Deep guide

The Thinking Partner

Most people use Claude as a production machine — give it a task, get an output, done. That's useful but it's not even close to what's possible. The real power is using Claude to think — to challenge your assumptions, stress-test your plans, and surface what you're not seeing.

The difference between producing and thinking

When you ask Claude to "write me a business proposal," you're asking it to produce. It'll give you a competent, generic document. You'll edit it a bit, send it, forget it.

When you ask Claude to "read my situation and tell me the three things most likely to kill this business in the first year," you're asking it to think. Now you get something you couldn't have written yourself — a clear-eyed analysis built on your specific details, identifying risks you've probably normalised because you've been staring at the idea for too long.

One replaces a task. The other changes a decision.

"The best conversations with Claude end with you thinking differently, not just with a document you didn't have before."

Eight techniques that turn Claude into a thinking partner

You've made a decision. You're fairly confident. Now get Claude to argue the opposite — not to change your mind necessarily, but to make sure you've actually considered the other side. Most people's biggest decisions have never been seriously challenged.

I've decided to [your decision]. Before I commit — argue the strongest possible case against it. Not the obvious objections I've already thought of. The ones I'm probably glossing over because I want this to work.

Use when: making a major decision you're already leaning toward — job offer, investment, business move, relationship decision.

Standard risk assessment asks "what could go wrong?" A pre-mortem is more powerful: it assumes the thing has already failed and asks why. This reframes your brain away from optimism bias and forces it into root-cause thinking. It consistently surfaces things the forward-looking version misses.

It's 18 months from now. My [plan/business/project] has failed badly. Not a small stumble — it's over. Work backwards. What most likely went wrong? Give me the top 5 causes, ranked by probability. Be specific to my situation: [paste your situation].

Use when: launching something new, making a significant investment, starting a project with real stakes.

There's a specific type of thing you consistently don't think about — your blind spots. They're invisible to you by definition. Claude, having read essentially everything written by humans, has a pattern library you don't. Asking it directly what you're probably missing almost always produces something worth hearing.

Here's my plan/thinking: [explain it]. What am I probably not thinking about? Not what I got wrong — what I haven't thought of at all. What do people in my situation typically overlook that turns out to matter a lot?

Use when: you've been working on something for a while and want a fresh set of eyes that's genuinely different from your own perspective.

The opposite of a strawman. A strawman is the weakest version of an argument you disagree with. A steelman is the strongest. If you're about to dismiss an idea, a competitor, a strategy, or a person's argument — ask Claude to give you its best version first. You might still disagree, but you'll disagree more intelligently.

I think [position/idea/person] is wrong/weak/not relevant. Before I dismiss it — give me the strongest possible version of the argument for [position/idea]. Assume the person arguing it is very smart and has good reasons. What's the best case they could make?

Use when: evaluating competitors, considering an idea you initially dislike, or understanding someone you disagree with.

Professionals — doctors, lawyers, financial advisors — are expensive and unavailable at 11pm. Claude won't replace your accountant, but it can help you walk into that meeting already knowing the right questions to ask, the things to push back on, and the options you haven't been offered. It's not about replacing expertise — it's about being an informed client.

My [advisor/doctor/lawyer] told me [advice]. I want a second perspective. What are the things they might not have mentioned? What questions should I be asking that I probably haven't? What would a different approach to this look like? I understand this isn't professional advice — I'm trying to have a more informed conversation with my professional.

Use when: you've received professional advice and want to understand it better before acting on it.

Every plan is built on assumptions. Most of those assumptions are invisible — you've stopped questioning them because you've been living with them. Having Claude list the assumptions embedded in your thinking forces you to see which ones are solid and which ones are doing a lot of load-bearing work without ever being tested.

Here's my plan/business/decision: [describe it]. List every assumption this depends on being true. Then rate each one: which assumptions are well-supported, which are untested, and which ones — if wrong — would collapse the whole thing?

Use when: planning anything that involves predicting how other people, the market, or the economy will behave.

Every situation looks different depending on who's looking at it. The same business proposal that sounds brilliant to you looks completely different to a bank, a potential investor, a sceptical client, or a competitor. Claude can occupy any of those chairs and give you their honest reaction — which is often more useful than another round of your own thinking.

Read this [proposal/plan/email/idea]: [paste it]. Now give me three separate reactions: (1) How a [senior banker / potential investor / sceptical customer] would read this. (2) What they'd immediately question. (3) What would need to change to get a yes from each of them.

Use when: preparing something that needs to persuade someone with a different set of priorities or concerns than yours.

Before you commit fully to something, Claude can help you design a small, cheap test that validates the most important assumption. This is the thinking pattern behind lean startup methodology — but it applies to personal finance, career decisions, relationships, and creative projects too. Test the key assumption before you bet everything on it.

I'm thinking about [big decision/commitment]. I don't want to fully commit until I know if [key assumption] is actually true. What's the smallest, cheapest, fastest test I could run in the next 2–4 weeks that would tell me whether this is worth pursuing fully?

Use when: considering a major commitment — starting a business, changing careers, making a large investment — that can be tested on a small scale first.

A real example: making a career decision

Here's how you'd chain these techniques together for a real decision — whether to leave a stable job to go freelance:

Full thinking workflow — career decision

Start with context: give Claude your full situation — income, savings, skills, dependants, risk tolerance, why you're considering it.

Run the Blind Spot Finder: "What am I not thinking about?"

Run the Assumption Audit: "List every assumption my freelance plan depends on."

Run the Pre-Mortem: "It's 12 months from now and this failed. What happened?"

Run the Minimum Viable Test: "What's the smallest way I could test this before quitting?"

Synthesise: "Based on everything we've discussed — what's your actual recommendation? Be direct."

That conversation takes 20–30 minutes and you'll come out of it with a clearer view of your decision than you'd get from weeks of thinking alone or a single conversation with a friend. Not because Claude is smarter than you — but because it has no stake in your decision, no emotional investment, and no social obligation to be supportive. It will just tell you what it sees.

Start here: Take one decision you're currently turning over in your head. Don't ask Claude what to do. Ask it: "What am I probably not thinking about?" That one question almost always changes the conversation.

Open Claude and try it →