We all at some point started using AI to help us understand the week’s news. At first, I was inclined to ask for a summary. The summaries were comprehensive, well-structured, and easy to read. But they weren’t particularly helpful. The problem wasn’t that the summaries were missing the news. They were full of news. The problem was that I hadn’t made it clear what I wanted to learn from it.
So I changed the question. Instead of asking what happened that week, I started asking which developments should change my thinking about how to build products, invest in platforms, manage risk, and lead teams. As I wrote in The Prompt Behind My Weekly AI Executive Brief, that shift changed the quality of the output and, more importantly, what I chose to pay attention to.
This also reminded me of something that extends far beyond AI. We put a lot of effort into trying to find better answers. We put far less effort into thinking about the questions that lead us there. AI makes this easier to see, because it’s so quick to respond. Ask a broad question, you might get a polished, broad answer. Ask for context and ask it to look at a specific decision, and the conversation can become much more useful. But the same is true every day, with product reviews, customer conversations, team meetings, and one-on-ones.
The value of asking thoughtful questions has always been there. I think it’s even more important as answers are easier to generate.
Questions are more than just a way to gather information
A lot of organizations use questions as a way to fill in gaps in a status update. Is the launch on track? Did the metric go up? When will the work be done?
These questions serve a purpose. Leaders need to understand progress. They need to make decisions. But if that’s the end of the conversation, there’s a lot we can miss.
Say a team mentions that revenue went up after a product change. That’s useful information. It’s also the start of the discussion. Why is it going up? Are customers finding what they need more easily? Is the experience better for everyone, or is it better for some and worse for others? What do we still need to know before we scale the change?
The follow-up can move the conversation from “we’re reporting a number” to “we’re understanding an outcome.” It can also highlight a risk or opportunity that wasn’t obvious from the first answer.
This is closely connected to what I wrote in The Undervalued Art of Creating Clarity. Teams can be working hard. They can be meeting milestones. But they can still have different views on what problem they’re trying to solve. Asking people to explain the customer problem, desired outcome, tradeoffs, can highlight those differences early enough to do something about them.
There’s a human value to questions, too. When someone shares an idea or concern, asking a thoughtful follow-up question shows that we heard them. We want to understand more. This is important in a product discussion, but it’s just as important in a one-on-one or difficult conversation with a colleague.
The follow-up is often where learning starts
I’ve noticed how easy it is to show up to a meeting with a list of questions. Then, we can just go through them one by one. We get an answer. We acknowledge it. We move on to the next one. We can get through everything on the agenda, but we might miss the most interesting part of the conversation.
With a follow-up, we’re asking ourselves to spend a little more time with the answer. “Can you tell me more about that?” might highlight a constraint that wasn’t in the presentation. “What makes you say that?” might help us understand the evidence behind a recommendation. “What are we missing?” might give someone an opening to share a concern that they were hesitant to raise.
The goal is to listen closely enough to know which answer we should pay more attention to. Some conversations need a direct decision. Some need more space to explore. Part of the work is knowing the difference.
This same habit can help us use AI better. The first response can often be a starting point. We can ask it to explain the reasoning. To distinguish evidence from assumptions. To look at it from another perspective. To identify what would change its conclusion. These follow-ups can expose a weak answer. But they can also help us see that our original question was too narrow.
But there’s a difference. In a conversation with another person, a follow-up can help build understanding. It can help build trust. With AI, it’s a way to explore and test the output. We need to be careful not to mistake the fluency of the response for evidence that it’s correct.
Give people the space to answer
The way we ask a question can influence what we hear.
“Are we on track?” might be the right question when we need a clear answer on a commitment. But in an early product discussion, it can end the conversation before we understand the work. “What’s going well? Where are you seeing friction?” gives the team more space to explain.
Open-ended questions are particularly useful when we don’t know what we’re looking for yet. A customer might tell us that the checkout is frustrating. If we immediately ask if there are too many steps, we might get an answer about steps. But if we ask them to walk us through what happened, we might find that the issue was uncertainty about delivery date. Or whether a payment went through.
Direct questions have their place, too. If a team needs to know if a dependency will be ready by Friday, an open-ended conversation might lead to different interpretations. Sometimes, the most helpful thing we can do is ask for a specific answer.
I think about this often in product work. As I wrote in Not Every Problem is THE Problem, the first idea that sounds promising is not necessarily the most important problem to solve. We need enough openness to discover the problem, followed by enough precision to make a decision.
The tone and timing change the conversation
The words we use in a question are only part of it. The tone, order, setting all play a role in how people respond.
A leader can ask “Why did this happen?” with a genuine desire to learn. Or they can ask it in a way that makes the team feel like they need to defend themselves. It’s the same sentence. But the conversation that follows can be very different.
I’ve also found that a large meeting isn’t always the best place to hear an unformed concern. People might be working through the facts. They might not be sure they have the full picture. Or they might be hesitant to challenge a direction in front of a group. Starting with a more open conversation. Creating space for different perspectives. Following up individually when needed. This can help bring those concerns forward.
But that doesn’t mean we shouldn’t ask hard questions. Leaders have a responsibility to understand risks. To challenge assumptions. To be clear on expectations. But we need to pay attention to whether the way we ask makes it easier for people to share what they know.
The same is true when we’re answering. If we want a team to be open about uncertainty, we need to be willing to say when we don’t know something. A conversation is better when people have the space to change their minds as they learn.
AI makes the question more visible
One of the reasons I find AI so useful is that it gives me an immediate response to how I’ve framed a problem. If I ask for a list of product ideas, I’ll usually get one. If I ask for a recommendation based on a specific customer need, the evidence available, and the constraints we face, I’m more likely to get something I can work with.
But a more detailed prompt can also bring a more detailed set of assumptions. If I tell AI that a drop in engagement is a messaging problem, it might help me come up with a better messaging plan. But it might not establish whether messaging is the issue.
This is where I want to slow down. Before I ask for solutions, I can ask what other explanations we should consider. What evidence would help us distinguish between them? After I get an answer, I can ask what’s missing. I can check the underlying sources. The responsibility is still on us to decide what matters. To judge the evidence. To choose a direction.
There’s also a risk that we take the convenience of AI into our conversations with people. When we’re used to immediate answers, we might be less patient with the time it takes for someone to think. To explain. To work through a concern. Some of the most valuable conversations don’t follow a straight line. They need space for a follow-up. A pause. A change in direction.
Stay curious
Earlier in my career, I often felt like being prepared meant having an answer ready. But as I’ve gained experience, I’ve learned that being prepared also means knowing what I need to understand before I decide.
The challenge is that experience can make us faster at recognizing familiar patterns. This is valuable. But it can also make us stop looking. I wrote about this in The Patterns We Stop Seeing. I watched my son and his friends find clues in an escape room that I’d missed. They were willing to test possibilities that didn’t occur to me.
I think this is a useful reminder for all of us. The more we know, the easier it can be to assume we already understand the question.
AI will continue to make information and answers more accessible. That’s a real opportunity for us to learn faster. To consider possibilities we might otherwise miss. But it also makes curiosity, listening, judgment more important in our conversations with each other.
The question I want to keep asking myself is: Am I making enough space to learn something I didn’t already expect?
