The Claude Prompting Framework: Outcome, Task, Context, Format ⚡

Claude is not the problem. The prompt is the problem. Most people open a chat, type one line like "help me get customers," and then decide the AI is overrated when it hands back a list of tips they could have found on page one of Google.

Here is what is actually happening. Claude has no idea what you are trying to achieve, no idea what your business is, and no idea what a good answer even looks like to you. So it does the only safe thing it can do. It gives you the average answer. A mid answer, because you gave it a mid prompt.

The fix is not a magic word. It is not "act as a world class expert." It is four pieces of information, in order, every time.

• 1. Outcome. Where am I trying to end up? Skip it and Claude optimizes for finishing the task, not for winning. You get technically correct, practically useless.
• 2. Task. What exactly do I want it to do right now? Vague in, vague out. You get advice instead of work.
• 3. Context. What does it need to know about my situation? Skip it and it writes for a generic person with a generic business. Sounds like everyone else.
• 4. Format. How should the answer be delivered? Skip it and you get five paragraphs of prose when you needed a spreadsheet.

The order matters

Outcome first. Always. When Claude knows the end goal, it works backwards from it and every choice it makes downstream gets better. If you lead with the task, it just does the task. If you lead with the outcome, it does the task in the way most likely to get you the outcome.

You do not need to label the parts. You do not have to write "Outcome:" and "Task:" like a form (though you can, and it works fine). You just have to include all four things.

This is the one that separates beginners from people who get incredible answers. If Claude knows the end goal, it works backwards to give you the best possible path to it.

Think about it like hiring someone. If you tell a new employee "write me a cold email," they write you a cold email. If you tell them "I am trying to book 10 discovery calls with agency owners this month, and cold email is the channel I am testing," they write you a completely different email. Same task. Different outcome. Way better output.

What an outcome actually sounds like

• Not an outcome: "Write a social post." Outcome: "I want this post to get people to click the link in my bio and join my email list."
• Not an outcome: "Help me with my resume." Outcome: "I am trying to land interviews for senior product manager roles at Series B startups."
• Not an outcome: "Summarize this report." Outcome: "I need to walk into a board meeting tomorrow and defend our Q3 numbers in under 5 minutes."
• Not an outcome: "Give me some marketing ideas." Outcome: "I want to add 1,000 paying customers in 90 days without increasing ad spend."

Notice the pattern. An outcome has a destination in it. Often a number. Often a deadline. Always a reason.

Here's what I'm ultimately trying to achieve:

[the real goal, with a number and a deadline if you have one].

Success looks like: [what you'd see if this worked].

Here's what I'm ultimately trying to achieve:

I want to sign 5 new coaching clients at $2k/month by the end of the quarter.

Success looks like: 5 signed contracts and a repeatable process I can run again.

If you cannot state the outcome in one sentence, you do not know what you want yet. That is not a Claude problem, that is a you problem, and it is worth 30 seconds of thinking before you type anything.

The outcome is the destination. The task is the next step. And the more specific you are, the less room there is for Claude to guess wrong.

"Find me 50 leads that are ideal customers for my business" beats "help me get customers" every single time. One is a job. The other is a wish.

The specificity ladder

• Wish: "Help me get customers." You get a list of generic marketing advice.
• Task: "Find me some leads." You get a vague explanation of how to find leads.
• Specific task: "Find me 50 leads that are ideal customers for my business." You get an actual list, though it may be off target.
• Specific and bounded: "Find me 50 e-commerce founders doing $1M to $10M in revenue who post on LinkedIn about ops problems. One row each." You get a usable list you can work from today.

Four things that make a task specific

• A number. 50 leads. 10 headlines. 3 options. Numbers force it to commit.
• A boundary. Who is in, who is out. What is off limits. What you have already tried and do not want repeated.
• A verb that means something. Write, rank, rewrite, compare, find, score. Not "help with," not "look at."
• A finish line. What makes this task done? Say it out loud.

Task:

[Verb] [number] [thing] that [criteria].

Do not include [what you don't want].

I've already tried [X], so skip that.

This is done when [finish line].

Asking for too many things in one prompt is the flip side of this mistake. "Write my landing page, my email sequence, and my ad copy" gets you three mediocre things. Do one task at a time and use follow-ups.

This is the one everyone skips. It is the one that matters most.

Everyone reads the first two steps and nods along. Then they skip this one, because it takes effort. And this is the step that makes the difference between an answer that sounds like the internet and an answer that sounds like it came from someone who works at your company.

Claude cannot read your mind, your CRM, your brand guidelines, or the conversation you had with your co-founder last Tuesday. Everything it knows about your situation is what you gave it. So give it a lot.

What to actually give it

• Who you are. Your bio, your business, your role, your customers. Stops it writing for a generic person.
• Brand voice. 3 to 5 pieces of your best writing, or a voice doc. It matches your rhythm instead of sounding like AI.
• The raw material. Docs, transcripts, spreadsheets, screenshots, PDFs. It works from your facts, not its guesses.
• What has been tried. Past campaigns, drafts you rejected, what flopped. Stops it handing you the thing you already know failed.
• Constraints. Budget, timeline, tools you have, rules you can't break. The answer becomes actually executable.
• The audience. Who reads this, what they care about, what they hate. Tone and angle land instead of missing.

