HR As Product Designers

Welcome to "AI@Work." In this course, you'll unpack insights from Iris McQuillan-Grace's Transform 2025 session, where she walked through how Oliver — a 5,000-person advertising agency operating across 46 countries — built an AI-powered feedback coaching bot that generated 19,000 feedback requests with a 97% completion rate, all without a single engineer on the project. Across three units, you'll practice applying human-centered design to find where AI belongs in People programs, mapping employee journeys to surface the real problem to solve, and structuring iterative pilots that ship fast and prove impact through data.

Let's start where Iris started: with the case that you're already equipped to lead this work — and the reframe that changes how you approach it.

You'll recall that Iris opened with a direct challenge to the room: "What if I told you everyone on this call is already a designer?" Her argument wasn't motivational fluff — it was a precise mapping of HR competencies onto the capabilities that human-centered design actually requires. As she put it, designing anything, "whether it's a bot or an app or an automation of some kind," demands "deep empathy, critical thinking, and the ability to effectively work across every facet of an organization." That's not an adjacent skill set for HR — it is the job description.

Building on this foundation, Iris urged People teams to adopt a product mindset: treating employees as consumers and HR programs as products to be "shipped, tested, and iterated" rather than policies to be pushed. This reframe matters because it shifts your orientation from compliance to experience — and from one-time launches to continuous improvement. As Iris emphasized, these aren't new capabilities to acquire; they're "just containers for things we are already doing."

So if you're already a designer with a product mindset, where do you start? Iris's answer was ruthlessly practical: identify your moments that matter across the employee experience, then prioritize the one that's most broken. At Oliver, that meant performance management — and the data made the case undeniable. With 46 countries running 46 different approaches, training reaching only 40% of employees, and HRBPs "coming to us very nicely with shivs in their teeth" about unsustainable coaching workloads, this was the moment where friction was highest, adoption was lowest, and manual effort simply couldn't close the gap. Iris described it plainly: "she was a trash fire and we needed to solve for X pretty quickly."

The lesson here isn't that performance management is always the right starting point — it's that the criteria for choosing are what matter. Highest friction, lowest adoption, inability of manual effort to scale. When you apply those filters honestly, one moment will stand out. And that's where AI earns its place — not because it's the shiniest tool, but because it solves a problem nothing else can.