A second opinion on advice

Questions to Ask Before You Take Anyone's Advice

The useful question is not "do they mean well?" — you can't verify that, and you don't need to. The checkable questions are structural: who carries the risk, what closes if you say yes, and what would need to be true for this advice to fit your life.

By Helm AI Inc · Published October 18, 2026 · 8 min read

Someone gives you advice. A boss says wait one more year and the promotion will come. A parent names the major, the city, or the career that will make you safe. A partner wants to move; an advisor wants the deal closed this week; a friend wants you to co-sign, co-found, or co-invest.

The instinct is to evaluate the person — are they genuine, do they really care, what's their angle. That instinct is understandable and mostly useless. You cannot see anyone's motives, you cannot prove them, and the whole conversation turns into a relationship fight instead of a decision. Meanwhile the advice might be excellent or terrible regardless of how sincere the person is.

There is a better frame. Judge the structure of the advice, not the character of the advisor: who benefits, who carries the downside, which options survive, and what would have to be true for the advice to be right for you specifically. None of those require mind-reading, and all of them are checkable.

This is also where the market is heading. A 2025 KPMG survey found that about a quarter of US adults had already used generative AI for concrete advice on things like health and travel, and Pew Research Center has found that roughly half of Americans are more concerned than excited about AI. People want a second read on the advice they're getting — and they're right to be cautious about anything that claims to know what's best for them.

Question 1: Who carries the risk if this goes wrong?

The boss who says "one more year." If you wait, he keeps a reliable employee for another cycle. If you wait and nothing changes, he has lost almost nothing; you have spent a year of your career on a promise that was his to make and yours to lose. The advice might still be correct — timing matters in every job — but the risk is asymmetric, and that asymmetry is a fact you can hold without accusing anyone of bad faith.

The parent who names your major. They get a child in medicine or law or civil service — a future they can picture and talk about at family gatherings. You get the coursework, the exams, the student debt, and the decades. Their exposure is reputational; yours is biographical.

The friend who wants a co-founder. They get a partner they already trust. You get the personal guarantee, the second mortgage conversation, and the version of your life where the company fails. Both of you may genuinely believe it will work. Only one of you is betting the same amount.

Question 2: What closes if you say yes — and what actually happens if you say no?

A genuine recommendation survives your hesitation. The person who really means it will usually add information, not pressure: here's what I'd want to know in your position, here's what I'm uncertain about. Pressure behaves differently. When you decline, the deadline tightens, the guilt arrives, or the relationship itself becomes the stakes. That reaction is information about the pressure — not proof of bad intent, and you don't need it to be proof of anything to take it seriously.

This is also the question that protects you from yourself. Half the advice people regret taking was advice they never actually agreed with; they just never noticed the alternatives closing. Writing down what saying yes costs you in foregone options makes the trade visible.

Question 3: What would need to be true for this advice to be good for you?

Worked example. Your uncle says take the stable job at the family business. What would need to be true? That you'd still be there in two years without resentment. That the family dynamics can absorb a bad quarter. That the stable salary actually covers your obligations. That you've genuinely tried the alternative and it didn't fit — not that you're avoiding it. Some of those conditions are checkable today, and some aren't. Now you know which ones you're betting on, and you can say so out loud: "I'd take this if I were confident about X and Y. I'm not yet. Here's how I'd find out."

That sentence — here is what would need to be true for this advice to be good for me — is the single most useful thing you can bring to any advice conversation. It keeps the relationship intact and moves the discussion from "who do you trust" to "what do we actually know."

A fourth question worth asking: what do they actually know?

You don't have to audit the person. One question usually does it: "Have you done this, or someone you watched closely?" The answer recalibrates everything — and people who have done it tend to give more hedged, more useful advice than people who haven't.

How to use this without turning every conversation into an interrogation

• Ask one or two questions, not all four. "Who carries the downside if this doesn't work?" and "what would need to be true for this to be right for me?" do most of the work.
• Ask sincerely. These are not gotcha questions. If they sound like cross-examination, you'll get defensive answers, and defensive answers tell you less.
• Answer them yourself first. You can run the whole checklist alone, on paper, before you ever talk to anyone. The conversation gets easier when you already know what you're checking.
• Keep the verdict about fit, not people. The output is "this doesn't fit my goals, risk, or timeline" — never "they don't really care about me." One is a decision you can act on. The other is a fight.

None of this makes you suspicious of everyone who cares about you. Most advice you'll ever receive is well-meant, and some of it will be the best input you get. The point of the questions is narrower: they let you tell the difference between advice you've weighed and advice you've simply absorbed — before the decision becomes expensive to reverse.