FORWARDED THIS? Subscribe here to get the next one.
ISSUE 10 · AI + COMPENSATION
Where AI has already moved into your pay decisions
TL;DR
I'm astonished by how quickly AI is being embedded into compensation processes compared to a year ago, despite how insulated and high stakes the domain is. What matters now is being able to explain a number on the day someone asks, and with pay transparency rules rolling out globally, that day will come sooner than you expect.
A NOTE FROM GIAC
I'm writing this three hours before boarding my flight from London to Madrid, this being the last leg of a relocation that's been going on since June. I cannot express enough how excited I am for this new chapter, but I am feeling so nostalgic if I think about what a city that I started to feel less fit in has given me over the past 8 years.
Packing has been unexpectedly useful for thinking in AI applied to Compensation, a far fetched analogy that surprisingly holds. Just like any comp team embedding AI into their processes now, every box I packed for the relocation forces the same question, which is whether I can actually justify keeping what's inside it, or if I should let go of it (mostly, charity donations).
That's close enough to what this week's edition is about.
Giac
THE BREAKDOWN
AI has already moved into three parts of your comp calendar
In house comp teams have been expediting technology advancement at a ridiculous high speed compared to the past 25 years of underinvestment in technology where just 5 years ago most teams were still running merit cycles on tech that came out in 1985 (your good old friend Excel).
Before the decision. Role matching, market pricing, scenario modelling.
During the decision. Budget allocation, manager recommendations, equity checks running while the numbers are still moving.
After the decision. Audit trails, compliance reporting, and the note that explains to one employee why their number is their number.
Most of us have only really noticed the first, because it is the work we would hand to a comp analyst first. The second and third are harder to hand to anyone, human or virtual.
The risk here has very little to do with AI itself. It is comp teams rushing to build or rent tools before they've firmed up their messy job architecture, cleaned up their data, agreed what their company's pay philosophy is, or decided who signs off what.
Many of the AI challenges I hear from comp teams are actually governance problems in disguise. But the curtain is being pulled on them now because AI has the potential to accelerate a weak process as happily as a strong one.
This has been the recurring story over my in house career years. People trying to assign multiple owners to single decisions or processes. I actually had someone suggest they would decide who we benchmarked against, which percentile we targeted and what budget went in front of the RemCo, and I would own execution only. As in, I decide, you do for me.
What holds up over the next few years is being able to meaningfully classify and price jobs in a way in which it will robustly stand up to reporting scrutiny for roles of equivalent value. And pay transparency legislation shouldn't be the main reason for that to be worth building now.
Four checks before a number leaves your hands
So I've written them up as a set of instructions you attach once, either as a skill you load or an .MD file you upload and then forget about.
It runs on any figure. A benchmark going to a hiring manager, a band about to be published in a job ad, anything an AI tool handed you that you're about to pass on.
The four checks go after the source, the stability, the edges and the stakes, in that order.
It ends on a verdict of trust, verify, or do not ship, plus one specific 10 minute fix. It takes about 4 minutes to run and it has stopped me sending a confident number I couldn't stand up more than once.
 |
ON THE RADAR
Only one in five companies are ready to hand a process over to AI agents. Deloitte surveyed 501 US senior managers and executives. 73% expect around half their business processes to be redesigned around AI agents within four years, and 61% expect those agents to run largely autonomously with people supervising. Read → the findings
Samsung trained its leaders on AI before it trained everyone else. Around 50 affiliate presidents went through hands on training first, and one team has since reported cutting the time it takes to produce a report by 40%. Read → the account
BCG puts the share of companies seeing real financial return from AI at 5%. The case made here is that organisations buy AI as a quantitative technology and then measure it with engagement scores and adoption percentages, which is why finance keeps sending the business case back. Read → the argument
RECENT EDITIONS
When you're ready, here are two ways Range can help you.
1. Grab the free resources
Everything free we've made now lives in one place, the Comp AI Vault. The guides, the prompt library and the frameworks, all in the open. Browse the Comp AI Vault →
2. Join the community
Find your people. Inside Range, reward pros learn to build with AI together, the prompts, the setups, the band workflows that actually hold up. Free is the signal, the room is the substance. Apply to join →
Reading this because a friend forwarded it? Get Range Weekly in your own inbox every Sunday. Subscribe free |
 That's me for this week. I'm mostly offline next week while a Madrid flat turns into furniture. Until next week, Giac | |
|
 |