Organise · Automate · Breathe
Seven ways to get more out of Claude than a better-worded answer. Flip each card to see why it works and a prompt you can use straight away.
Most people use Claude like a search engine with better sentences. You ask, it answers, you copy and paste. That's useful, but it's not the half of it.
The real value is in the back-and-forth — using it to work through something you haven't figured out yet, to pressure-test a decision, to ask the question you haven't quite formed. This is how I use it every day, and it's changed how I think, not just what I produce.
Start with the messy question.
Claude doesn't need a perfect prompt. Give it the rough version — the half-formed problem, the thing you can't quite articulate. Starting messy gets you to useful faster than waiting until you know exactly what to ask.
Give it a role before you ask your question.
Claude performs differently depending on the context you give it. The role shapes the lens. Be specific — the more defined the role, the more focused the thinking you get back.
Ask it to challenge you, not just help you.
By default, Claude tends toward helpful and agreeable. That's not always what you need. You're not looking for validation — you're looking for the thing you haven't seen yet.
Tell it what NOT to do.
Constraints make Claude more useful, not less. No bullet points, no hedging, give me a direct answer — Claude follows constraints well. The problem is most people don't give any.
Treat the first response as a draft, not a verdict.
The conversation gets more useful as it goes, because Claude builds on what you tell it. People who get one response and stop are leaving most of the value on the table.
When something sounds authoritative, ask where it's coming from.
Claude can sound confident whether it's certain or guessing. It will often flag uncertainty it didn't volunteer upfront — but only if you ask. Especially important for anything financial, legal, or NZ-specific.
Use it to organise your thinking, not replace it.
Claude is excellent at taking a brain dump and making it coherent. But the thinking is still yours. The context, the nuance, the knowledge of your own business — that only comes from you.
Click any card to flip it. Or use the buttons above to flip all at once.
Claude is a thinking partner, not a source of truth. I've had it give me an answer with complete confidence that turned out to be wrong — not maliciously, just wrong in the way a smart person can be wrong when filling in a gap they didn't flag. The problem is it doesn't always flag the gap.
You're asking about something specific to New Zealand — legislation, market norms, tax rules. A generalist AI trained mostly on offshore content doesn't always know what it doesn't know about our market.
You're asking for statistics, citations, or research. Claude can produce plausible-sounding sources that don't exist. Always verify independently before using anything factual.
The answer sounds surprisingly clean and certain. That's often the moment to push hardest. Confident answers on complex questions deserve a follow-up.
You're in territory where Claude's training data might be thin — niche industries, very recent events, or highly local specifics. It will fill the gap without telling you there was one.
The habit to build: when something matters, ask where it came from, and verify it independently. That's not distrust — that's just good thinking.
And it's not only about facts. If you're working through a strategy or decision, Claude will often give you a solid answer — but it's the answer from one angle. Ask it to stress-test the idea: "Tell me why this wouldn't work, and what would need to change so that it could." That's when you start getting the full picture.
The skill isn't learning to trust AI. It's learning to think with it from all sides.
Used well, Claude is like having a thinking partner available at any hour who never gets tired of your half-formed problems.
Used poorly, it's a confident answer machine that occasionally makes things up.
The difference is whether you stay in the driver's seat. You bring the context, the judgment, and the critical eye. Claude brings the pattern recognition, the structure, and the willingness to push back when you ask it to. Together, that's more useful than either of you alone.
Want help building this into how you work?
This is exactly what an Ops & Automation Partner engagement looks like.