Great discussion. I am not sure AI disrupts the principles of performance review anymore then excel or PowerPoint did. Fundamentally, most organisations compete through some form of differentiation- mainly price or premium positioning. So the job of the performance cycle is to align the outcomes of people’s activities with delivering that and provide feedback on how effective that activity was. Performance reviews should always distinguish between being effective and being active, whatever tools are being used to enhance performance.
Very good points! But maybe it is easier now to be "active" in terms of generating output with AI models? And if that is true, the role of the manager(s) is increasingly important to make that distinction?
Yes, that may be true. I think the issue is that people lend too much intelligence to outputs of LLMs. So they end up sounding like many management consultants, lots of great words and theory but absent of any context or tacit knowledge. There will be many lost hours on the amount of noise that people will create using LLMs, due to the ease in creating artefacts of corporate theatre.
I think we need to counter the vendor hype and educate people on what an LLM actually is and importantly isn’t. This is a role HR can play in how we design work and integrate AI into job design.
An interesting take but I think the performance standard has changed and that is what we haven’t recalibrated yet in terms of expected output. Our job descriptions haven’t changed (enough), our leveling matrixes haven’t changed etc but the expected output on our teams has quadrupled if not more. Specifically on how the expected turn around time for closing PRs or doing code reviews. We are shipping more than ever and in turn, our expectation of what success is in the age of AI has increased but we haven’t set that standard yet because we can’t get anyone to agree on the ROI of AI use yet from a finance standpoint but we don’t want to give up the power it gives our teams. I could write an essay on this 😂
Great discussion. I am not sure AI disrupts the principles of performance review anymore then excel or PowerPoint did. Fundamentally, most organisations compete through some form of differentiation- mainly price or premium positioning. So the job of the performance cycle is to align the outcomes of people’s activities with delivering that and provide feedback on how effective that activity was. Performance reviews should always distinguish between being effective and being active, whatever tools are being used to enhance performance.
Very good points! But maybe it is easier now to be "active" in terms of generating output with AI models? And if that is true, the role of the manager(s) is increasingly important to make that distinction?
Yes, that may be true. I think the issue is that people lend too much intelligence to outputs of LLMs. So they end up sounding like many management consultants, lots of great words and theory but absent of any context or tacit knowledge. There will be many lost hours on the amount of noise that people will create using LLMs, due to the ease in creating artefacts of corporate theatre.
100% agree with that take!
But how do we "combat" that?
I think we need to counter the vendor hype and educate people on what an LLM actually is and importantly isn’t. This is a role HR can play in how we design work and integrate AI into job design.
Yepp.
Can I also squeeze in that we probably should re-think performance reviews anyways? Feels outdated...haha!
An interesting take but I think the performance standard has changed and that is what we haven’t recalibrated yet in terms of expected output. Our job descriptions haven’t changed (enough), our leveling matrixes haven’t changed etc but the expected output on our teams has quadrupled if not more. Specifically on how the expected turn around time for closing PRs or doing code reviews. We are shipping more than ever and in turn, our expectation of what success is in the age of AI has increased but we haven’t set that standard yet because we can’t get anyone to agree on the ROI of AI use yet from a finance standpoint but we don’t want to give up the power it gives our teams. I could write an essay on this 😂