Where to put context so you only do it once

Pasting the same background into every new chat gets old fast. Claude has three places to store it permanently, and using them properly is the single biggest upgrade to your setup.

Projects (for context that belongs to one body of work). A Project is a folder with a memory. You upload files once and every chat inside that Project can see them. Make one per client, per product, per book, per anything you will come back to. Put your brand guidelines, your customer research, your past work, and a short "how I like things done" note in there.

PROJECT INSTRUCTIONS (paste into the Project's custom instructions)

You are working with me on [project].

My business: [one paragraph].

My customers: [who they are, what they want].

• Be direct. Cut filler.
• If an idea is bad, say so.
• Never write in em dashes.
• Default to [format I usually want].

The files in this project before answering.

Styles (for how it should sound). A Style is a saved voice. Give Claude samples of your writing and tell it to match. Once it is saved, you pick it from the dropdown and every response comes out in your voice. Set one up for your personal writing and one for anything client facing.

Memory (for facts about you that never change). Anything true about you across every conversation goes here. Your role, your business, your tools, your preferences. Say "remember this" and it sticks.

Fastest way to set this up: upload a few of your documents and say "Read these and write me a one-page context brief about my business that I can paste into future chats." Then paste that brief into your Project instructions. Ten minutes of work, permanent payoff.

Bad context vs good context

Same task. Watch what changes.

WITHOUT CONTEXT

Write me a cold email to potential clients.

WITH CONTEXT

Outcome: book 10 discovery calls this month.

Task: write one cold email I can send to e-commerce founders.

I run a 3-person agency doing email marketing for DTC brands.

Average client pays $4k/month. My best case study: took Brand X

from $12k to $61k/month in email revenue in 90 days.

My prospects get 20 agency pitches a week and delete all of them.

They care about revenue per subscriber, not 'engagement'.

I've attached 3 emails that got replies before, match that tone.

Short sentences, no jargon, no 'I hope this email finds you well'.

Under 120 words. Subject line + body. Give me 3 versions with different angles.

The second one takes 90 seconds to write. It is the difference between deleting the output and sending it.

You would be amazed how many people get a wall of text back, sigh, and then manually turn it into a table. Just ask for the table.

Format is the cheapest of the four pillars. It is usually one sentence at the end of your prompt, and it saves you all the cleanup work on the other side.

Format is more than file type

• Structure. "Give me a table with 4 columns: name, why they fit, contact, priority."
• Length. "Under 200 words." "One page max." "Three bullets each."
• Reading level. "Explain it like I'm in seventh grade." "Write it for a CFO."
• Deliverable. "Build this as a spreadsheet." "Give it to me as a document I can download."
• Tone. "Casual, like a text to a friend." "Formal, this goes to a board."
• Order. "Rank them best to worst and tell me why the top one won."
• What to leave out. "No intro, no summary at the end, just the list."

Format: [structure], [length], [reading level], [what to leave out].

Format: a table, 10 rows, plain English, no intro paragraph.

Format: 5 bullets, under 100 words total, explain it like I'm in seventh grade.

Format: build it as a spreadsheet I can download. One tab. No commentary.

"Explain it like I'm in seventh grade" is the most underrated line in prompting. It is not dumbing it down. It forces the model to strip out jargon and hedging, which is where bad answers hide.

Here is the master shape. Every good prompt you ever write is some version of this.

THE MASTER PROMPT SHAPE

What I'm ultimately trying to achieve: [goal, number, deadline].

Right now I need you to: [specific verb + number + criteria].

About me/my business: [short paragraph].

What I've tried: [past attempts].

Constraints: [budget, tools, rules].

I've attached: [files].

Give it to me as: [structure, length, reading level, what to skip].

You can label the parts or write it as normal sentences. Claude does not care about the labels. It cares that the four pieces of information are in there. The labels are just for you, and they do make it easier to spot which one you forgot.

Book 15 sales calls in the next 30 days for my bookkeeping service.

Find me 50 companies that fit my ideal customer profile and rank them

by how likely they are to need me right now.

I do bookkeeping for construction contractors in Texas doing $2M-$20M

in revenue. They come to me when they're growing fast and their books

are a mess. Attached: my client list and pitch deck. Skip anyone

already using a big 4 firm.

A table: company, city, est. revenue, why they fit, the signal they

need help now, priority score 1-10. Sort by score. No intro.

Why this one works. Outcome: 15 calls in 30 days, so it prioritizes urgency over volume. Task: 50 companies, ranked, so it produces a list instead of a lecture. Context: the niche, the revenue band, the buying trigger, plus real files. Format: named columns and a sort order, so it lands ready to work from.

I have to present our new pricing model to the sales team on Monday

and get them bought in, not confused.

Turn the attached pricing doc into a 5-minute talk track I can deliver

from memory.

The sales team hates change. Last time we changed pricing, three reps

quit. Their main fear is that higher prices mean they miss quota. The

new model actually raises their commission on every deal over $50k.

