 words
- One clear ask and next step
Return subject + final email.

3) Formatter — format

Reformat the text below into:
- A 5-bullet TL;DR
- A tidy outline with H2/H3
- A 3-step call-to-action
Keep my tone. No extra commentary.

Bonus for creators: Instant Analogy Generator — analog

“Explain the concept I paste using 3 simple analogies (one technical, one everyday life, one humorous).”

How to build the habit

  • Pick 3–5 prompts you use constantly.
  • Turn each into a snippet and force yourself to use each once/day for 2 weeks.
  • Add options inside the snippet ({trim level: ...}) so you can adjust on the fly without digging up a different custom GPT.

Chaining = chef’s kiss

  • Use the launcher to open a new Claude chat with your text prefilled → type colgpt → your prompt expands instantly → Enter.
  • No copy-paste marathons. Just type and go.

Habit #3: The Prompt Multiplier Method

(Use this when the stakes are high.)

Frontier models are so good that even lazy prompts can work
 until the decision matters.

If you’re spitballing ideas for a team stand-up? Fine, be lazy.

If you’re deciding whether to refinance a mortgage? Maybe don’t gamble.

The 6 ingredients of a strong prompt

  • Task
  • Context
  • Example
  • Persona
  • Format
  • Tone

Do we always include all six? Honestly, no. But we should when money/reputation/risk is on the line.

I've seen many "prompt anatomy" posts for specific models, for reasoning, etc. In my experience, this simple basic framework worked better in all situations.

Two scenarios

  • Low stakes (lazy is fine):
  • “Act as a senior PM. Given this goal, suggest improvements.” → You’ll still get decent feedback.
  • High stakes (multiply your prompt):
  • Instead of guessing, ask your model to rewrite your basic ask into an optimized prompt for itself.

Now, 80% of my prompts are produced by another AI. Gemini 2.5 Pro and reasoning models in general are good prompt engineers.

Copy this “Prompt Multiplier” once and save it as a snippet (multi)

You are a prompt engineer for {MODEL}.
Rewrite my rough prompt into 3 optimized versions tailored to {MODEL}'s strengths.

Each version must include:
- Clear Task, Context, Persona, Format, Tone
- Inputs you need from me (as a short checklist)
- A suggested evaluation rubric
- If quantitative, include a small comparison table

Return all 3 versions titled V1/V2/V3.
My rough prompt: """{paste here}"""

‍Example (high-stakes money decision)

Rough prompt:

“Bank offered 4.2% on a refinance. What factors should I consider?”

Multiplied output might include:

  • Persona: “You are a neutral mortgage advisor.”
  • Format: “Return a concise decision guide + a table comparing monthly payments at 3.7/4.2/4.7.”
  • Inputs checklist: credit score, loan term, break-even horizon, closing costs, prepayment plans.
  • Tone: plain, practical, no fluff.

You’ll instantly see a night-and-day difference. And because it’s high-stakes, the extra step is worth it. Save your favorite version as a snippet for next time.

Bonus: Embed AI “Triggers” Where You Already Work

If you have to remember to use AI, you won’t. So don’t remember—embed.

My setup

  • Notion “Prompts” database: every useful prompt gets its own page with examples. I link those pages anywhere I need the nudge.
    • In my 1:1 doc with my manager, a pinned link to “Refine Unstructured Thoughts” sits at the top—so I use it before every meeting.
  • Calendar reminders: my weekly “Summarize snippets for leadership” event includes a direct link to the exact ChatGPT/Claude project.
  • Projects: I bookmark my AI project pages and jump in via launcher keywords (gbt p, claude p).

Result: AI is always one click away at the moment of need. No willpower required.

[Screenshot Placeholder]

Tiny Dialogues (Because You’re Probably Thinking This)

“Isn’t this overkill?”

No. This is removing pebbles from your shoes. Over a week, those pebbles are hours.

“Do I need paid tools?”

No. Raycast + PowerToys Run + Beeftext are free. Paid tiers and Thinking mode help, but the habits work regardless.

“Will these make me dependent on AI?”

They make you dependent on good process. You’ll still think—just faster, with better output.

However, DO make sure that your day is not filled by “you talking and using AI output mindlessly”. Make sure you're using your brain in whatever you're doing. Using AI all day without you using your neurons is a serious threat to your brain.

TL;DR Checklist

  • Everywhere Shortcut: set gbt, claude, gem, plus static jumps (claude p, gbt p).
  • Browsing Shortcut: add gbt search for web-enabled runs.
  • Thinking Mode: paid users, default to Thinking; add “Think hard about this” when depth matters.
  • Text Expanders: create colgpt, polish, format, analog (and your top 1–2 job-specific prompts).
  • Prompt Multiplier (multi) for high-stakes tasks; save your best version.
  • Embed Triggers: Notion prompts DB, calendar links, project bookmarks.

Quick Examples You Can Use Today

  1. Workday reality:
    • You highlight a messy paragraph in Docs → Cmd+C → Option+Space → claude␣ → paste → Enter → type colgpt → Enter. Done.
  2. Design crit:
    • Type analog after pasting an image description. Get 3 analogies you can use to sell a direction to stakeholders.
  3. Weekly status:
    • Paste raw notes → format → get TL;DR bullets + a clean outline you can drop into email or Slack.

If you implement just one of these today, make it the Everywhere AI Shortcut. It’s the highest-leverage unlock because it kills the “Ugh, I’ll do it later” moment.

Until the next one,

— Charafeddine

Share this article on: