Three Things People Analytics World in Zurich Taught Us About Where HR Is Actually Heading

Written by Jessica Phillips, CEO and Founder of PowerUp.

I spent a few days at the end of February at People Analytics World in Zurich, in a room with roughly 70 People Analytics leads from across Europe and beyond. The sessions covered everything from agentic AI governance to predictive attrition modelling.

Here are three themes that I think deserve serious attention from anyone leading a People function or a People Analytics team right now.

1. Experimentation is everywhere. Strategy is not.

One statistic from the conference stopped the room: 39% of organisations cite a lack of clear AI strategy as the single biggest barrier to delivering value in HR (Insight222).

This is worth sitting with for a moment, because the absence of strategy doesn't mean the absence of activity. Quite the opposite. What multiple speakers, including Microsoft themselves, and many of us are seeing across most organisations is what you might call Stage 1 adoption - AI being used to smooth personal productivity: drafting communications, summarising meetings, generating first-draft job descriptions. Useful, certainly. Transformative, not yet.

Genuinely mature adoption - where AI agents operate autonomously across end-to-end workflows, making or informing decisions in real time - remains rare (less than 1% potentially). The leap between these stages isn't primarily technical. It's organisational and, fundamentally, philosophical.

The question most HR functions haven't yet answered isn't "what can AI do?" It's the harder one: what should humans remain uniquely responsible for?

This distinction matters enormously. When AI is layered on top of existing processes without redesigning those processes, we risk overloading staff, not enabling them.

Microsoft Switzerland's COO, Ann Jameson, spoke publicly about this concern: organisations are adding AI-assisted tasks to people's existing workloads rather than rethinking the work itself. The result is a workforce that is simultaneously more technologically capable and more cognitively strained.

The implication for People Analytics leaders is clear: before building the next model or deploying the next tool, the more valuable investment may be in facilitating a clear organisational conversation about where human judgement is genuinely irreplaceable - and designing AI adoption around that answer rather than despite it.

2. The era of "nice-to-know" analytics is over.

The Insight222 research presented at the conference contained a finding that should concentrate minds: only 11% of People Analytics teams are currently classified as top performers. Eleven percent.

What separates the leading tier from the rest isn't team size, budget, or technology stack. It's the powerful capability of connecting people data directly to business outcomes. Among top-performing teams, 88% consistently measure the financial or business impact of their work. 90% can demonstrate a direct link between their insights and improvements in business performance. Among lower-performing teams, these figures drop to a fraction of that.

One example from a European global retailer illustrated this: their analytics team found that stores with higher employee engagement scores had a higher average basket size than those in lower-engagement stores ($48 vs $40). This commercial insight is exactly what positions a People leader as a strategic value add to any board or executive meeting.

This is the shift that defines the difference between a People Analytics function that is consulted and one that is consequential.

The teams stuck on what conference speakers affectionately called the "reporting hamster wheel" - producing headcount updates, absence rates, and turnover figures on demand - are doing necessary work, but they are not building influence. The teams that have broken out of that cycle are asking a different set of questions: What decisions does our leadership need to make? What would they need to believe to make them differently? And what evidence could change that?

For HR leaders, the practical implication is to audit your current analytics output against a simple test: if you stopped producing it tomorrow, would a business decision get worse? If the honest answer is no, it may be worth redirecting that effort toward something that would. Getting strategy right matters, but strategy without demonstrable impact is still just a plan.

3. Data quality is the bottleneck.

It came up in virtually every session and the consensus in the room was clear: poor data quality is the single largest constraint in People Analytics today, and AI is making it more urgent, not less.

The principle is straightforward: AI does not correct bad data. It amplifies it. In a field where those outputs may inform decisions about hiring, retention, or workforce investment, "confidently wrong" carries real organisational risk.

Leading teams at the conference were taking a notably unglamorous approach to this problem. They are investing in data infrastructure before they touch advanced analytics. They are consolidating records into a single system of truth rather than managing multiple disconnected platforms that each tell a slightly different story. And they are applying rigorous data quality testing as a precondition for any modelling work.

The organisations making the most progress with AI-powered analytics are the ones that did the less visible, less celebrated work of getting their data house in order first.

For HR leaders, the question worth asking your analytics team is not "what can our data tell us?" but "how much do we actually trust our data?" The gap between those two questions is often where the real work begins.

A note on the next generation

One conversation surfaced in a few sessions concerns the development pipeline for the next generation of talent. If AI progressively absorbs entry-level tasks, the traditional entry-level model for building skills faces a genuine structural challenge.

But there's a more interesting dynamic at play here too. Younger professionals entering the world of work today often arrive with a native fluency that many senior practitioners are still actively developing. The "skills gap" in most organisations is usually discussed as a deficit among junior staff. It may, within a relatively short period, look quite different - with newer entrants holding capabilities that are in higher demand than the more established skills of those managing them.

How organisations navigate that inversion - without either undervaluing experience or underutilising new capability - strikes me as one of the more interesting leadership questions of the next few years.

Final thoughts

The opportunity for People Analytics in 2026 is in developing the kind of intelligence - human and artificial, working in genuine collaboration - that helps organisations understand why things are happening, not just what is happening, and intervene at the point where it actually matters. Moving from measuring symptoms to diagnosing causes is where the real winners will live.

If anyone has any thoughts on the above, I'd welcome any reactions! (Or if anyone in Spain can point me toward a decent Laugencroissant, please do get in touch. Zurich has left its mark...)