Alonda Williams is the CEO of Big Brothers Big Sisters of Puget Sound.
Last year, a peer CEO told me she’d been using AI for eight months and still couldn’t explain to her board how she benefitted beyond saving time writing. She wasn’t an outlier. Surveys suggest that while many organizations are now using AI in some form, only a small fraction describe their use as strategic. That gap isn’t a technology problem. It isn’t a training problem. It is a leadership problem and it has a solution.
Most leaders are using AI to answer questions faster, to be more productive. They are in search of “getting time back” so that they can do other things. What they’re not doing is using AI to ask better questions. And I contend that the distinction turns out to be the difference between AI as a productivity tool and AI as a genuine leadership advantage.
The bottleneck isn’t the tool. It’s the quality of the question the leader brings to it.
I came to this observation from an unusual angle. I’m a nonprofit CEO who spent years in tech at firms like Microsoft, Verizon and Qualcomm before moving into mission-driven leadership. I’ve watched both worlds adopt AI with the same pattern: enormous excitement and investment in the capability, almost no investment in the human skill that determines whether the capability works. That skill is curiosity. Specifically: the ability to ask the question underneath the question.
AI removed three costs that made good questions too expensive
Before AI, certain questions were too costly to ask, not intellectually, but practically. Let’s say a leader is curious about whether they’re actually the right person to lead their organization through its current challenge. That question has always existed. Before AI, asking it required either a trusted advisor with full context and availability, or the willingness to voice a vulnerability that carries real professional risk. Most leaders skipped the question. Not because they didn’t know it mattered, but because the cost of asking was too high.
AI removed three costs simultaneously:
- The social cost. You can ask AI something you can’t ask a colleague without damaging the relationship.
- The time cost. No appointment constraint, no availability window, no relationship to manage.
- And the judgment cost. The inner critic that filters out the question before it ever gets asked. AI doesn’t react. It doesn’t protect the relationship. It doesn’t need you to perform confidence before it will engage. It just responds to what you actually ask.
That’s not a small thing. That’s a structural change in the cost of intellectual honesty. And most leaders haven’t figured out how to use it yet.
The risk hiding inside the tool
Here’s what makes this more than an optimistic argument: AI also creates a specific and underappreciated risk for leaders who don’t navigate it consciously. When AI can draft every communication, summarize every report, and generate every set of options, a leader can go an entire week generating outputs without doing any of their own thinking. The cognitive muscle of forming an initial opinion, sitting with ambiguity, arriving at a judgment through genuine discernment atrophies quietly when it isn’t used.
MIT Media Lab researchers recently coined a term for what happens to the brain under habitual AI use: cognitive debt. In a four-month study, participants who used ChatGPT to write essays showed significantly weaker neural engagement than those who wrote independently. Further, when AI users were later asked to write without the tool, their brains remained less activated, as if they had adapted to outsourcing the effort. The researchers described the dynamic as analogous to financial debt: AI assistance offers immediate benefits while potentially creating long-term costs.
The study was conducted with students writing essays. But the underlying dynamic maps directly onto leadership. The cognitive muscle for forming an initial opinion, sitting with ambiguity, arriving at a judgment through genuine discernment atrophies quietly when it isn’t used. As the lead researcher noted in the wake of the study going viral, the question isn’t whether to use AI, it’s how to use it to augment intelligence rather than replace it. That distinction is exactly what most leaders are currently missing.
I speak quite a bit to nonprofit audiences about AI. I’ve started asking leaders to audit their own AI usage. List the tasks you’ve used AI for in the last two weeks. Then ask yourself: which of those tasks required my judgment, my values, or my relationships and which ones were essentially processing work? The ratio is usually uncomfortable. Not because leaders are lazy, but because AI is so frictionless that the boundary between thinking and not thinking disappears without anyone noticing.
The leaders who will thrive in an AI-augmented world are not the ones who know the most about AI. They are the ones who bring the best questions to it. That’s a learnable skill. And it starts with understanding that AI’s highest use is not answering your questions faster, rather it is helping you discover which questions you should have been asking all along.
What this looks like in practice
The most powerful prompt I’ve found for this is also the simplest. Before asking AI to generate options or frameworks for a decision, ask this instead:
- “I’m going to describe the most important decision I’m facing right now. Don’t give me options or a framework. Tell me: what is the question I haven’t asked yet that would change how I’m thinking about this?”
Try this once. The question that comes back is almost always one the leader recognizes immediately and has been avoiding, sometimes for months.
A second prompt that reveals the risk rather than the opportunity:
- “Here are the tasks I’ve used AI for in the last two weeks: [list them]. Which of these required my judgment and values and which ones were processing work? What does the ratio tell me about how I’ve been using my own thinking?”
The list itself is often diagnostic before AI says anything. Leaders who struggle to name five tasks learn something. Leaders who easily name 20 learn something different.
Neither of these prompts teaches a leader to use AI better. They teach a leader to use themselves better and to recognize that AI is only as valuable as the quality of thought the leader brings to the conversation.
The leadership question AI cannot answer
There is a category of question that sits entirely outside AI’s capability. McKinsey’s research on leadership in the AI era put it clearly: only human leaders can determine why we work and what we’re trying to achieve. AI cannot answer those questions. It can surface them, challenge them, and hold them without judgment but the answers require a leader who is curious enough to ask and honest enough to sit with what comes back.
That’s the capacity worth developing. Not AI fluency, though yes, that matters. Not AI governance, yes that matters, too. The foundational skill is the willingness to bring your honest and most uncomfortable questions to a tool that will engage without flinching and to trust what comes back..
There is one more dimension to this shift that rarely gets named in the AI productivity conversation.
For most of history, the ability to ask hard questions with expert help attached has been a function of access. Access to advisors. Access to networks. Access to the rooms where those questions were welcome. That access was never evenly distributed — not across organizations, not across sectors, and not across communities.
AI changed that, too. The nonprofit leader without a strategy consultant on retainer, the first-generation executive without a mentor network, the leader of a small organization without a research team — they now have access to the same quality of thinking partnership as anyone else. The question that was previously too expensive to explore is now free for everyone.
That is not a small thing. That is the most significant democratization of intellectual access in a generation.
AI made curiosity free. The leaders who spot that first are going to use it in ways others won’t. And right now, there’s a huge opportunity for all of us, no matter sector, scale or role, to level up. By asking better questions, we can get more out of AI and deepen our own thinking and leadership.
Editor’s note: GeekWire publishes guest opinions to foster informed discussion and highlight a diversity of perspectives on issues shaping the tech and startup community. If you’re interested in submitting a guest column, email us at [email protected]. Submissions are reviewed by our editorial team for relevance and editorial standards.
