Ask most people what Satya Nadella did to turn Microsoft around and they'll probably tell you about Azure. About the huge bet of moving to the cloud, ultimately creating the pathway to all the AI investment now happening, and one of the main triggers for the share price that quietly went from afterthought to trillion-dollar juggernaut. All true. All also the wrong answer.


I was there watching it play out from afar. At the time, my role was with Bing - the scrappy underdog living inside an organisation that was used to being the gorilla in the room. And Bing was already, in a strange way, living in the future the rest of the company hadn't arrived at yet. While the enterprise side of Microsoft was still selling in three-year cycles mailing out shrink-wrapped discs of plastic - we were shipping features weekly, sometimes daily, watching the competition and responding in real time. Everywhere around the organisation were posters with a picture of a cloud and the phrase “we're all in," which, being British, always struck me as faintly absurd - all in what, exactly? But it was because what Ballmer and Kevin Turner were really asking the rest of the company to do was to fundamentally change their culture.


Then Nadella arrived with a new purpose statement – “to empower every person and organisation on the planet to achieve more” - which you could, if you wanted to be cynical about it, dismiss as nebulous. Achieve more of what? But what it actually did was force a shift from "we're the biggest, you'll buy our stuff because we think it's brilliant" to something more humble, better suited to a world where Amazon and Google were suddenly real competition. He gave that shift a name: growth mindset, borrowed openly from Carol Dweck's book, which more or less percolated through the entire company. It didn't stay inside Microsoft, either - I remember my son's primary school talking about growth mindset within a couple of years. And look at where we all are today, I doubt that there’s many people who don’t know what it means or how to do it.


What doesn't make the case studies, though, was the small, secret ingredient to making this huge change actually stick, they made a tiny change to how we were measured.


Before, you were assessed almost entirely on your contribution to your own team - fine in principle, except it had an ugly unintended side effect. If two Microsoft employees met for the first time, the unspoken first move was to quickly check all of your KPI’s against theirs: do our KPIs overlap? If they did, a conversation and likely lifelong friendship would ensue, if they didn’t it would be “lovely to meet you, but I never want to see or hear from you again.”  Because if working with you doesn't move my numbers, I can't justify the time, however useful you might be. It is, I believe, why the caricature of Microsoft was this company of silos where each silo secretly hated each other.  This was brought to life in the infamous cartoon from industry insider and cartoonist, Manu Cornet.


The fix wasn't a values poster,  or a set of principles in printed in 3 feet high vinyl letters behind reception.  Instead it was ten percent of your rating being tied to your contribution to Microsoft as a whole, not just your team - a change driven through by Kathleen Hogan, who'd go on to lead HR through the whole transformation. Suddenly I had explicit permission to walk into a building I had no reason to be in and say, "I think I might have something useful for you." The more interesting second-order effect was what it did to management. The old-school, “bums-on-seats” managers hated it - if you can't tell me exactly how to measure something, I can't manage it. The managers who thrived were the ones comfortable enough to say "I've got a rough calibration for this, and I can defend it to my own management" - which turns out to be exactly the managerial muscle effectiveness requires, because you can't measure impact the way you measure units shipped or licenses sold.


For a while I had a role built for exactly this: evangelism and strategy, translating between sales teams who needed language (and a smidge of charisma) that the material out of Redmond didn't give them, and technical teams who needed someone who could turn deep technical knowledge into something a business customer could understand and use. At the time, our Reading offices were five buildings joined by long corridors, each one taking a few minutes to traverse, and there were entire days where I'd go in, have maybe one meeting, and spend the rest of the day walking building to building having conversations - not sat in front of Outlook on a laptop, because what's the point of that? Schools for the future. Early mobile strategy, back when almost nobody had worked out what the iPhone actually meant. None of it showed up as productivity. All of it was the culture working exactly as it was designed to.


I think about that a lot right now, because HR leaders are standing at more or less the same door, except it's generative AI knocking this time instead of the cloud, and I suspect a lot of organisations are about to make the Ballmer-era mistake all over again: treat it as a rollout. Buy the licences, run the training, measure usage, call it done.


That's the technology-shaped version of the problem. The actual problem is human, and it's the one HR is uniquely placed to solve. Are people willing to be beginners again? Do they feel safe enough to say "I don't know how to do this yet" out loud, in front of their manager, without it costing them anything? None of that shows up on a usage dashboard.


The HR job was never to manage the human resource more efficiently. It's to deliver the right human experience to the people you actually care about - your customers and your employees - because that experience is what drives engagement, and engagement is the force multiplier that decides whether you ever see any benefit from transformation.


Which really means that the most important people in your organisation's AI transformation aren't the ones asking for it, and they're not the ones delivering it. They're the people creating the conditions in which it can happen. That's HR.


You've done this before. The question worth asking this week isn't "how do we roll out AI." It's "what's our growth mindset initiative this time round, and what's the ten percent that makes it real" - and who in your organisation is going to lead it the way HR did the last time everything changed at once.