
Can AI Replace a Colleague?

Introduction
The hype around AI is not new. But few have examined what happens when AI becomes not just a tool but an actual team member. A new field study from Harvard Business School and the Wharton School now offers clear answers. Its findings should prompt every organisation in a regulated industry to pause and consider how AI fundamentally affects our teams, skills development and ways of working.
In this blog, we explore what the experiment found, why it matters for compliance and legal teams, and what to do the next time you consider whether AI should “join the team”.
Background: A large-scale, real-world AI experiment
The study “The Cybernetic Teammate: A Field Experiment on Generative AI Reshaping Teamwork and Expertise” (Dell’Acqua et al., 2025) is based on a field study involving 776 professionals at Procter & Gamble (P&G). Participants worked on real product development innovation challenges and were assigned to four groups:
- Individual without AI
- Team without AI
- Individual with AI (GPT-4 via Azure)
- Team with AI
The aim was to evaluate AI’s role in three key areas of teamwork:
- Performance
- Sharing expertise
- Social engagement and motivation
Finding 1: AI can match, and sometimes exceed, team performance
One of the most striking findings was that individuals using AI performed as well as two-person teams without AI. In terms of quality, they even scored slightly higher, at 0.37 standard deviations above the baseline. Teams using AI performed best of all and were most likely to produce solutions in the top 10%.
“This means that AI is not just an aid; it can actually function as a ‘cybernetic teammate’.”
(Dell’Acqua et al., 2025)
For compliance and legal work, where complex tasks often require several perspectives, this means that AI can replace some aspects of collaboration and coordination without compromising quality.
Finding 2: AI breaks down functional silos
A traditional challenge in regulated organisations is siloed thinking: legal focuses on risk, business on growth and IT on systems. This was clearly visible in the study:
- Without AI: R&D experts developed technical solutions, while commercial staff produced marketing-driven proposals.
- With AI: Both groups produced balanced, cross-functional solutions.
AI enabled individuals to think beyond their own expertise. This is crucial for compliance teams, where traditional divisions of responsibility can hold back innovation.
Finding 3: AI creates more positive experiences
Contrary to some expectations, the study showed that AI not only improves quality but also increases job satisfaction:
- AI users reported more energy, enthusiasm and motivation.
- They also reported less frustration and stress.
In a compliance role, where work is often solitary, detailed and risk-sensitive, AI can therefore provide social support and mental relief.
What does this mean for compliance and legal teams?
1. Rethink team structures
AI allows certain types of work to move from a group to an individual without reducing quality. Organisations can therefore reduce meetings for analysis and decision support, simplify coordination between specialist functions and restructure workflows to become more agile. Teams do not always need to be large or cross-functional to deliver high quality; sometimes an individual supported by AI tools may be enough.
2. Build prompting skills
The study showed that even inexperienced users could perform well with AI. Training is likely to improve results further. Organisations should:
- Offer prompting training for compliance and legal teams
- Develop internal AI support for policy work, risk analysis and documentation
3. Integrate AI into collaboration platforms
AI is most effective when used within existing workflows. Integrating AI into tools such as Teams, SharePoint or document management creates:
- Better accessibility
- Higher adoption
- More natural collaboration between people and machines
4. Focus AI use on quality, not quantity
The study shows that AI does not necessarily save time, but it does produce better solutions. Organisations should therefore look beyond efficiency and set goals for higher-quality policies, analysis and documentation. By using AI to generate more cross-functional proposals and improve understanding of both risk and business, compliance work can become more strategic and valuable.
Conclusion: AI as a colleague, not just a tool
The study shows that generative AI can act as a cybernetic teammate: it helps you think more intelligently, more broadly and with greater engagement. It does not replace human judgement, but complements it in a way that makes you both more effective and happier.
For compliance and legal teams in regulated industries, where demands for collaboration, accuracy and efficiency are especially high, this is a wake-up call.
It is time to say:
“Yes, AI can join the team.”
Source:
Dell’Acqua, F., Ayoubi, C., Lifshitz, H., Sadun, R., Mollick, E., Mollick, L., Han, Y., Goldman, J., Nair, H., Taub, S., & Lakhani, K. R. (2025). The Cybernetic Teammate: A Field Experiment on Generative AI Reshaping Teamwork and Expertise. Harvard Business School Working Paper 25-043.
Get a demo
See how North House can enhance your organization's governance. Schedule a demo today!
Customer Support
North House provides governance solutions that adapt to your organization's evolving needs.