Spoken language, not slides. Under 700 words. Lead with what's in it

for them. Explain it like I'm in seventh grade. End with the three

objections they'll raise and my answers.

Why this one works. Outcome: buy-in, not comprehension. Completely changes the angle it writes from. Task: a talk track from memory, not a summary. Different deliverable entirely. Context: three reps quit last time. That one sentence reshapes the whole thing. Format: spoken language, word count, reading level, and the objections handled up front.

I want this landing page to convert cold traffic from Instagram into

email signups. Currently at 4%, I want 10%.

Rewrite the headline, subhead, and the 3 benefit bullets. Give me 3

versions of each.

Audience: people who follow me for AI tips and have never bought

anything from me. They're skeptical of anything that smells like a

course. The freebie is a real guide, not a teaser. Attached: current

page + my 5 best-performing captions for voice.

Just the copy. Label each version. No explanation unless I ask.

Under 30 words per headline.

Why this one works. Outcome: a number to beat (4% to 10%), so it optimizes for conversion, not for pretty writing. Task: exactly which elements, and exactly how many versions. Context: the skepticism is the whole problem, and it now knows about it. Format: "just the copy, no explanation" saves you scrolling past three paragraphs of preamble.

Your first prompt is a draft, not a contract. The people who get the best results from Claude are not the ones who write one perfect prompt. They are the ones who push back.

When the answer is weak, do not start a new chat and try again with slightly different words. Diagnose which of the four pillars was missing and patch that one.

• Answer is generic and could apply to anyone. Context failed. Say: "You don't know enough about my situation yet. Ask me 5 questions that would let you give a much better answer."
• It gave advice instead of doing the work. Task failed. Say: "Don't tell me how. Just do it. Produce the actual deliverable."
• It's technically correct but doesn't help me win. Outcome failed. Say: "Remember the goal is [X]. Redo this optimizing for that, not for completeness."
• It's a wall of text. Format failed. Say: "Same content. Give it to me as a table with 4 columns and no intro."
• It's too safe and hedged. Task failed. Say: "Pick one and defend it. No 'it depends'. Tell me what you'd actually do."

Three follow-ups worth memorizing

1. THE INTERVIEW (fixes missing context fast)

Before you answer, ask me 5 questions that would let you give me a

much better answer. Then wait for my replies.

1. THE CRITIQUE (makes the second draft much better)

Now critique what you just wrote. What's the weakest part and why?

Then rewrite only that part.

1. THE STAKES (kills the hedging)

Assume this has to work or I lose the client. What would you change?

Be blunt.

The interview prompt is the single fastest way to fix a bad answer. Claude usually knows exactly what it is missing. It just will not tell you unless you ask.

• Treating it like a search engine. One line in, one line out. Claude is closer to a very fast, very well read employee who happens to know nothing about your business. Brief it like one.
• Skipping context because it takes effort. People spend 40 minutes fixing a bad output rather than 3 minutes uploading the files that would have prevented it.
• Cramming five tasks into one prompt. Write my page, my emails, my ads, my captions, and my bio. You get five mediocre things. One task per prompt, then follow up.
• Never telling it the format. If you find yourself reformatting the output by hand, you forgot pillar four. It costs one sentence.
• Starting over instead of pushing back. A new chat throws away everything it just learned about your situation. Stay in the thread and correct it. Only start fresh when the conversation has clearly lost the plot.
• Being polite instead of clear. "Could you maybe help me look at some ideas for possibly improving my marketing?" is four hedges in one sentence. Say what you want.
• Accepting the first answer. The first draft is almost never the best one available. One "critique this and rewrite the weakest part" gets you most of the way there.

Do not paste in sensitive data you would not want stored. Client financials, passwords, private medical info. Redact before you upload. Context is powerful but use your judgment.

Every one of these is built on the four pillars. Fill in the brackets and go. The full guide has 12; here are four to start.

1. FIND CUSTOMERS

Outcome: [X] new customers in [timeframe].

Task: find me [number] companies that match my ideal customer

profile and rank them.

Context: I sell [what] to [who]. They buy when [trigger]. My best

client is [example]. Skip [who to exclude].

Format: table (company, why they fit, buying signal, priority 1-10).

Sort by priority.

1. COLD OUTREACH

Outcome: book [number] calls this month.

Task: write [3] versions of a cold email.

Context: I do [service] for [audience]. Proof: [best result you've

gotten someone]. They get pitched constantly and delete everything.

Attached: 3 emails that got replies. Match that tone.

Format: under 120 words. Subject + body.

1. EXPLAIN IT SIMPLY

Task: explain [topic] to me.

Context: I know nothing about it and I have to

[use it / present it / decide on it] by [when].

Format: explain it like I'm in seventh grade. Under 300 words. One

analogy. Then tell me the one thing most people get wrong about it.

1. BUILD MY CONTEXT BRIEF (do this once)

Task: read everything I've attached and write a one-page context

brief about me and my business that I can paste into any future chat.

Format: cover who I am, what I sell, who I sell to, my voice, my

goals, and what I don't want. Then wait for my replies.

Outcome. Task. Context. Format. If your answer is bad, one of those four is missing. Go find out which one.