HR tech CEO says becoming AI native revealed the work that employees actually want to do
See how one HR tech CEO used AI to remove admin work, reveal what employees actually value, and help HR teams focus on strategy, growth, and human judgment.
AI Adoption: Leapsome has integrated AI tools, with employees creating AI agents for automating various tasks.
Organizational Challenges: Early AI adoption exposed inefficiencies, revealing the potential to enhance roles by reducing administrative burdens.
Leadership Insights: Leaders should engage with AI to automate processes and maintain focus on strategic roles and human interactions.
Performance Overhaul: Leapsome revamped performance reviews using AI, enhancing feedback quality and reducing manager workload.
Data Integration: Seamless AI use requires connected data systems; disparate tools can hinder effective AI implementations.
Jenny Podewils is cofounder and co-CEO of Leapsome, an HR and AI platform used by over 2,000 companies globally. Her previous experience at a fast-growing clean-tech startup and as Chief of Staff in a rapidly changing industry led her to focus on a core question: How do we build organizations that align, learn quickly, and perform at a high level?
We caught up with her to find out more about how she's answering that question with AI. Here's what she had to say.
Designing to Learn Fast and Perform Well
I’m Jenny Podewils, cofounder and co-CEO of Leapsome. I grew up in Berlin, Germany, as the daughter of two academics with regular international guests and thought-provoking dinner conversations.
Keep Reading—and Keep Leading Smarter
Create a free account to finish this piece and join a community of forward-thinking leaders unlocking tools, playbooks, and insights for thriving in the age of AI.
Have an account? Log In
This shaped my interest in tackling big global challenges. During my university education in Switzerland and the UK, I discovered that impact-oriented businesses were the right way to make an impact.
My roles at a fast-growing clean-tech start-up and as Chief of Staff at a rapidly transforming media company hooked me on a question. How do we build organizations that align, learn quickly, and perform at a high level?
I became deeply interested in scaling alignment, designing environments that learn fast and perform well, and building better organizations grounded in real data. Humans spend roughly 80,000 hours of their lives at work. This time should be impactful. That belief became the foundation for Leapsome.
Kajetan von Armansperg and I founded Leapsome in 2016 to help others build high-performing teams, design exceptional employee experiences, and craft successful cultures using people-centric HR software.
Leapsome’s comprehensive HRIS, talent, and AI platform serves over 2,000 people-centric and successful customers globally. I am based in NYC, where we have one of our global offices.
As a leader, I always listen and ask the hard questions. I trust owners, and I am decisive about what I own. I have very high standards, and I value urgency and impact.
Making the Choice to Be Fully AI-native Internally
We embraced bringing AI's benefits to our customers early, shipping impactful AI features in our product quickly. We also made a deliberate choice to become fully AI-native internally.
We have built the vision, set the policies, and put the infrastructure and tooling in place early. One hundred percent of our team uses AI tools or agents in their daily work, and more than 60% of our employees have built at least simple agents to automate some tasks. So, AI usage and experimentation are part of our regular work and roles at Leapsome.
We continue to build AI into processes across the platform — from performance to surveys, or payroll to workflows. It’s exciting to see how we can support our customers to leverage AI impactfully, since it's directly accessible where their HR context “lives”.
How AI Adoption Reveals Hidden Opportunities at Work
When we rolled out AI tools internally and saw the adoption curve — our entire team uses AI regularly, and most build their own agents — I expected friction: some skepticism, some resistance from people who'd rather stick to what they knew.
Instead, I found a mixture of excitement and relief. People wanted to shed the work that accumulated around their actual jobs: initial drafting, summarizing, chasing, data-pulling, and formatting. When AI absorbed that, something opened up — people gained more time for fulfilling work that requires critical thinking and human judgment.
That told me something uncomfortable. We weren't just giving talented people boring work. We'd built systems that slowly buried the best parts of their roles under administrative scaffolding and most people quietly accepted this was just how things were. AI made that visible. Once visible, no one wanted to go back. The potential is huge, and it’s exciting!
