What team compositions do you think work well with AI? What roles or functions are must-have on an AI-enabled team?

Different organizations may require different team compositions, but certain elements stand out. Someone must guide the AI strategy; simply encouraging teams to experiment isn’t enough. We promote experimentation and reduce roadblocks, but eventually, centralization is needed for cost and management purposes. Interdepartmental representation also plays a valuable role. Our AI platform team brings together data science and IT, involving stakeholders from across the organization. While all teams should participate in learning and process improvement, someone needs to steer higher-level corporate strategy.

In regulated industries, oversight is a major concern. Someone on the team should understand where the guardrails are and what controls exist within AI systems. Hallucination is always a risk, and while excitement about AI is widespread, understanding the boundaries—where it’s safe to use and where it isn’t—is necessary. A dedicated person or team ought to take ownership of this oversight.

Team compositions keep evolving as transformations continue. What I’ve observed is that everyone on the team needs to be able to use AI; this is a prerequisite. Those who understand how AI works and how to build with it are positioned on the engineering side. Currently, our teams include several AI engineers and a product manager, led by a director. The manager role is being phased out. Ongoing training remains a priority.