Pinned Tweet
“AI won’t work for us. Our HR processes are too complex.”
I hear this a lot.
But complexity is exactly where AI becomes useful.
Different countries.
Different bonus rules.
Different exceptions.
Different employee populations.
AI shouldn’t force you into a standard process.
It should learn yours.
AI won't eliminate the need for compensation professionals.
But it should eliminate a lot of the repetitive work they do.
Building the same report again.
Recalculating budgets.
Finding the same pay data.
Answering the same manager questions.
I'd rather have HR spend that time on decisions that require experience and judgment.
33
I don't want AI deciding who deserves a raise.
I want AI doing the work that helps a manager make a better decision.
Modeling budgets.
Checking pay history.
Finding outliers.
Identifying potential inequities.
The recommendation can be automated.
Accountability cannot.
1
108
A better AI compensation prompt has 3 things:
1. A specific employee population.
2. A clearly defined compensation measure.
3. A request to explain the calculation.
"Show engineering managers in Austin whose base pay is below midpoint, and explain how you calculated it."
Specificity makes analysis easier to validate.
2
67
Your compensation team shouldn't need days to answer a question about merit spend.
Stello's AI Compensation Agent can analyze budgets, salary ranges, and employee compensation data using plain-English requests.
If your team is preparing for a merit cycle, DM me "AI" for a walkthrough.
2
6
104
Here's another prompt I'd use before closing a compensation cycle:
"Show employees who will remain below their salary range minimum after the proposed increases."
That tells you whether your merit budget is actually correcting pay problems or simply distributing increases.
3
123
Before approving a merit cycle, ask:
"Which high performers are receiving below-average increases?"
It's a simple question.
But answering it across hundreds or thousands of employees can reveal allocation decisions that deserve another look.
Pay equity analysis should happen before approvals, not after.
4
116
An AI agent that answers compensation questions needs to understand permissions.
A manager should see their own team's salaries.
A CHRO may need company-wide visibility.
The same question should return different data depending on who's asking.
Enterprise AI needs access controls, not just accurate answers.
1
4
104
Imagine a manager asking:
"If I give one employee a 6% raise, how much budget remains for the rest of my team?"
Today, that question often means opening another spreadsheet.
An AI agent connected to live compensation data can answer it directly.
Small improvements like this change the entire planning experience.
We've built an AI Compensation Agent at Stello called Iconic.
You can ask it to model merit budgets, apply compensation rules, and calculate the impact on your employee population.
No rebuilding spreadsheets for every scenario.
DM me "ICONIC" if you'd like to see it in action.
1
7
113
The best time to catch a compensation budget problem is before managers start planning.
Model the scenarios first.
What if the budget is 3%?
What if it's 4%?
What if promotions have a separate pool?
The earlier you test the assumptions, the fewer surprises you'll have later.
1
7
115
One AI prompt every compensation leader should try:
"Show me the total cost of a 3.5% merit budget, broken down by department."
Then ask:
"What changes if we give higher increases to top performers?"
This is where AI becomes useful for actual compensation decisions.
1
5
135
A general AI chatbot can explain what a merit matrix is.
An AI compensation agent should be able to answer:
"What happens to our total payroll if we increase the merit budget from 3.5% to 4%?"
The difference is access to your actual data and the ability to run the calculation.
1
5
164
After years of leading HR teams, one thing became clear to me:
The most expensive compensation mistakes often start as small operational problems.
An outdated salary range.
A missed market adjustment.
An inconsistent offer.
A delayed pay equity review.
Fixing them early is almost always easier than explaining them after someone resigns.
2
6
305
Before approving a sign-on bonus, I'd check two things:
Is the salary competitive?
And will this offer create pay inequity with existing employees?
At Stello, we help teams compare offers against market data and salary ranges.
DM me "OFFERS" if you're reviewing your hiring compensation strategy.
5
57
Exit interviews tell you why someone left.
Stay interviews tell you what might make someone remain.
I'd rather have the second conversation while there's still time to act.
Especially with high performers whose compensation hasn't kept pace with the market.
5
72
Giving every employee the same percentage raise feels simple.
But simplicity isn't necessarily fairness.
An underpaid high performer and an employee already above market may need very different decisions.
A good merit budget should reflect performance, market position, and business priorities.
3
56
One hiring mistake can create years of compensation problems.
You offer a new hire more base salary to close the deal.
Six months later, experienced employees discover they're earning less.
Now you have a retention problem.
A one-time sign-on bonus may have been the better solution.
2
3
101
One early warning sign of turnover:
New hires earning almost as much as experienced employees doing the same work.
Pay compression often develops gradually.
Then suddenly, your most experienced people realize they could earn more by leaving.
It's worth checking before the next hiring round.
3
3
102
Gallup estimates replacing an employee can cost 50% to 200% of their annual salary.
That changes how I think about compensation adjustments.
A $10,000 market correction can look expensive in a budget meeting.
Until you compare it with the cost of losing and replacing that person.
2
4
68