The real lesson: adoption is the least of the problems. What you discover once the admin clears away is where the real work begins.
People wanted to shed the work that accumulated around their actual jobs: initial drafting, summarizing, chasing, data-pulling, and formatting. When AI absorbed that, something opened up — people gained more time for fulfilling work that requires critical thinking and human judgment.
How Leaders Can Lean into AI
It’s an important part of my role as a leader to dedicate time to my own continued education and hands-on experimentation with AI. I believe it is crucial for leaders today to carve out this time.
The opportunity for leaders and professionals in the People function is to lean into their role as a strategic partner to the business by automating manual processes, overcoming fragmented data, and generating insights faster. This also opens up capacity for the important people and business partnering work that should not be automated.
I've had to let go of the idea that "face time" with managers is the primary way employees grow. AI coaching, when done well, with human oversight, can provide consistent, timely guidance that even the best managers can't scale.
To be clear: that doesn't replace managers, it extends their reach and ensures employees get support between formal check-ins.
How We Overhauled Performance Reviews with AI
We overhauled performance reviews, both internally and in our product, with AI. It fundamentally changed how we think about the entire process.
The old way: Managers spent hours writing review drafts, often starting from scratch and struggling to recall specific accomplishments from months earlier. A cumbersome, slow process requiring heavy administrative lifting harms results and timelines! Reviews risk becoming a compliance exercise rather than an alignment and growth conversation.
The AI-augmented way: We built an AI assistant that drafts review feedback using context from meetings, goals, and prior feedback. The AI coaches users through the process, surfaces relevant context — what the employee accomplished, where they struggled, how they progressed — and suggests language managers can edit, refine, and personalize based on their human reasoning, empathy, and personal connection with direct reports.
Results: The median time managers spend writing reviews dropped significantly. More importantly, quality improved because managers edit and add nuance instead of staring at a blank page. Employees report that the feedback feels more specific and actionable.
We've applied the same logic to employee Q&A. HR fielded the same policy questions repeatedly via Slack, email, and sometimes from the same person asking more than once. Our AI Helpdesk agent now answers those questions in plain language, pulling from our knowledge base. HR regains that time for work requiring human expertise.
We also applied this to post-survey analysis. We used to spend weeks synthesizing open-text responses, often incompletely. AI now surfaces themes and patterns within days, so leaders can act while the feedback is fresh.
The common thread: AI absorbed the administrative load, so people could focus on work requiring human judgment, such as the conversations, the decisions, and follow-through that no one can delegate.
Why AI Potential Often Exceeds Organizational Readiness
People hear "AI will transform HR," and they expect a magic button. AI is only as good as the data, workflows, and governance around it. If your employee data lives in five disconnected systems, or your processes are informal and inconsistent, AI can't solve that.
Worse, it might surface the wrong data and cause, at best, confusion, and, at worst, a real issue.
We address this by designing AI embedded into our product as your system of record, with all your context tied to your processes and workflows — not bolted on as a separate feature.
Our AI coaches employees within the goal-setting process. It assists with reviews inside the review cycle. It answers policy questions from a defined knowledge base. The structure creates the guardrails, and the AI operates within them.
We also default to human-in-the-loop. Our AI drafts, suggests, and summarizes, but it doesn't take action without a person approving. We believe HR decisions involve context, nuance, and consequences that AI shouldn't handle autonomously. Leaders who want AI to "just do it" without oversight are setting themselves up for legal, ethical, and trust problems.
Going Deep Rather Than Wide with AI Integration
A year ago, we had a lot of tools. Now we have fewer, and they're more embedded. We've gone deep rather than wide. We use AI features throughout our workflows: review drafting, survey summaries, our AI-driven coach, policy Q&A, meeting transcription, and data insights.
These aren't separate AI add-ons, we build them into where the work happens, which ensures adoption.
