From Pilots to Performance: What HR Must Get Right to Scale GenAI

A growing number of HR teams are exploring the use of GenAI. But for some, the early excitement is starting to turn into frustration. Tools are being tested, pilots are underway, but the business impact remains limited.

The issue is not lack of interest, it is lack of readiness.

In recent conversations with some of my business partners on GenAI strategy and implementation, I wrote down five ideas which stood out for me as particularly relevant to HR. They help explain why some organisations are starting to generate real value from GenAI, while others are stuck in perpetual experimentation.

1. Readiness, not adoption, is what separates leaders from the rest

GenAI can no longer be considered a “novelty”. The real question is whether the organisation, and the HR function in particular, is ready to embed it into daily operations.

This means that, more than running pilots, HR needs to have the operating model, data foundations, use-case alignment, and trust mechanisms in place to make GenAI part of how the function delivers impact.

The smart HR teams are asking themselves: How prepared are we to scale this? What gets in the way? And what’s the roadmap?

2. Capability building is now the primary bottleneck

One of the clearest items to address is that people, not technology is the issue here.

For HR to lead AI adoption across the organization, it needs to show within the function that we can walk to the talk, and take responsibility for the skills uplift that makes it viable. This includes:

• Confidence in using GenAI tools across the employee lifecycle

• Understanding AI's strengths and limits

• Working alongside data teams to co-design experiments that matter

Upskilling can’t be limited to tech teams. It must include HR, line leaders, and employees. And that uplift needs to be built into transformation plans from the start.

3. Governance needs to come from HR, not just IT or Legal

AI governance is often treated as a compliance issue. It shouldn’t be. It’s a design issue: what will get built, what will get automated, and what should be controlled.

HR needs to be involved in:

• Setting responsible use policies

• Framing ethical guidelines for workforce applications

• Creating feedback loops to learn from early deployments

If AI is going to be used for hiring, performance management, learning, or pay, then HR must have a say in how it should and shouldn’t work.

4. From pilots to systems: impact only comes from scale

Most HR functions are running disconnected pilots. Some in recruitment. Some others in learning.

Yet more in performance management. But most of these are not being scaled, integrated, or assessed against business value.

The difference between activity and impact is whether these tools are embedded in how the function works. That requires a systemic view of how these fit together, rather than a disconnected series of experiments.

To make real progress, organizations must build repeatable infrastructure: common data layers, integrated workflows, and clear criteria for scaling what works.

5. HR must manage GenAI as a portfolio of strategic initiatives

The last point is about mapping AI initiatives to value improvement, not just process improvement.

This means aligning GenAI to:

• Workforce productivity goals

• Capability-building priorities

• Talent risk areas

• Key business transformations

• Retention and engagement of key staff

Closing 

Treating GenAI as an innovation side projects is unlikely to drastically change the game. Instead, HR needs to manage GenAI deployment it as a strategic portfolio.

The performance gap between organisations that are scaling GenAI and those that aren’t is likely to widen quickly. What will make the biggest difference is the ability to move from pilots to use case to system, from tool to capability, and from experimentation to accountability.

For HR, that means leading on readiness, skills, governance, and alignment.

It’s time to move from experimenting and start building.