FORWARDED THIS? Subscribe here to get the next one.

How to redesign the junior comp role so it still teaches

I spoke at a conference back in September and someone asked what happens to P1s and P2s comp roles when AI takes over their work, given that the work is how they learn to operate at a higher level.

My answer on the day was that the pipeline started narrowing when CompTech arrived, long before AI. Teams needed fewer people handling spreadsheets, so fewer P1 seats opened up. That part is already happening. In the US occupations most exposed to AI, entry level roles fell from 29% of job postings to 10%, according to Indeed's Hiring Lab.

Most of the companies I've worked with barely hired P1s anyway, outside junior engineering and admin roles. What I can't answer yet is where your P3s come from in three years if AI takes the P1 and P2 work, or whether next year's headcount plan still counts roles AI has already half absorbed. I've listed a few more of these questions at the bottom of this week's page.

The second problem gets far less attention. Whoever is still in the seat learns less, because the groundwork that used to teach them is the first thing AI takes. Job matching and survey submissions were never glamorous, but they were where most of us learned how jobs actually differ from their titles. We used to hand juniors an annual compensation report review or survey submission and hope they'd learn something, and now we'll have to teach them on purpose.

I don't think this is the end of junior roles. Senior people will keep handing work down. But the role needs redesigning, so I had a go at one.

I took a typical P1 comp analyst's month and sorted every task into one of three columns. Some go to AI now, and the analyst checks the output instead of producing it. Some stay by hand for at least the first year, because that's where the judgement gets built. The rest get redesigned so AI does the first pass and the analyst does something with it that a senior then pushes back on.

Before you open it, try this. Pull up last week's work for the most junior person on your team and mark which tasks taught them something. It takes five minutes and it's slightly uncomfortable.

Which tasks would you move to a different column? Hit reply and let me know. I'm expecting arguments about job matching.

ON THE RADAR

  1. Graduates in the majors most exposed to AI now start on 13% lower pay. US Census Bureau working paper, September. For the most AI-exposed tenth of college majors, the odds of landing a first job fell 5 percentage points after ChatGPT arrived, and first-quarter earnings fell 13%. About half of that is lower pay inside the industries still hiring them, and the rest comes from graduates drifting into lower paid sectors like retail. It's US only and not peer reviewed yet. Read → the Census paper

  2. Anthropic modelled what AI could do to knowledge workers' pay by 2030. Published September. Three scenarios, ranging from wages for cognitive workers 0.4% above a no-AI path to 11.5% below it, and unemployment for the same group anywhere from 2.9% to 17.9%. It's a model built on assumptions, and it comes from an AI lab (the one behind the tool I build most things with). Read → the Anthropic scenarios

  3. Fed governor named wage pass through as an AI risk. Governor Lisa Cook described a downside where productivity gains aren't passed on to wages because workers have little bargaining power. She also said recent graduates are finding first jobs harder, possibly partly because AI can take on entry level tasks. Read → the FED governor’s speech

FROM THE COMMUNITY

The first Working Groups started last week. Six teams of three to five members, from the US West Coast all the way through Dubai and to Australia, are spending six weeks building something real together, ending in a demo day in November. They're warming up on a pay equity analysis done with AI in a way that the employee file never leaves anyone's own laptop, which spares you from having to run it past IT.

Midpoint Ltd pay equity analysis hint

I deliberately planted a few rotten seeds in the inputs and methodology for the warm up assignment. While AI can do the analysis and processing for us, it’s a good reminder that it doesn’t replace human discernment.

If you’re keen on finding out more about what members are building as part of the working groups, I’ll share more updates over the upcoming weeks. You’re welcome to follow along.

RECENT EDITIONS

P.S. Know someone who manages a junior comp analyst? Forward this to them. The table's free on the same page.

Heads up: this newsletter will start going out on Wednesdays instead of Sundays.

Grab the free resources
Everything free we've made now lives in one place, the Comp AI Library. The guides, the prompts and the frameworks, all in the open.
Browse the Comp AI Library →

How blocked are you from using AI at work?
Find out in 2 minutes. The permission gap is an ongoing survey that not only gives you a score. It tells you where you sit against other comp and benefits teams that have already participated to the study.
How do I compare against other total rewards teams? →

Join other 2,000+ compensation and total rewards professionals. Get Range Weekly in your own inbox every week.

Subscribe free

Keep Reading