Organizing Around Value, Not Technology: Melissa Perri on the AI Team Trap

In a recent LinkedIn post, product management expert Melissa Perri warns against a common organizational pitfall: building teams around new technologies, such as Artificial Intelligence, rather than customer outcomes. This approach, she argues, inevitably leads to silos and a disconnect from the very users the technology is meant to serve.

The Perils of Technology-Centric Teams

Perri likens the current excitement around AI to past organizational trends, such as the focus on platform, data, or mobile teams. She contends that when teams are structured around a specific technology, their primary objective becomes optimizing that technology, as that’s what they are measured on. This focus, she explains, causes them to drift away from the core customer problem.

“When you organize a team around a technical component, it optimizes that component, because that is what it owns and what it is measured on. It drifts from the customer problem, because no one on it is close enough to feel the drift.”

This, according to Perri, is the essence of the “build trap” – creating a polished product that ultimately no one needs or asked for. She emphasizes that this failure mode is consistent, regardless of the technology in question.

Organizing for Value Streams

Perri advocates for an alternative organizational structure, one that aligns with principles espoused by experts like Matthew Skelton in his work on Team Topologies. The core idea, as Perri explains, is to organize teams around the flow of value to the customer.

She elaborates on this by quoting Skelton’s approach, highlighting the importance of starting with the end-user. “Start with the value consumer, understand the value, then set up the organization so that value is easy to deliver,” Perri states, referencing Skelton’s perspective. Perri herself structures her teams similarly, prioritizing value streams over architectural components.

GitHub Copilot: A Case Study in Success

As a concrete example of getting it right, Perri points to GitHub Copilot. She recounts a conversation with Mario Rodriguez, GitHub’s CPO, where the focus was not on the AI model itself, but on product decisions that enhanced the developer experience.

“Ghost text instead of a popup. A 100-millisecond latency target to keep developers in flow. Those were product judgments from a team living inside the customer’s experience, not a model team optimizing accuracy off to the side.”

Perri argues that these decisions demonstrate a team deeply embedded in the user’s workflow, making product choices based on direct customer experience rather than abstract technical optimization. This, in her view, is the antithesis of the technology-centric silo.

The Guiding Principle: The Customer

In conclusion, Perri urges organizations to reconsider their team structures, especially as they embrace AI. She poses a critical question for leaders to consider:

“So before you build a team around AI, ask which value stream it belongs to. The technology is never the organizing principle. The customer is.”

Her overarching message is clear: successful product development hinges on organizing teams around the value they deliver to the customer, not the tools they use to deliver it. This principle, she stresses, remains constant even with the advent of powerful new technologies like AI.