I’ve always been curious about how AI might reshape the way we work. I came across an interesting study by researchers from Harvard and Wharton, done at Procter & Gamble (P&G), about generative AI in teamwork. Here’s what I found interesting:
What’s the Big Idea?
The study wanted to see if AI could act like a real team member. They tested how AI affected product development teams at P&G. These teams included technical folks from R&D and people from the commercial side.
They split 776 employees into four groups:
Team Setup | AI Used? |
---|---|
Individual | ❌ |
Team of Two | ❌ |
Individual | ✅ |
Team of Two | ✅ |
Everyone worked on real business problems, such as how to make people switch to new products or how to encourage new user habits.
Things that Stood Out
1. AI Boosts Performance (Even when you are working alone !)
The most surprising thing? People working alone with AI did just as well as two-person teams without AI. The AI didn’t just help; it practically replaced the need for a second teammate. People using AI wrote more detailed ideas faster than traditional teams.
Takeaway: With the right AI, you might be as productive alone as you’d usually be in a team.
2. AI Breaks Down Expert Silos
Normally, R&D people pitch technical solutions, while commercial folks suggest marketing-driven ideas. But AI helped everyone offer balanced ideas, mixing business and technical thinking. Even employees who rarely worked on product ideas before suddenly performed at expert levels thanks to AI.
Takeaway: AI might blur job boundaries, making cross-functional knowledge easier to use.
3. AI Feels Like a Teammate
People enjoyed working with AI. They reported more excitement and enthusiasm and felt less frustrated or anxious. It was like having a helpful colleague by your side, not just a search engine.
Takeaway: AI isn’t just efficient—it can genuinely improve the work experience.
Interesting Results
Metric | Individuals (No AI) | Teams (No AI) | Individuals (+AI) | Teams (+AI) |
Quality (Relative) | Baseline | +24% | +37% | +39% |
Time Spent (compared to baseline) | Baseline | Similar | -16% | -13% |
Emotional Impact | Mixed | Positive | Very Positive | Very Positive |
However,
Interestingly, people using AI weren’t always confident that their solutions were better. They were objectively stronger, yet they felt unsure. This means there’s still work to do in helping users trust their AI teammate.
Closing Thoughts
This study got me thinking about our teams. If AI helps individuals match team performance, how should we build future teams? Maybe smaller groups with strong AI support can move faster and more creatively.
Also, breaking down traditional expertise boundaries could change how we design roles. AI might enable our developers to think like marketers, and vice versa, creating more versatile teams.
Lastly, the emotional benefits of AI were a nice surprise. A tool that’s efficient and also improves people’s day-to-day happiness at work? What more could you wish for 🙂
What’s Next?
As we explore AI at work, let’s think about:
- Training our teams to interact smartly with AI.
- Reconsidering team structures, given how effective AI is for solo tasks.
- Learning how to trust AI outcomes confidently.
This isn’t about replacing people. It’s about enhancing what we can do. I’m excited about what comes next.
What are your thoughts on AI and teamwork?
References
And yes, to make sure this blog made sense, I had help from my AI friend.