Our foundation is our own tool, and being our own customer clarifies things. It keeps us honest. We use it for everything: HRIS (employee records, documents, absences, payroll prep, workflows) and talent management (performance reviews, goals, engagement surveys, feedback, learning, 1:1s) and AI. The data lives in one place, which matters more than it sounds.
Beyond that, we've built custom agents in Dust and Notion for cross-team use. Gemini, Claude, and Lovable are all in the mix, depending on the task. And every team has its own stack layered on top. We don't try to standardize everything, but we're intentional about what sits at the foundation.
The biggest shift over the last year has been consolidation. We had too many point solutions, each doing one thing well but creating data fragmentation and constant switching costs. Fewer tools, deeper integration, and AI embedded across workflows rather than siloed in one app.
Why an AI Development Coach Can Be a Game-changer
We use our own Development Coach internally. It's an AI agent that gives employees and managers real-time guidance on growth, without waiting for a formal review cycle or manager availability.
Managers at growing companies are stretched. They can't check in with everyone, and employees who need direction often don't get it until a review rolls around. The Development Coach changes that cadence.
Employees can ask questions about their goals, get specific next steps, and understand what "great" looks like in their role. Managers can use it to prepare for meaningful conversations instead of reactive ones. It draws on context from goals, feedback history, and company frameworks, so what it surfaces is specific, not generic.
I care most about what happens in the background: the signals. When employees ask the AI questions we'd never hear otherwise, we learn something — not about individuals (it's aggregated and anonymous) — but about where clarity lacks, where people feel stuck, what managers aren't addressing. That kind of early signal is hard to get any other way.
Agents go beyond work, too. At home, I have built several custom AI agents to make running a family easier. For example, I created a weekly meal planner that considers the healthy eating principles I set.
Why Peer-led and Practical is the Best Way to Teach AI Literacy
We made using AI structural, not optional. Every team member has access to AI tools, not just engineering or product. AI fluency can't be a privilege of certain roles or functions.
We built literacy through sharing, not mandatory training. People demo how they use AI in their workflows. It's peer-led, practical, and sometimes messy, but that's what makes it stick. Someone shows something that worked, and someone else adapts it for their context. That spread builds real adoption faster than any top-down rollout.
What I've noticed: the more people experiment, the more they find work they want to automate, tasks they'd come to accept as "just part of the job" for years. AI gives people permission and opportunity to question that.
On governance, we're explicit about data privacy. Everyone knows what data can and can't go into external AI tools. Non-negotiable. Clarity on the guardrails gives people confidence to push the limits within them.
As for what "AI-ready" looks like: I'd describe it less as a checklist and more as a culture. Employees who experiment without fear of judgment. Teams who ask "Should a person be doing this?"
Leaders who model AI usage visibly, not as a performance, but because they use it. I try to be explicit about that myself. If I'm asking my teams to build AI fluency and I'm not practicing it, that's not leadership. It's theater.
People demo how they use AI in their workflows. It’s peer-led, practical, and sometimes messy, but that’s what makes it stick.
Why AI is Only as Good as the Data It Accesses
Data access is key. AI is only as good as the data it accesses.
A shiny, standalone AI agent outside your HR system can write a job description, sure. It cannot tell you if engagement is dropping in a specific team, if a high performer hasn't had a development conversation in months, or if a manager's direct reports consistently underperform their own peer reviews.
This insight requires historical, connected data — HRIS, goals, feedback, reviews, engagement, and learning, all in one place. Fragmented point solutions, no matter how AI-powered, cannot deliver that. We've spent years building to solve that problem.
Our roadmap focuses on the agentic side. We build agents that handle bounded, specific tasks: drafting documents, answering policy questions, nudging managers on follow-through, while humans retain decision authority.
These agents run in the background. The goal is to extend what a lean HR team can accomplish. This allows an HR leader running people operations for 200 employees with one or two colleagues to stop spending their day on administrative tasks and start focusing on work that drives outcomes.
