A 24-step guide to using Claude as a thinking partner for your business: diagnosis, customers, market, pricing, financials, sales, a pre-mortem and cheap tests. Every step has a prompt you paste, plus a context template and a master prompt.
What you build
| Part | What it is | What it does |
|---|---|---|
| Context | One business context document | Your numbers, audience and goals in one place, reused in every chat |
| Diagnosis | A ranked list of problems | Finds the biggest constraint before you pick tactics |
| Research | Customer, market and competitor analysis | Patterns from real customer words, with every guess labeled |
| Strategy | Positioning, offer, pricing and a financial check | Options with criteria, plus break-even and three scenarios |
| Tests | A pre-mortem and cheap experiments | Pass and fail numbers set before you spend money |
| Review | A weekly review and metrics check | Turns what happened into the next three actions |
Where this comes from
Most people use Claude to write emails and captions. If you run a business, the better use is as a thinking partner: something that diagnoses, compares, argues with you and turns a vague idea into a plan with a test attached. This guide is that workflow, written down as 24 steps. It is my own structure, built from the way I work through a decision, not from a published framework.
What you build
You do not build software. You build a repeatable loop: your business context in one place, a diagnosis, research on customers, market and competitors, a strategy with a financial check, and a weekly review that feeds the next round. The table above is the loop in order. Each step has one prompt you paste, with your own numbers in it.
What it deliberately does not do: make your decisions. Every recommendation in the guide is treated as a hypothesis, and the guide ends each stage with a cheap test and a pass or fail number to set before you run it.
Do you need this? Be honest first
It is for you if you run or are starting a small business and want structure for your thinking: one person, a small team, a side project that is getting serious.
It is not for you if you want Claude to hand you a business, or if you will not give it real numbers. Generic input gets generic strategy, and no prompt fixes that. If you have no customers and no idea yet, talk to ten people first and come back.
What you need
- A Claude account. The prompts are plain text and also work in other assistants, but they were written for Claude.
- Your own numbers: price, costs, customers per month, hours per week, budget. Blanks are fine.
- About 20 minutes to fill in the context template and run the first diagnosis, then one step at a time.
What is in the zip
claude-business-strategist/
claude-business-strategist-guide.pdf the full guide, 24 steps, to read or print
claude-business-strategist-guide.md the same text in markdown, easy to copy prompts from
README.txt where to start
Inside the guide: a fill-in business context template, the 24 prompts, follow-up prompts that make any answer better, a master prompt you can save in a Claude Project, a before-and-after example, and a list of the mistakes I would avoid.
What was verified and what was not
Checked: the guide's structure, that every step has its prompt, that the numbering matches, and that the text has no broken placeholders. The example business in the guide, a bakery, is invented for illustration.
Not checked: I have not tested every prompt across every Claude model or every kind of business, and I have not measured any business result from using them. Claude can state market sizes, prices and competitor details that are wrong, so check any number it gives you unless you supplied the data. Treat the prompts as a starting structure and tell me what broke.
Read the prompt before you run it
Hand each prompt to Claude first and ask what it plans to do with it. A prompt costs a minute to fix. A wrong answer acted on costs a month.
