The Best Thing AI Agents Give a CPO Isn't Efficiency. It's Time

That's a People problem wearing a technology costume. And what strikes me about it isn't the technology at all — it's the quiet admission underneath it. If the shape of every role is shifting, the person best placed to lead that is the CPO. Yet most People leaders I speak to are spending their week on work that has very little to do with people.

I sit on both sides of this. I run an AI recruitment business, and I do HR consulting as a CPO. So I get to watch what actually happens when a People function starts using agents properly, rather than talking about it in a strategy day.

The unglamorous truth about the CPO week

Ask any CPO where their time goes and you'll get a version of the same answer. Policy questions that have already been answered forty times. Chasing managers for review submissions. Rebuilding the same headcount report in a slightly different format because the board asked for it differently. Drafting the letter. Checking the letter. Sending the letter.

None of that is why anyone took the job.

The work that actually moves an organisation — the workforce design, the capability building, the honest conversation with a leader who isn't leading — is the work that gets pushed to the back of the week. Every time.

What agents are genuinely good at right now

This is where I'd caution against the vendor pitch. Agents are not yet a substitute for judgement. But they're very good at the layer sitting underneath it:

Answering the same question, well. A policy agent grounded in your actual handbook handles the leave, expense and entitlement questions that eat your team's inbox — and it doesn't get tired by Thursday.

Doing the first pass. Screening summaries, interview scheduling, drafting the first version of a position description, pulling the attrition numbers by team before you've asked twice.

Keeping the process moving. Onboarding checklists, review chase-ups, compliance reminders. The nagging nobody enjoys and everyone needs.

Handled well, that's not a small saving. It's the difference between a People function that reacts and one that plans.

The part CPOs shouldn't hand over

Here's where it gets interesting for our profession.

If agents become part of how work gets done, someone has to own the questions that follow. Who is accountable when an agent gets an employee decision wrong? What are we automating, and what have we deliberately chosen to keep human? How do we tell people honestly what these tools are doing with their data?

Those are People questions as much as technology ones. Governance, ethics and trust have always been part of our remit, and this is simply the newest place they show up. The best outcomes I've seen come from People and technology leaders answering these together, early — with the CPO in the room rather than briefed afterwards. I've written before about what an AI governance remit actually looks like, and it increasingly belongs right beside the People agenda.

Where I'd start

Pick the one recurring task in your week you most resent. Just one. Automate that, watch it properly for a month, and use what you learn to decide the next one.

And if the capability isn't in your team yet, you don't necessarily need a permanent hire to get moving — fractional AI leadership is how a lot of Australian organisations are taking the first step. If you're still working out which roles you're actually missing, my AI roles explained guide is a good place to start.

The point was never to make HR faster. It was to make it human again.

‍