Building resilient teams: how to help people take a punch and do hard things

Executive overview

Product leadership puts teams under constant pressure — from critical feedback, shifting strategy, and the fear of being judged. Most managers prepare teams to succeed but not to recover when things go wrong. Hilary Gridley, Head of Core Product at Whoop, shares a practical playbook drawn from cognitive behavioral therapy, behavioral psychology, and hard-won experience.

The core insight: resilience is a teachable skill — action reverses negative spirals, not rumination.

Taking a punch: counter-programming bad narratives

• When someone feels criticised or misread, the instinct is to litigate — don't.
• Instead, ask: what is one action I can take that demonstrates the opposite of what I fear this person thinks of me?
• Defending yourself draws more attention to the mistake; forward action changes the narrative.
• Example: after accidentally laughing at a serious suggestion, Hilary sent a follow-up note proposing a related idea rather than explaining herself.
• This is behavioral activation from CBT applied to work: act first, feel better second — not the other way around.
• Give people agency by focusing them on the next step, not on what another person thinks.

Teaching your team to do hard things

• Most people gravitate toward problems they're likely to succeed at; the hard, important problems go unsolved.
• Fear of judgment is the main blocker to speaking up in meetings, taking risks, and putting work out there.
• Teaching people to take a punch reduces the fear of getting punched — which unlocks bolder behavior.
• Challenge negative thinking directly when you see it; don't just validate the spiral.
• Let the pity party happen briefly, then redirect to action as fast as possible.

Building a shared mental model of leadership

• Understanding what the CEO thinks is less useful than understanding how they think.
• Share a weekly rundown with your team: here's what I heard in key meetings, here's my interpretation, here's what I'll do differently.
• When you see feedback repeated across reviews, name the pattern and explain the reasoning behind it.
• This enables your team to make decisions independently without approval chains slowing everything down.
• Approval chains exist because people have different mental models — not because of bad process.

The magic questions technique

• Instead of asking open-ended questions to understand someone's thinking, make a statement and ask "do you agree?" or "is that right?"
• This surfaces mental models faster, calibrates judgment, and avoids creating dependency on you for answers.
• Works with executives, legal teams, and your own reports.
• Requires genuine openness to being wrong — or the technique backfires.

Disagreeing with leadership without undermining your team

• First, try the "what if I'm wrong?" exercise — ask what would have to be true for the other person to be right.
• If you still disagree, be candid but respectful; explain the rationale from their point of view even if you don't fully share it.
• Never tell your team a decision is stupid — it removes agency and poisons execution.
• Frame it as: "I might not fully agree, but here's why it makes sense from their perspective, and we'll find out together."
• You are probably not the protagonist in the story of the company — and accepting that makes you more effective, not less.

Building habits and reward loops on your team

• Habit formation is more effective than education models (teach, assess, enforce).
• Three levers: consistency (start tiny, every day), reduce friction (don't force AI into deadline-heavy work first), reward loops (powerful, immediate, emotional).
• Shout out the behaviors you want to encourage — working late gets rewarded inadvertently; make space for the behaviors you actually want.
• Ask in one-on-ones: "Is there something you're doing every day that brings you joy?" If not, treat it as a problem.
• Model taking care of yourself publicly so your team has permission to do the same.

Using AI to accelerate learning

• AI compresses the feedback loop for building judgment — the thing that normally requires years of low-level reps.
• Build custom GPTs that think like you so your team can get near-real-time feedback on demand, not just in one-on-ones.
• Create scenario-based practice tools (e.g., LSAT-style logical reasoning questions set in PM contexts) for deliberate skill-building.
• Start AI adoption with fun, low-stakes use cases to build habit before applying it to high-pressure work.
• The concern about lost entry-level learning is real, but AI can make that learning faster and more targeted, not absent.