Turning AI skeptics into believers

Remember a year ago when the agentic AI bombshell entered the villa? It may not have been love at first sight, but agentic AI is changing the game. Talk of managing a team of agents has now come to the forefront of leadership and board meetings. And everyone is scrambling to find the imaginative use case that balances human and machine most effectively.

In our last piece, we laid out HR's big, self-perpetuating AI problem: Buy-in requires seeing AI transform real HR work. But that transformation requires the kind of deep, agentic use most of HR hasn't tried yet. Belief needs proof, and proof needs belief. That's the circular problem we left off with, and the one this piece is built to answer.

To solve it, we need to start building toward the reality we want to see for HR in the next two years. And, hint, hint, it has to do with agentic AI.

But first, let’s look at current use. How often and for what purposes are HR professionals using AI at work? And what do those common uses tell us about where HR could go?

The way HR uses AI hasn’t changed much

AI use is nearly unanimous, with 98% of HR professionals saying they’re using AI for something in 2026.

When we asked what they’re using it for, out of seven possible use cases, most people picked between four and five uses. Nearly everyone in our sample says they use AI for creating content (like onboarding materials, policies, emails, etc), brainstorming ideas (like strategy and best practices), and synthesizing information (like meeting notes).

So, the most-selected use cases involve basic task completion – same as what we found in 2025. Time-saving? Yes. Transformative? No. Most HR professionals are stuck in one-off, high-edit use cases, and that can easily make AI feel unremarkable rather than transformative.

There is a sizable minority, though, who break that pattern. These HR professionals have gone beyond asking AI to help them do a task to letting AI run a set of tasks for them. When we quantified patterns of use, two stuck out.

The two AI use cases that make a difference

We bucketed respondents by how many of the seven use cases they apply AI to. There's a gradual climb in reported benefit as people add more use cases, but the biggest jump comes from something more specific than volume.

Those using AI for all seven use cases report more meaningful impact across the board. We’re calling them the power users.

Power users have seen more AI transformation (by 19% points) and are more likely to be using AI to reshape roles (by 23% points) than peers who are using AI for 5-6 use cases. They also seem to be reaping a personal benefit. 84% of power users reported they’ve become a better leader with AI, which is 33% points higher than peers who are only using it for 5-6 things.

Those gains are not a result of simply doing more with AI. We think the leadership improvement stems from doing something differently with it. Those last two use cases – automating HR tasks (like HR Help Desk) and workflow support (like task nudging) – are examples of agentic AI usage. They show us that a sort of cliff forms around those who are letting AI act on its own instead of just assisting. So, we asked people directly how they felt about that.

Comfort with agentic AI is predictive of transformation

When we asked HR professionals directly whether they’re comfortable with agentic AI acting autonomously on their behalf, only 24% said yes. What’s more, a whopping 52% said no.

To understand what that comfort level actually predicts, we split respondents into two groups: those who agreed or strongly agreed, and everyone else – those who disagreed, strongly disagreed, or sat on the fence.

The comfortable group was more than twice as likely to report positive transformation of core HR processes, and outperformed the other group by 19 to 27 percentage points across every other measure of organization-level impact.

These big deltas show us that comfort with agentic AI is worth watching, especially because HR's fence-sitters (i.e., the group the preceding article called the most swayable) are looking for exactly this kind of evidence. So, what are the few who do use it actually deploying agents to do?

How are HR professionals using agentic AI?

While genuine agentic use is still rare in HR, the qualitative responses on our survey surfaced clusters of agentic work. This is what believers look like:

Sarah Ostrom, Global Shared Services Director, People Operations and Culture First Community member, built a 1-on-1 OKR tracker that grew into a culture-aware coordinator, catching things like regional celebration norms (For example, in the US, baby showers are common, while in Israel, it’s considered bad luck to celebrate early). She set the agent up to pull from Slack, email, calendars, and shared policy docs to function as an always-on coordinator. "The value isn't any single task," Sarah says. "It's having a system that keeps me ahead of things I would otherwise have to hold entirely in my head." Sarah’s deployment makes her whole AI tool context-aware.

Other leaders are reframing AI enablement altogether. Kate Sargent, Founder & CEO of KJS Coaching & Consulting and Culture First Community member, shared, "One of my partners and I built out an ‘AI fluency for Execs’ training that takes the idea of AI as tech out of the equation and balances language, leadership, and communication."

Others are using it more covertly, with results that speak for themselves. One anonymous commenter said, "I built my own executive coach that walks me through a weekly exercise. My team has noticed. My boss has noticed. They don't know that it's AI helping, but they see the outcome."

That said, most of what people call agentic is still a wish list. But leaders do appear to be stepping up to the agentic plate first.

Leaders be leadin’

Do the examples above sound familiar? HR executive leaders are the most comfortable with agentic AI, and the most bought-in that AI strategy should be owned by HR.

But just because they’re ahead doesn't mean they’re at ease. Despite leading on comfort with agentic use, and on conviction that HR should own how AI gets used, leaders are also the least confident it'll actually work out.

On the forward-looking measures (e.g., Will their org use more AI next year? Will it support roles rather than replace them? Is there a real plan to hand off transactional work?), leaders are up to 18% points less favorable than employees at other levels.

The people highest up the leadership ladder are the most cautious about whether or not it will pay off. That's either a warning sign, or it's exactly the kind of hesitation that comes from actually understanding the stakes. Are HR leaders worried about transformation not happening, or about what happens after it does? The qualitative data suggests it's worth taking seriously either way.

What HR actually wants AI transformation to look like

Across levels and regions, HR professionals describe the same aspiration with striking consistency: AI absorbs the administrative and operational weight so they can focus on the work that actually requires human judgment. They want to spend more time developing the strategic partnership, doing complex problem-solving, and having the conversations that can't be templated. It is a version of transformation that fundamentally alters the composition of many HR roles.

Notice that the vision is a far cry from AI replacing HR. It is also far from simply using AI to make HR faster at the same work. Instead, the vision centers around AI changing the composition of the job itself, shifting from transactional to relational and strategic experiences.

The same comments surfaced real concerns that complicate the vision somewhat. HR professionals pointed out that using AI in performance reviews risks encoded bias rather than corrected bias, and that using AI for decision support could become a crutch without critical-thinking discipline to match. The potential entry-level pipeline thinning today in ways that won't surface as a mid-level capability gap for another three to five years surfaced as well.

These concerns describe what pursuing agentic AI use responsibly actually requires and maybe why the leaders closest to the work are the least eager to celebrate early.

Next steps for HR leaders

The loop is simple: Skeptics need proof, and proof requires agentic use most of HR hasn't tried yet. That's what's kept HR stuck.

Leaders are closest to breaking through. They're the most comfortable with agentic AI, and the closest to the transformation everyone else is waiting to see.

To grow agentic AI use at your organization, start small.

1. Pick a single HR process and fully commit to automating it.
2. Run at it together rather than directing individuals to pick their own separate, smaller experiments.
3. As one team, rework the process from the ground up using an agent, so you can show (not tell) that AI can fundamentally change how HR works.

That's the whole mechanism: Leaders go first, visibly, and skeptics get the one thing that turns them into believers. Not a pitch or a vibe sesh. Proof.