Pull
The call already happened. The transcript is raw material, and it is usually richer than the note you would have written from memory.
Day 2
One call in. A stack of finished assets out. Day 2 gives you a repeatable line for turning real source material into things the team can use.
Use a transcript if you have one. A Slack thread, client question, or rough note works too.
The call already happened. The transcript is raw material, and it is usually richer than the note you would have written from memory.
Hand the source to Claude. Ask for stories, angles, objections, examples, phrases, and assets worth building.
Pick the best one and turn it into a real thing: landing page, carousel, cheat sheet, prompt pack, or client resource.
Layer in the names, voice, examples, colors, and magic moment. Now it feels like it came from this team.
Drop it where people can use it. Then save the move so the next source does not start from scratch.
A finished, on-brand asset from material you already had.
What you are doing today
Start with one real source. Do not make the source perfect. Claude can help you find the useful parts after you give it the raw material.
A transcript is ideal. If you do not have one nearby, use a Slack thread, client question, messy note, or rough content idea.
Use the prompt below. Claude will sort the raw material into asset candidates and tell you which one is easiest to ship first.
Pick one asset and ask Claude to make the first version. Rough, specific, and useful beats polished and generic.
Copy this
You are helping me turn one real source into useful assets. The source might be a call transcript, Slack thread, client question, messy note, or rough idea. Goal: Build an asset map from the source, then help me create one first draft. Rules: - Work from my actual source material. - Do not invent details. - Ask for missing context only when it changes the output. - Keep the language plain. - Make practical assets a team could actually use. - When you make a claim, point back to the source detail that supports it. Step 1: Ask me to paste the source material. Step 2: After I paste it, extract: - Stories or examples - Questions people asked - Objections or hesitations - Useful phrases - Teaching points - Client or cohort insights - Repeated patterns - Any obvious asset ideas Step 3: Create an asset map with 8-12 asset candidates. Use this format: ## Asset Map | Asset idea | Source detail | Who it helps | Why it is useful | Effort | Use effort labels: Low Medium High Step 4: Pick the top 3 assets to build first. For each one, explain: - Why this one first - What format it should take - What source details it should include - What would make it feel specific to our team Step 5: Ask me which one I want to build. Step 6: After I choose, create the first draft. Make it usable, not theoretical. End by giving me: 1. The draft 2. A short checklist for improving it 3. A suggestion for where to ship it
Working with Claude
Everything here bolts onto how you already work. Keep your way of doing things. Claude just makes the useful parts easier to pull out, shape, and reuse.
Claude only knows what you tell it. The more it knows about the person, goal, constraint, and quality bar, the better the result.
A wall is usually the next thing to ask about. Paste the confusing part, send the screenshot, or ask what the error means.
Your calls, transcripts, notes, questions, and messy drafts are usually the richest material.
Claude builds from the words you give it. Fuzzy in, fuzzy out.
Anything you repeat is a workflow. A workflow you save once becomes easier the next time.
Know your tools
You just need the map. Start in chat, grow into Cowork, and use Projects and Skills when the same context or workflow keeps coming back.
Regular conversation. You ask, it answers. This is where you start, and where plenty of good work still happens.
Claude works beside a folder or real workspace. Give it the goal, source material, and constraints. It does the multi-step job and reports back.
You will live hereThe most powerful version. Still plain English, but it works with files, sites, automations, and build environments.
The default handles most work. Use faster, lighter models for quick throwaway jobs. Use the strongest model when the thinking is harder.
Set one up for a client, cohort, or repeated body of work. Drop the files and instructions in once so each chat starts with context.
Think of it as an onboarding doc for a new teammate. Build one after you have done the job twice.
Example from Day 1
One teaching call became a web playbook, PDF, analytics sheet, carousel direction, and prompt pack. That is the move. Start with source material that already exists, then make one useful thing from it.
Reference build
Use this as proof of the pattern. It came from real Brick by Brick material, not a blank prompt.
Interactive web resource
Before you call it done