AI Guides › Claude Mastery
Claude For Big Decisions: Slowing Down On Purpose
By Nigel Guy · 2 min read
A big decision gets the same session pace as a small one — a quick back-and-forth, an answer that feels settled, a sense of relief that it's been thought through. The speed is the problem. Claude will happily produce a well-reasoned-sounding conclusion in the same five minutes it would take to answer a trivial question, and a fast, fluent answer to a big decision feels like clarity when it might just be the first coherent thing that was said.
The rule: the size of the decision should set the pace of the session, not the other way round — a big decision deserves a session that's deliberately slower than the conversation naturally wants to be.
The mechanism
- Separate the exploration session from the decision session, on purpose. Use one conversation to lay out the landscape — options, trade-offs, unknowns — with no pressure to land anywhere. Use a later, separate session to actually decide. Collapsing these into one sitting is the single biggest driver of decisions that feel resolved too fast.
- Ask it to argue against the conclusion you're leaning towards, specifically, before you accept it. Not a generic "any concerns?" — a direct request to make the strongest case for not doing the thing you already want to do.
- Name what would have to be true for this to go badly, and check, honestly, whether you actually believe those conditions don't hold. This surfaces the assumptions the fast version of the conversation quietly skipped past.
- Sit on it between sessions. A big decision reasoned through once, fluently, in one sitting hasn't been pressure-tested by time — coming back a day later with fresh eyes catches a different class of problem than reviewing the same reasoning immediately.
- Write the decision down with its reasoning, before you act on it. Not just the conclusion — the reasoning that led there. If you can't state why, in writing, you likely haven't finished deciding, you've just stopped talking about it.
What to skip
Skip using a single chat session as if it were a deliberation process — a good back-and-forth is a useful input to a decision, not the decision-making process itself. And skip treating "Claude and I talked it through" as equivalent to having consulted the people the decision actually affects; for anything involving other people's stake in the outcome, that conversation still needs to happen separately, with them.
Guardrails
- Slowing down helps you notice weak reasoning and unexamined assumptions. It does not add information you don't already have access to — Claude isn't a source of facts about your specific situation beyond what you've told it or it can look up.
- For decisions with legal, financial, medical, or safety consequences, this process is a way to think more clearly before you consult a relevant professional — not a substitute for consulting one.
- The instinct to slow down should scale with how hard the decision is to reverse, not with how important it feels in the moment — a decision that's easy to undo doesn't need this treatment even if it feels weighty right now.
All 751 AI guides · JulieMango plans from £17/mo