What Does Good Performance Even Mean Now?

What the performance review misses now

Welcome to FullStack HR, and an extra welcome to the 33 people who have signed up since last week.

If you haven’t yet subscribed, join the 11300+ smart, curious HR folks by subscribing here:

---

Happy Tuesday,

Talking about it out in the void still, but the podcast is getting more and more listeners! Two episodes on there that haven’t been published here.
One about why your AI consultants need to be practical and another one about whether we are in an AI bubble or not. My aim is to keep this up during the summer as well. Very much liking the format so far!

Listen on Spotify here.
Listen on Apple Podcaster here.

But let’s get on with today’s article that is a bit of a thought experiment. This is not a finished model, and I don’t have the answers. It’s more of an idea and a bit of a call to action that we need to rethink this.

Let’s get to it.

---

I just listened to a podcast with Nate, whose Substack I highly recommend.
(Thanks Tue for the recommendation!)

Nate didn’t talk about performance reviews directly, he talked about what you as an employee need to showcase in this AI era to prove your value and your worth. And that connects right back into the performance part of things for me.

We’ve been measuring tasks for 70 years

Performance reviews have been around for what, 70 years? I wasn’t part of the initial phases here because I’m not that old, but for as long as I’ve been in the workforce, they’ve looked pretty much the same.

Sure, there have been various attempts at doing it slightly differently, but the basis of it has stayed the same. I’ve been fortunate enough to work with forward-thinking companies, so the review was maybe named something else, but if you look at what we did it was a standard process where you got evaluated once a year against a goal or some value the company had. And at its core it was fairly task-based.

Whenever I’ve been part of setting up different performance review structures, the team and I did our best to make them as objective and as reliable as humanly possible. But we also know that humans are poor raters in general. We have recency bias, we have tons of other biases, which all also play a part in this. But no matter how we twisted and turned this, most managers still landed in the tasks.

”How many tasks did you achieve in the past year, did you reach the projects you set out to reach, did you deliver what you said you’d deliver.”

That’s my experience working with I don’t know how many organizations and managers over the years, and I think it’s roughly the same right now.

What happens when the task runs without you

Now what’s happening is that AI can supercharge that process, so you can take on more projects, get through more tasks and be more productive. And when I talk about it like that it sounds great, we get more stuff done. But then there’s the Nate conversation about human judgment, and how do you as an employee show that you’ve used your judgment in the process.

Because if the AI can do the task, let’s assume it can, then what are we rewarding? Take my bookkeeping. It’s more or less fully automated right now, I have an agent gathering all the receipts and another uploading them. The task runs without me, but it still requires me. I’m the one who has to make sense of it all. Does my manager know I’m still needed? Do we as an employer even understand what’s going on in that space anymore?

I think we need to steer away from that mindset of tasks being the output we value people on.

I know a lot of companies aspire to evaluate against their values instead, but as said I do think we still fall back on tasks to a large extent. So how do we move away from it, and where does the human factor come into play? How do we know there’s a human in the loop at all? And do we even need to attribute the right things to the human and the right things to the AI, or should we just reward the human for using the AI in the first place?

Using AI vs using it well

Because there’s a difference between rewarding someone for using AI and rewarding them for using it well. It’s fairly easy today to open ChatGPT or Claude or Gemini or Copilot and get a task done, and I’d guess most of you reading this already have that skill. But I’ve been in plenty of discussions with managers where someone used AI and created more work further down the line. The task got delivered, the output got a thumbs up, it did the thing it was supposed to do, and then it created a ripple through the organization afterward that someone else had to deal with.

I think this requires more from managers, not less.

It’s easy to think AI will make your life easier as a manager, that you can just let it do stuff for you, which you of course can, but leading people who work with AI takes more conscious effort, not less. Because you’re the one who has to make sure the people you lead are not just using the tools, but using them in the right way to get valid outputs, and that they validate what the AI gives them.

The CHRO who saw it coming

One of the first proper AI conversations I had was with a CHRO, well over three years ago now. She’d trained her whole HR team early on, which is how we met. A year later she called me and said she had two people who used to be standard performers before AI, and now they were superstars, using it in everything and getting so much done. And she had two people who used to be high performers, not using AI and falling behind. It was performance review time, so what was she supposed to do, how was she supposed to attribute her pool of money across these people.

Back then I didn’t push as hard on the value question as I would today. We talked about output and whether they were providing value, but it wasn’t as focused on value as I’d be now. We landed on her wanting to reward the behavior of using AI, that was the answer at the time. Looking back, that was probably an early signal of what’s coming even more now.

So what is the review even for?

I think it’s worth reopening the whole thing and asking what the performance review is even for. Usually the answer is that it’s the one structured moment where an employee gets to talk to their manager somewhere during the year and gets some help to grow. But is that the best vehicle for it even now? Should we let AI support that growth together with the manager, should it be a combination of the two, or should we scrap the annual review altogether and build something else? Or is it even enough for now to say that as long as you’re part of the development process and using AI, that’s good enough?

I don’t have the answers here. I think different organizations will land in very different places on this, and that’s fine. But is it worth having the conversation? 100%. What does good performance look like in the age of AI?

I don’t have a clean answer for you, I just think we need to stop assuming the old one still fits.