Promoting the loudest voice in the room is easy. Rippling AI helped Ignite Reading find everyone else.

The Challenge

Sabrina Kidd is the Lead Director of People Operations at Ignite Reading, an education nonprofit. Twice a year — in January and July — the organization runs a promotion cycle. And twice a year, Sabrina faces the same structural challenge: the people most likely to be discussed aren't always the people who most deserve to be.

Part of the problem is the halo effect. Managers going into a promotion conversation often anchor on the most recent few months of someone's performance — what's fresh, what's visible, what's been presented in a staff meeting. The employee who hosted the all-hands last week. The person who introduced a new tool and got some attention for it. These are the names that surface naturally, even when they're not necessarily the strongest performers across the full review period.

The other part of the problem was structural. Ignite Reading's 360 performance evaluation cycles are staggered by department to avoid asking teams to take on heavy review work during their busiest quarters. That means performance data across the organization isn't all collected at the same time. Going into a promotion cycle, someone in People Ops would need to gather evaluation results, compensation history, tenure data, and last raise dates — across roughly 75 employees eligible in any given cycle — and stitch it all together by hand.

Doing it thoroughly enough to surface the right people, especially those whose contributions happen behind the scenes, would take up to two weeks. The risk wasn't just time. It was that the behind-the-curtain performers, the ones doing essential work that doesn't get applause in a meeting, would simply be left out of the conversation.

The Solution

Going into this promotion cycle, Sabrina took a different approach. She opened Rippling AI and submitted a single, precise prompt:

“Show me employees who ranked in the top 25% of our performance ratings for full-time salaried employees, but were in the bottom 50% for compensation for their role and tenure. Overlay trends such as tenure and last compensation adjustments.”

The prompt was designed to find a specific profile: high performers whose compensation hadn't kept pace with their contributions. In other words: employees who might be doing excellent work, quietly, without the recognition or pay to match and who, if nothing changed, might be a flight risk.

Rippling AI worked through the performance, compensation, employment type, and tenure fields simultaneously, cross-referencing data that would previously have required pulling from multiple separate reports and manually comparing the results. Within minutes, it returned a complete analysis.

Out of 150 full-time salaried employees, 19 met the criteria: top-quartile performers sitting in the bottom half for compensation relative to their role and tenure. Of those 19, ten had been with the organization for more than two years and hadn't received a raise in over eight months. Rippling AI also identified which departments had the highest concentrations of at-risk employees, flagged six individuals as particularly urgent based on their tenure and performance trajectory, and generated a set of recommended actions: conduct stay interviews, review compensation structures, ensure inclusion in the upcoming promotion cycle, and establish a more regular cadence for compensation reviews.

What made the result especially useful was the layering. The department-level breakdown gave Sabrina a way to cross-reference the findings against employee satisfaction survey data — asking not just who was underpaid, but how those specific departments were doing on workload, burnout, and capacity. The data didn't just answer the question she asked. It opened the next set of questions worth asking.

The Impact

The time savings were significant — what would have been a week to two weeks of manual analysis, tracking data across sources and building a picture employee by employee, was done in a single prompt. But for Sabrina, the more meaningful outcome was what the analysis made possible.

Promotion decisions at most organizations are driven by recency and visibility. The people who get promoted tend to be the people who are most top-of-mind. Rippling AI gave Ignite Reading a way to make those conversations evidence-based instead — to walk into the room not with a gut feeling but with a list of names, backed by data, representing employees who had earned a seat at the table but might not have gotten one otherwise.

The department-level breakdowns added another dimension of value. Knowing that customer experience and tutor operations had higher concentrations of underpaid high performers allowed Ignite Reading to prioritize proactively. The team connected that data to their employee satisfaction surveys before those employees reached the point of leaving.

One to two weeks of work reduced to minutes. Cross-referencing performance rankings, compensation levels, tenure, and raise history across 150 employees — a process that would have taken up to two weeks manually — was completed in a single prompt.

Evidence over instinct. Promotion conversations shifted from recency bias and visibility to data-backed identification — giving the quiet contributors the same shot as the employees who happened to present at the last all-hands.

Flight risk, identified proactively. 19 high performers in the bottom half for compensation were surfaced before the promotion cycle — with six flagged as most urgent, and recommended actions ready to act on immediately.

A foundation for equitable decisions. Department-level breakdowns gave Ignite Reading a starting point not just for this cycle, but for building a more systematic, ongoing approach to compensation equity — one that doesn't depend on someone having the time to do it all by hand.

For Sabrina, the point was never just efficiency. It was fairness. The people who do essential, invisible work deserve to be in the room. Now there's a way to make sure they are.