AI as a Creative Teammate


Summary: 
AI acts as a creative teammate in group settings as it does for solo work. Thoughtful facilitation ensures teams leverage AI effectively in workshops and critiques.

I don’t need to tell you that generative AI can be useful for ideation, writing, or communication — you probably already know and believe that. In speaking with hundreds of UX practitioners about their uses of AI, it’s clear that many see it as a companion in their work. But what happens when we bring AI into group settings? Does it enhance or diminish the collaborative advantage? The results of a recent study shine some light on these questions.

AI Experiment with 776 Procter & Gamble Employees

A 2025 study by Fabrizio Dell’Acqua and colleagues compared the quality of ideas generated by 776 working professionals at Procter & Gamble who were assigned to work as individuals or in pairs (“teams”), some with AI and some without.

The quality of their ideas was rated by independent evaluators who did not know whether the solutions came from an individual or a team or whether AI had been involved.

The ideas generated by the four groups — individuals without AI, teams without AI, individuals with AI, and teams with AI — were given quality ratings by independent evaluators. The ideas generated by individuals without AI received the lowest quality ratings and were used as a baseline. The chart shows the relative increase in solution quality for the different groups relative to this baseline. The individuals and teams with AI received the highest ratings. The ideas from teams without AI were rated higher than the ideas from individuals without AI. (Adapted from the original article by Dell’Acqua and colleagues)

There were a few interesting results when comparing the 4 conditions.

Individuals vs. Teams Without AI

On average, when neither teams nor individuals were using AI, the teams produced higher-quality ideas than the individuals. This isn’t surprising since we already know that groups hold a strong advantage over individuals.

Individuals with AI vs. Individuals or Teams Without AI

On average, individuals working with AI produced higher-quality ideas than those coming from either individuals or pairs with no AI help.  This finding suggests that even working alone with AI can replicate many of the benefits of working with a group — something you’ve likely experienced.

Individuals with AI vs. Teams with AI

On average, the teams working with AI outperformed the individuals working with AI, but only barely.

The End of Team Ideation?

Does this third finding suggest that perhaps the era of workshops is ending?

From this data and my own observations, individuals working with AI can often outperform individuals or teams who do not use AI and can fiercely compete with teams using AI. In this study, the individuals working with AI were the fastest to come up with their solutions and used far more detail than individuals and teams who did not use AI, and just about as much detail as the AI teams.

But organizations do not implement all ideas; they implement the best ideas. The AI-powered teams generated a disproportionately high number of the top 10% highest-quality ideas in the study. In fact, while the individuals working with AI had impressive ideas on average, even AI-absent teams had more ideas in the top 10% percentile. There may be something powerful in a team, after all, that AI seems capable of enhancing.

Bar graph showing quality of top 10% solutions. Teams with AI produce the highest-quality solutions, outperforming all other conditions.
This figure shows the number of solutions in the top 10% percentile of solution quality for each of the four ideation approaches. This is where teams with AI really excelled, outperforming individuals with AI as well as individuals and teams without AI.

How Do Teams Collaboratively Use AI?

AI has likely been useful for you personally, but I’ll bet you haven’t yet experienced a well-facilitated workshop that leveraged it. This isn’t surprising since most LLMs are built for individual use. To effectively leverage the power of AI while working with a team, AI cannot be treated simply as a tool. We need to think of it as a teammate. This requires intentional facilitation approaches that differ from those used in traditional workshops or from individual AI use.

You can integrate AI into any kind of workshop, discussion, or critique you facilitate. However, it’s safest to rely on AI for activities like ideation, where “truth” is not the primary criterion for success (as opposed to something like data analysis) because of the ever-present hallucinations inherent in LLMs.

Next in the Series

Over a series of articles, I’ll share tips based on my experience facilitating both in-person and remote groups using AI. In the next article of this series, I’ll provide my recommendations for preparation steps that will help you integrate generative AI into your next workshop. In the third and final article, I’ll share my strategies for facilitating AI-based workshops.

Reference

Fabrizio Dell’Acqua, Charles Ayoubi, Hila Lifshitz-Assaf, Raffaella Sadun, Ethan Mollick, E., Lilach Mollick Yi Han, Jeff Goldman, Hari Nair, Stew Taub, Karim R. Lakhani, 2025. The cybernetic teammate: A field experiment on generative AI reshaping teamwork and expertise (March 28, 2025). Harvard Business School Strategy Unit Working Paper No. 25-043, Harvard Business School Technology & Operations Mgt. Unit Working Paper No. 25-043, Harvard Business Working Paper No. No. 25-043, The Wharton School Research Paper, Available at SSRN: https://ssrn.com/abstract=5188231 or http://dx.doi.org/10.2139/ssrn.5188231