“With having performance, pay, and HRIS data in one place, the biggest win is definitely consolidation. Simply having to put in that data once has removed a whole category of mistakes, and also leads us to faster people decisions.”
How charles unified their people data and rebuilt their competency framework around AI-driven execution, with Leapsome
“With having performance, pay, and HRIS data in one place, the biggest win is definitely consolidation. Simply having to put in that data once has removed a whole category of mistakes, and also leads us to faster people decisions.”
charles is an AI-native conversational marketing and commerce platform that helps consumer brands attract, convert, and retain customers across messaging channels, primarily WhatsApp, using AI agents and marketing automations.
“charles is a conversational AI platform. We're based in Berlin, and we have about 48 employees. We help consumer brands convert, retain, and attract customers through messaging channels, primarily WhatsApp, with AI agents and marketing automations.”
Joy Vujic is charles' people lead, and at this size, the role is a function of one. She covers hiring, culture, performance, and compliance as a team of just one, so every strategic question about the workforce and every operational task that keeps it running land on the same desk, on the same day. Being a people lead at an AI company shapes how she works, too: she relies on AI for the automations, drafting, and analysis that would otherwise eat her week, specifically so she can spend more of it on the parts of the job that still need a person.
Before Leapsome, Joy ran her HRIS on Personio and handled performance and engagement work separately. Personio covered the HRIS basics well enough, but not the part of the people function charles needed most as it scaled: engagement surveys, a competency framework built to grow with headcount, and performance reviews tied to that same underlying data instead of living apart from it. Joy ran both systems side by side for a stretch before deciding one platform made more sense than two.
“We soon realized that Personio wasn't able to give us what we needed in terms of the people development and enablement side. So engagement surveys, a competency framework that actually would scale, and then the performance reviews that were connected to that. Leapsome was pretty much stronger in all of those, which is why we decided to then become the first HRIS customers of Leapsome, to consolidate all the data and only run with one platform going forward.”
The cost of running two systems wasn't abstract. Tenure, title, reporting line, performance data, and compensation data all depend on one another, so a single change- a promotion, a new manager, a comp adjustment- meant updating more than one place and checking that nothing had drifted out of sync in between. Consolidating onto Leapsome meant Joy could pull a report or make a people decision without first reconstructing which system held the current version of the truth.
“What makes Leapsome so special is most definitely the customer success and support team. They've been super responsive and supportive through every situation throughout the migration, which is a lot of work, especially working solo. That's definitely something that wasn't the case with our previous HRIS.”
With performance, pay, and HRIS data now living in one system, nothing about charles' underlying workforce changed. What changed is how much manual reconciliation used to sit between that data and a decision, and how many chances that reconciliation created for a small error, a title entered twice, a comp figure updated in one place but not the other, to become a wrong call on leveling or pay. Entering the data once, in one place, closed off most of that risk.
Joy uses the same consolidated data for compensation prep, comparing what candidates say they expect against what charles already pays internally, and pulling market benchmarks faster than she could when that data lived in two disconnected systems. The comp decision itself still sits with her, but the groundwork that used to take a day now takes an afternoon.
Since charles is an AI-native platform, the company is, of course, progressive and AI-forward. However, because the potential for ‘output’ per person is so high, the scarce thing now is the judgment to know when AI's output is correct, the verification to catch it when it isn't, and the accountability to own the decision either way. That distinction has reshaped hiring and reviews at charles.
“Output per person has definitely gone up; no one can deny that. But AI multiplies everything, also the bad stuff. What has become scarce is really judgment, verification, and accountability. That's ultimately changed how we hire and how we evaluate things.”
It carried straight into hiring first. A take-home assignment shows what a candidate can produce, often with AI's help, but not how they think, so charles now runs live working sessions and case studies where a candidate has to explain a decision as they make it, including whether they trusted the AI in front of them or checked its work.
“We rely a lot more now on live working sessions and case studies that require the candidates to explain themselves. Why did they make that decision? Why did they choose to go that route? While doing that, we can actually assess whether they are using AI responsibly or they're trusting it blindly.”
Joy also rebuilt leadership competency frameworks in Leapsome around that same principle. Task execution counts for less because AI can now do a version of it. Decision-making, communication, and developing other people count for more, since those are the parts of the job that still sit with a person.
“AI can take on a lot of the execution part. What really matters is how people show up for each other, how they make sound decisions. We've redesigned our leadership competency framework in Leapsome around that: less task execution, more decision making, communication, and people development.”
Reviews now also carry a simple self-rating on AI adoption, deliberately kept out of scoring and promotion decisions. Its only function is to get a manager and their report talking about what's actually been automated and where someone is still stuck, which Joy has seen turn into wider knowledge-sharing across teams.
Freeing up time from admin and reconciliation was never the end goal for Joy; it was what made room for the part of the job she considers non-negotiable. When someone leaves charles, she doesn't default to a like-for-like rehire. She runs exit interviews to understand what felt frustrating or redundant in the role, then checks with the team and manager whether the position still needs the same skill set now that AI can handle more of it than it could a year earlier. That's a judgment call made with people, not a process run on a system.
Some parts of the role were never about process to begin with, and Joy is deliberate about keeping them that way even as everything around them gets automated. She names specific moments in an employee's time at charles that she makes a point of not letting AI take over: a wedding gift, greeting someone in person on their first day, watching a match together as a team.
“The human layer shouldn't go missing. There are moments that cannot be replaced by AI: getting someone a wedding gift, saying hello on their first day in person, watching a World Cup game together. I try to really embrace those moments.”
🪄 Start transforming your workplace today
Explore how Leapsome can help your organization, too, unify your people data and build a competency framework fit for AI-powered teams.
👉 Chat with an expert
Give them a platform that saves time for HR — and builds growth for employees.
Get your personalized walkthrough — you’ll know in one call if Leapsome fits.
