The Before/After Method

Everyone has burned an afternoon trying prompts and still can't say whether it paid off. Before/After fixes that with a deliberately dumb comparison: measure the work the way it used to happen, then measure it the way it happens now. No vanity metrics, no guesswork.

What is the Before/After Method

It's simple: for each task, you compare how you'd do it without AI with how you do it with AI. The method reveals exactly where AI adds value and where it gets in the way.

It's not about prompts. It's about process. The prompt is only the connector between the two halves of the comparison.

The Before: how you do it today

Pick a real task from your work. Write the step-by-step of how you do it today, without AI. Note:

• How long does it take?
• What's the most tedious part?
• Where do you get stuck?
• What do you need to double-check?

That list is your baseline. Without it, every AI answer looks impressive — because you have nothing to compare it against.

The After: how to do it with AI

Now redo the same task with AI. Use the Perfect Prompt Formula. Compare:

• How long did it take?
• Did the quality improve or worsen?
• What did the AI do well? What did it do poorly?
• What did you need to adjust?

The official prompt engineering guides from OpenAI and Anthropic make the same point: ask for context, task, constraints, and output format. In Before/After, you find out which of those four were missing from your prompts — by looking at the result, not at the theory.

Example: writing a weekly report

Before (without AI) — 45 min

1. Open weekly data (5 min)
2. Write summary (15 min)
3. Review and adjust tone (10 min)
4. Double-check numbers manually (15 min)

After (with AI) — 12 min

1. Paste raw data into chat (2 min)
2. AI generates structured summary (30 sec)
3. You review tone and accuracy (5 min)
4. AI checks numbers and formatting (2 min)
5. You approve final version (2 min)

The savings are not the important part. What matters is that the output is now reproducible: every week you run the same flow and get the same quality. Time saved turns into routine, and routine turns into an edge.

How to apply today

1. Pick ONE task you run every week.
2. Write down the Before, with time per step.
3. Do the After with AI and time it again.
4. Record the difference — and what is left for you to review.
5. Repeat with two more tasks.

Three tasks in, you have a personal AI mini-system running — no course, no hype, no fluff. If a task doesn't improve, that result is data too: maybe AI is the wrong instrument there. To make that call with more structure, see the Ladder Rule and the definition of the perfect prompt formula.

Frequently asked questions

Does Before/After work for any task?

It works for recurring tasks, where the gains compound. For a one-off question — convert a currency, summarise a page — the comparison isn't worth the note-taking. Focus on what shows up every week.

Do I have to time everything precisely?

No. An honest estimate is enough, as long as you use the same yardstick on both sides. What you cannot do is remember the Before and stopwatch the After.

What if the quality gets worse with AI?

Then write that down: the method is just as good at confirming a gain as it is at catching a loss. Usually the problem sits in the input — incomplete data, missing constraints, no requested format. Fix the After and measure again.

How long before the method pays off?

The first task already exposes rework you hadn't noticed. The habit of comparing usually sticks by the third or fourth task. After that, opening an AI chat without thinking about the before feels wrong.

For the full walkthrough, the definition of the Before/After Method keeps the model in reference form — start by applying it to the two tasks that eat the most of your week.