Large language models (LLMs) are moving beyond just responding to prompts and are starting to act as conversational partners that ask their own questions to shape the discussion. Instead of only giving direct answers, many newer tools now: - Ask clarifying questions when a request is ambiguous. - Offer follow-up questions to deepen the conversation. - Suggest trending or relevant questions other users are asking. For example: - WaLLM, a WhatsApp-based chatbot, doesn’t just answer queries. It offers follow-up questions, lists of trending and recent queries, and even a “Top Question of the Day.” - OpenAI’s DeepResearch and tools like Manus are designed specifically to ask users questions to better understand their needs before responding. For business leaders, this shift means AI is no longer just a passive support tool. It can now: - Help structure conversations. - Surface issues you might not have thought to ask about. - Influence how decisions are framed by the questions it raises. This creates opportunities to reimagine how teams explore problems and options, but it also introduces new risks if leaders let the AI’s line of questioning drive the conversation without oversight.

Research comparing more than 1,600 executives with 13 leading large language models shows that AI systems tend to ask a different mix of questions than human leaders. Key pattern from the research: - AI models often **overemphasize interpretive analysis** (for example, “What does this data mean?” or “How should we understand this trend?”). - They often **underweight productive and subjective questions**, such as: - Productive: “Who will do what, by when?” “What resources do we need to execute?” - Subjective: “How will stakeholders feel about this?” “What are people’s concerns or motivations?” Because each type of question surfaces different information, this imbalance can: - Create blind spots around execution details and stakeholder impact. - Skew discussions toward analysis and away from action and ownership. - Subtly steer outcomes by focusing attention on certain dimensions of a problem and not others. On top of that, the research finds that **models vary widely from one another** in how they ask questions. So, the specific tool you choose can influence: - Which issues get airtime in meetings. - How risks and trade-offs are framed. - Which options feel more “natural” or well-developed. If managers simply follow the AI’s line of questioning, they risk: - Over-indexing on what the model is good at (interpretive analysis). - Underexploring what the model tends to neglect (execution, people, and politics). In short, AI-driven inquiry can be helpful, but if it goes unchecked, it can quietly reshape the decision-making agenda in ways that don’t match your organization’s real priorities.

Managers can get real value from AI that asks questions, as long as they stay intentional about how they use it and keep final judgment firmly in human hands. Practical guidelines: 1. **Treat AI questions as inputs, not instructions** Use the AI’s questions to broaden your thinking, but don’t let them define the full agenda. Ask yourself: - “What is this model not asking about?” - “Which stakeholders, risks, or execution details are missing?” 2. **Actively probe for underrepresented perspectives** Because many models underweight productive and subjective questions, deliberately add: - Execution-focused questions: “What would implementation look like?” “What capabilities do we need?” - Stakeholder-focused questions: “How will customers, employees, and partners react?” “Who gains, who loses?” 3. **Compare outputs across different AI tools** The research on 13 leading models shows they vary widely in their question patterns. To reduce bias from any single system: - Run the same prompt through more than one model when the stakes are high. - Compare the types of questions each model raises. - Use the differences to spot blind spots and refine your own inquiry. 4. **Be deliberate about when to use inquiry-driven AI** These tools are especially useful when: - You’re exploring a new problem and want to reimagine the space of options. - You need help structuring complex issues or surfacing alternative angles. They are less suitable as the primary driver when: - You’re making sensitive decisions with major people or political implications. - You need deep contextual knowledge of your organization’s culture, history, or constraints. 5. **Retain ownership of the decision process** Make it explicit in your team that: - AI is there to support thinking, not to replace leadership. - Human leaders are accountable for which questions matter most and which trade-offs to make. By using AI’s questions as a way to rethink how you explore problems—while consciously filling in the gaps around execution and stakeholders—you can benefit from these tools without letting them quietly take over the conversation.