Why AI Forces a Reckoning About HR's Purpose
For HR leaders, this moment is more than a technology transition. AI forces a reckoning about HR's purpose.
A lot of people I talk to went into HR because they cared about people. They wanted to build cultures, develop leaders, and create environments where people could do their best work.
Somewhere along the way, many became process administrators. Systems accumulated. The role bent around them. If your HR technology values process over people, that dynamic shapes the role over time. Tools shape behavior. Systems shape roles. Worth naming, and worth disrupting.
Some practical pointers for using AI in HR:
- Start with the work you hate, not the shiniest feature. Where do you spend time on repetitive, low-value tasks? Apply AI there first. The first draft. The summary. The follow-up chase. Automate those tasks and protect your time for work that requires you.
- Default to human-in-the-loop. AI should assist, not decide, especially in HR, where decisions affect people's careers and livelihoods. Clearly set and uphold that expectation.
- Move faster than you're comfortable with. Organizations that build AI fluency now will gain compounding advantages. Waiting for AI to "mature" is a choice with a cost.
- Bring your team with you. Junior employees and frontline managers often see the clearest opportunities. Adoption is a culture challenge as much as a technology one.
- For leaders more broadly: Personally model AI usage, invest seriously in your managers, and keep purpose front and center. AI changes how work gets done. It doesn't change why it matters.
How HR Roles Will Evolve in the Next 5 Years
Within five years, I believe HR will fully shed its process administrator identity, and companies still treating it that way will feel it in their business results.
The shift is already underway. HR leaders at forward-thinking organizations aren't measured on whether reviews happened or surveys went out. They're measured on outcomes:
- Is retention improving where it matters most?
- Are managers developing their people?
- Can we trace a link between engagement, performance, and productivity to act on?
AI makes that shift possible by reliably handling the operational layer, so HR leaders can focus on strategy, diagnosis, and intervention.
The HR team of the future will be smaller, more senior, and significantly more influential. Not because headcount gets cut, but because the role expands in scope and earns credibility at the leadership table. Less time in the administrative layer, more time in the room when business decisions get made.
Organizations still running HR on fragmented systems, separate tools for HRIS, performance, engagement, and learning, will face a measurable disadvantage. AI generates useful insight from connected data. It generates noise from siloed data. Fragmentation is becoming an expensive legacy, and the reckoning is closer than most people expect.
The version of HR that wins in five years looks a lot more like what most people signed up for when they entered the field.
How to Keep AI Human-centered in HR
If HR implements AI poorly — deploying it to eliminate accountability, using it to scale bad decisions faster, or treating it as a substitute for the difficult human conversations that build cultures — we don't just waste the opportunity. We erode trust in ways that are hard to rebuild.
A malfunctioning AI isn't the main risk. The risk is AI that works exactly as designed but points in the wrong direction. Reviews that happen faster but feel less human. Policies that are "answered" by a chatbot without anyone checking whether the answer is accurate or fair. Decisions that appear objective because they came from a system, when the system reflects the assumptions of whoever built it.
Our philosophy is clear: AI should never replace human judgment on decisions that affect someone's career, livelihood, or well-being. AI drafts, suggests, surfaces. A person reviews, decides, and owns the outcome. Human judgment is non-negotiable.
The HR leaders I most respect aren't only asking "how much can we automate?" They're asking "what should never be automated, and why?" That discipline is what separates AI that builds trust from AI that erodes it.
It's also one of the most recurring conversations in People over Perks, our community for HR professionals — and honestly, some of the most grounded thinking on this topic comes from there. If it's a question you're working through in your own organization, I'd encourage you to join us.
As I said, people spend roughly 80,000 hours of their lives at work. That time should be meaningful. AI should help make it more so.
Follow along
You can follow Jenny Podewils' work on LinkedIn, and check out Leapsome!
More expert interviews to come on People Managing People!
