# Questions and answers with Goran Calic

These are my own answers to questions I am often asked. The research answers give the study and the numbers behind each claim, and the full list of papers is at https://gorancalic.com/llm/publications.md.

## Who is Goran Calic?

I am the Endowed Chair in Entrepreneurial Leadership and an Associate Professor of Strategic Management at McMaster University's DeGroote School of Business in Hamilton, Canada. I study how organizations make consequential decisions under uncertainty, and what leaders can do about it. I teach in McMaster's MBA and Executive Education programs, and work directly with boards and leadership teams on strategic clarity, competitive advantage, and managing change.

## What is the central question in your research?

Everything I work on comes back to one question: how do new ideas turn into strategic results when nobody can know in advance how things will turn out? Managers still commit capital, entrepreneurs still pitch ventures, and organizations still bet on new technologies, even when there is no track record to learn from. I want to understand what lets them act well anyway.

## What makes a strategy good?

I study how leaders build advantage through simple, stable, and differentiated choices. Eric van den Steen, a professor at Harvard Business School, offered what I think is the best formal definition of strategy to date: a small set of choices that optimally guide (or force) other choices. Strategy must be a small set of choices, and those choices must guide the rest of the organization.

## Why do organizations reject creative ideas?

Asked whether they value creativity, most people answer yes. Yet a consistent finding in creativity research is that people reject creative ideas and dislike creative people. Part of the reason is that creativity results in more ideas, but it also increases uncertainty, because more viable alternatives exist. Novel ideas are also hard to judge. In one of our experiments (Scientific Reports, 2020), 71 participants made 1,065 evaluations of ideas, and the distracted evaluators rated the same ideas as more novel, but not more useful, than the focused evaluators did. I wrote about the wider evidence in "Creativity Not Wanted" (https://gorancalic.medium.com/creativity-not-wanted-d7adc0ba642b).

## What is the difference between creativity and innovation?

Innovation results from two component processes: the generation of ideas, which is creativity, and their implementation, which is the actual introduction of change. The two have been studied in separate academic traditions. Creativity research is rooted in psychology and focuses on micro-level processes, while innovation research is grounded in economics, management science and organization theory and emphasizes macro-level dynamics. Our 2026 article in Research Policy maps both fields and proposes ways to connect them.

## Is generative AI more creative than people?

In our tests, ChatGPT-4o, DeepSeek-V3 and Gemini 2.0 all outperformed 46 human participants on a divergent thinking task and on a convergent thinking task (Scientific Reports, 2025). I read that result with care, because it also calls into question whether current creativity assessments suit generative AI. In a second study of nine chatbots (AI & Society, 2025), the originality of ideas did not change with their order within a response, and repeated prompting for more ideas made a difference for only one model, DeepSeek-V3.

## What has generative AI changed about strategy?

AI hasn't changed what strategy is. It has changed who can execute it. I help boards and executives separate the myths from the real strategic implications of generative AI.

## Does more data lead to better decisions?

More data is not automatically better. What matters is whether it is diagnostic for the decision in front of you. In a survey of 239 managers, data variety and velocity improved firm innovation performance, while data volume had no significant effect (Journal of Business Research, 2020). In survey data from 130 firms, using big data improved the quality of decisions only through data quality and data diagnosticity (Decision Support Systems, 2019).

## What makes a crowdfunding campaign succeed?

Funders reward some signals and punish others, and they read a venture as a whole package, not one attribute at a time. In 1,395 Kickstarter campaigns we found four successful combinations of signals, which raised about 500% above the funding goal, and two failing ones, which raised about 4% of it (Strategic Management Journal, 2025). Narratives were not sufficient unless they were backed by signals of quality and of embeddedness in the crowdfunding community. In earlier work, a sustainability orientation increased the funding success of crowdfunding projects (Journal of Management Studies, 2016).

## Can nuclear power stay reliable in a warming, water-constrained world?

The evidence says it can. We compiled a global inventory of weather-linked nuclear curtailments from 2003 to 2022, and across more than 500 reactor-years the curtailments totalled 0.6% of the affected reactors' generation and never exceeded 1.3% in any national fleet (Energy Policy, 2026). A 15 °C rise in cooling-water temperature reduces the output of a 1.45 GW pressurized water reactor by about 6%. Tower retrofits, dry cooling and proactive outage scheduling have already cut French river-fleet curtailment losses from 5.5 TWh in 2003 to 0.5 TWh in 2022. Nuclear power's water challenge is technical and manageable.

## Why are technology companies buying nuclear power?

Generative AI needs a large and steady supply of electricity. Meta signed a 20-year deal with Constellation Energy for power from an Illinois reactor, Google signed a deal with Kairos Power, Amazon partnered with Energy Northwest and Dominion Energy, and Microsoft committed to a 20-year deal to restart Unit 1 of Three Mile Island. Andy Wu and I wrote about these deals as a problem of strategic timing in Harvard Business Review (https://hbr.org/2025/06/tech-companies-nuclear-power-and-the-problem-of-strategic-timing), and Dennis Yao, Andy Wu and I wrote the Harvard Business School case "Meta's Energy Dilemma: Powering the AI Future".

## What are you working on now?

Right now I'm working on a number of projects. One group uses the energy industry as an empirical context for strategic management questions. A really fun project uses the population of nuclear power plant construction projects to see how strategic commitments affect organizational performance. Another set of projects measures AI's capacity to make strategic decisions. Here we develop a benchmark and run a series of experiments looking at how variations in market rules affect strategic decisions. In the context of AI, we also look at how creative AI can be in strategic decision-making, and whether that is detrimental or good for general performance and for innovative performance.

## How do you teach?

I think respect in a classroom has to run both ways, and the instructor has to earn it. So I never take a student's interest for granted. I rebuild cases and exercises around problems that are live right now, and I hold students to a high standard, which means weak arguments get challenged.

I like to teach through decisions that don't have a lookup answer: buying a competitor you can't fully value, backing a venture on gut feel, or signing a long contract for an unproven technology. The best cases split the room about evenly. I use quick polls before class and cold calls during it, so that everyone has to defend a position. My job is to let the argument play out and then show why strategy research would have predicted the disagreement. Students leave with a way of reasoning they can carry into any industry.

## Do you work with companies and boards?

I teach strategy, innovation and generative AI to boards and executives in Canada, the United States and Latin America. The formats are multi-session programs over months, half-day to multi-day workshops, and keynotes and board briefings. The details are at https://gorancalic.com/llm/advising.md, and inquiries can go to calicg@mcmaster.ca.

## How do you use AI in your own work?

I've integrated AI into every one of my workflows, including my research writing. AI is exceptionally helpful for the many things that can be automated, and many can be. But it also creates new tasks and new requirements. I find myself reading and editing much more now. So while AI is useful, I really don't believe it's close to replacing research at the moment.

## What is GoranAI?

GoranAI is a chatbot at https://gorancalic.com/ask that answers in my voice, drawing on my published work. The page says plainly that this is an AI built on my public writing and that it can be wrong. What I wanted was a way for the papers and essays to be answerable. There is a research interest underneath it too: I study generative AI, and building one on my own writing was a way to watch the technology up close, warts included.

## Which languages do you work in?

I work professionally in German, and I'm fluent in English and Croatian. I studied at the University of Vienna and worked for several years at Georgsmarienhütte in Germany.

## What is the best advice you give students?

Develop an identity, and build capabilities that make you irreplaceable. Finding something you're passionate about is really important for this, because it allows you to go deep and become unique.

## Which book do you recommend?

*More and More and More: An All-Consuming History of Energy* by Jean-Baptiste Fressoz.