Breaking Internal HR Silos

In the previous lesson, you explored why clarity beats false comfort during transformation and why modeling vulnerability is the leadership posture this moment demands. But even the most candid communication falls flat if the people team's own structure works against the change it's trying to lead. This lesson turns inward and unpacks what Inna Landman's team discovered about how HR's traditional silos, undefined foundational terms, and transaction-heavy legacy systems quietly undermine AI's potential before it ever reaches the workforce.

Landman described a real-time aha moment during a dinner with her team where several leaders each explained the AI solutions they were building:

Landman: "I'm going to create an AI solution for the manager to help with compensation conversations. And then I'm going to create an AI solution for managers to help with performance discussions. And I'm going to help a manager figure out how to deal with recruiting."

The group paused and recognized the trap: "we can't design a solution for a manager to go to five different places". Each Center of Excellence (COE) was solving its piece of the puzzle in isolation, replicating the very silos AI was supposed to dissolve.

This connects directly to Kyle Forrest's observation:

Forrest: "The HR function has had a history of oftentimes being accused of transforming for the sake of HR."

The fix isn't incremental — it requires putting the manager's end-to-end problem at the center of design, not the COE's ownership boundaries. Building on this, Landman described her HR team's internal AI hackathon as "one of my most proudest moments in my entire career", not because leadership directed the innovation, but because "the innovation was already in the room". Forty percent of her organization spent a week or two building real solutions with existing tools, proving that the creative energy to break silos already exists when you create the platform for it.

Even the most elegantly designed cross-COE solution will fail if the data underneath it is inconsistent. As the conversation explored, leaders need to "know how the technology works" well enough to understand what agentic AI — systems that can autonomously make recommendations and execute multi-step workflows — requires before it can deliver trustworthy recommendations. Landman called this "the unsexy work" aligning on the foundation: establishing precise, organization-wide definitions for terms like what constitutes a promotion or what counts as regrettable attrition. Without that foundation, agentic AI layered on top of inconsistent definitions produces outputs that different teams interpret differently, eroding the very trust the tool was supposed to build.

Compounding this challenge is the administrative burden that keeps the people team locked in transaction mode rather than shaping strategy:

Landman: "I would love my team to stop loading EIBs and all these things and start being strategic."

Landman was candid that relying on legacy HR providers to deliver this transformation may itself be a mistake:

Landman: "I have to figure out a different solution in order for me to have an actual agentic workflow for how we do work."

Leaders need to understand how the technology works well enough to recognize that this foundational alignment isn't optional. It's the prerequisite that makes everything else possible. It's not glamorous, but it's the difference between AI that the organization trusts and AI that creates new confusion.