Good Prompting: four quick checks that keep AI useful
We all love the moment when a chatbot shoots back a brilliant idea. But flashy text is not the same as real progress.
Before you trust any AI answer—whether it’s writing copy, fixing code, or planning your quarter—run it through four simple checks. Skip one and the output can look smart while steering you straight into a ditch.
1 Integrity — solid facts, no fluff
Ask the model where its statements come from and what could make them wrong. If it can’t show sources or plausible counter-points, you’re holding fluff, not knowledge.
2 Momentum — a clear next step
A wall of text is useless without motion. Make the bot outline one concrete action and then challenge its own plan once. That small loop turns a static paragraph into a moving project.
3 Relevance — fits your real limits
Great ideas die when they ignore budget, time, or risk. Tell the AI your constraints—“$1k spend, two-week deadline, low legal risk”—and force the answer to adjust. Relevance keeps you from pitching moon-shots you can’t fund.
4 Impact — a metric that moves in 30 days
End with, “What number will this change, and how will we measure it a month from now?” If the bot (or you) can’t name a metric and a check date, the plan won’t pay off—and you’ll have no way to learn why.
Why these four matter together
Integrity stops self-delusion. Momentum kills analysis paralysis. Relevance keeps the plan grounded. Impact proves the work paid off. Clear all four gates and you’ve turned raw tokens into a live experiment that either wins cash or teaches you fast.
Real value shows up when the answer survives every layer above.
Drop-in prompt
Copy-paste this into any AI chat mid-thread and let the model audit itself:
Act as a self-auditor. Score your answers above against these four checks:
1 Integrity – facts & sources clear?
2 Momentum – next step defined and improved?
3 Relevance – fits my budget / timeline / risk?
4 Impact – metric named and how to measure in 30 days?
Reply with four lines: Check | Pass/Fail | 1-sentence fix.
Run the audit, tighten the weak spots, and ship the work. Good prompting isn’t magic—it’s just disciplined thinking, sped up by a very fast assistant.