AI Hasn't Fixed Teamwork

By 2025, artificial intelligence has become as normal in our offices as email. We use it to write code, draft emails, and summarise meetings. But if you feel that your team is still struggling with the same old problems such as misunderstandings, dropped tasks, and confusion over who is doing what, you are not alone. A fascinating new study published this month, titled “AI Hasn’t Fixed Teamwork, But It Shifted Collaborative Culture”, confirms exactly this suspicion. The researchers observed a software organisation from the early days of the AI boom in 2023 through to 2025. Their findings are a wake-up call for founders and managers everywhere.

The AI Paradox: Faster Individuals, Confused Teams

The study found that, although AI tools made individuals significantly faster at tasks such as coding and writing, they did not address the underlying issues of collaboration. The core challenges of teamwork, such as accountability and fragile communication, remained largely unresolved.

AI tools are excellent at getting things done. They can generate a report in seconds. But they cannot tell you who should generate that report, why it matters, or who needs to approve it. The research highlights that, despite our hopes that AI would act as an ‘intelligent coordinator’ to align our projects, it has mostly remained a tool for personal productivity. The result is a team of individuals working hard in isolation, often without a clear idea of how their efforts fit together.

The Missing Link: Role-Based Work Structures

This is where the concept of role-based work structures becomes crucial. While AI provides the speed, your organisational structure must provide the direction. In a traditional position-based company, you have a job title (e.g. ‘Marketing Manager’). This title is often broad and static. When work becomes complex, especially in the fast-paced world of 2025, it can be unclear whether the ‘Marketing Manager’ or the ‘Content Lead’ is responsible for the new AI newsletter. In a role-based organisation, people are not defined by their titles, but by the roles they own. A role is a clear, focused set of accountabilities. One person might hold five different roles. For example, you might have the ‘Social Media Publisher’ and ‘Event Planner’ roles. If the social media work becomes too much, you can hand that role to someone else without changing your job title.

The authors of the study noted that ‘ambiguity around accountability’ was a persistent issue that AI had failed to resolve. Role-based structures address this issue by posing one fundamental question: ‘Which role is accountable for this?’ When using a role-based structure, you are not just creating an organisational chart; you are building a dynamic map of accountability. This provides clarity, as everyone knows exactly what is expected of them. It also improves agility when priorities change (as they always do), and provides a clear path to address tensions.

AI Needs Structure to Thrive

The researchers found that, by 2025, efficiency had become the norm and the use of AI was a sign of professionalism. However, for a team to be truly professional, more than just fast tools are needed. It also needs clear rules of engagement. Adopting a role-based approach provides the ‘scaffolding’ that the research suggests is missing. You create an environment in which:

• Accountabilities are transparent.
• Overlaps are reduced.
• Resources are allocated to the right tasks.

We cannot automate our way out of bad management. As the study shows, technology alone cannot fix organisational culture or improve communication. If you want to build a team that will have a real impact in the AI era, start by clarifying your organisational structure. Foster a culture of clear accountability, ensuring that every team member really owns their role and purpose.