The phrase “Ask ChatGPT” (or Gemini, or Claude) has become commonplace, and people are increasingly relying on large-scale language models for answers and prescriptive advice. The same generative artificial intelligence technology helps leaders ask better questions, which often leads to more effective problem solving.
A Digital Fellow at the MIT Digital Economy Initiative, he has long supported the idea that individuals and teams can benefit from asking the right questions. Many of the questions were in quick succession. This method, which he calls “question bursting,” can open up new avenues for insight and impact, especially when people or teams get stuck.
Catalytic questions meet AI
In a recent webinar, Gregersen detailed how he transformed this largely analog exercise into a digital one by leveraging the speed and dexterity of LLM to enhance the human process of crafting high-impact questions. (Gregersen explores this approach in detail in his MIT Sloan Executive Education course, “Question Are the Answer: A Creative Approach to AI-Enhanced Investigation, Insight, and Impact.”)
Rather than leaving humans completely out of the picture, Gregersen said, question burst catalyst This tool shows how humans and AI can work together to enhance research and reverse the LLM trend of following the path of least resistance. This tool is available as a Gem in Gemini and as a Custom GPT in ChatGPT.
“LLM usually gives you an average meme-type response,” Gregersen said. “Question Burst Catalyst tools work with humans to maximize effectiveness.”
Questions are answers: A creative approach to AI-powered research
live online
Register now
How the question burst catalyst works
There are four steps associated with the Question Burst Catalyst tool.
- Define the issue. When a user describes a problem, the tool creates a more concise version of it and gives the user the option to accept or modify it. In a webinar demonstration, Peter Hurst, Senior Director of MIT Sloan Executive Education, posed a 49-word question about the role of AI and humans in organizations, summarizing the tool as, “How can we ensure that the adoption of AI does not crowd out investment in critical human, organizational, and business capabilities?”
- Humans think first and generate relevant questions. The tool prompts users to generate their own short, challenging questions in quick succession before the AI provides them. Gregersen said this order was a deliberate choice based on MIT IDE’s “human-first” approach to AI. Teams that outsource question generation to AI risk shrinking their own catalytic research capabilities, he said.
- Add personas. Users can assign personas to the AI to generate questions through a different lens, such as historical figures, fictional characters, professional archetypes, or even non-human perspectives. The goal is to steer the AI away from generic responses and towards more unexpected and provocative questions as the user continues the process of question bursting.
- Generate insights and action plans. The tool compiles and categorizes all questions by theme, prompting users to identify the one question most worth addressing. From there, action options are generated organized by type (exploratory or experimental) and risk level (low or high).
In our webinar example, a low-risk option would have been to interview five colleagues from different departments and ask them, “Which human capabilities will be made more valuable, not less, thanks to AI?” A riskier option was to reallocate some AI training budget for explicit development of human capabilities.
Actual usage example
The Question Burst Catalyst tool is designed to counter LLM’s tendency to return bland, unflattering results that tell users what they want to hear, Gregersen said.
This tool can be used by individuals to prepare for meetings, business teams to improve collaboration and drive better results, and other scenarios. It is being used at scale by the likes of Chanel, FM and Procter & Gamble, where administrators are piloting AI-enhanced tools across select leadership development programs within secure enterprise environments.
“Shifting the frame from answers to questions is a powerful thing, especially in this world of AI where we have very sophisticated answers and even more sophisticated questions,” Gregersen says.
Hal Gregersen I am a digital fellow of MIT Initiative on the Digital EconomySenior Lecturer and former Executive Director of Leadership and Innovation at MIT Sloan MIT Leadership CenterCo-Founder of Consulting Group Innovator DNA. He is the author of “.Questions Are Answers: A Breakthrough Approach to the Toughest Problems in Work and Life”, co-authored with Clayton Christensen and Jeff Dyer, “Innovator DNA: Master the five skills of a disruptive innovatorHe has written books such as ‘
Gregersen’s question burst catalyst The tool is available as gem of gemini and as a habit GPT for ChatGPT.
