Latest / Deep Tech Germany – AI, Robotics & Frontier Innovation by Startuprad.io™ / E 779 — Leapsome: Agentic AI Built on 10 Years of Data

Transcript

E 779 — Leapsome: Agentic AI Built on 10 Years of Data · Deep Tech Germany – AI, Robotics & Frontier Innovation by Startuprad.io™

0:00
Speaker1: Ten years, five of them without a single year of venture capital.
0:04
Speaker1: Reportedly $10 million in annual recurring revenue before the first outside
0:10
Speaker1: money, then $16 million in one round.
0:13
Speaker1: A core product rebuilt from scratch in 2024 in the middle of the AI boom.
0:19
Speaker1: And today, AI agents running HR for more than 2,000 organizations, while
0:25
Speaker1: new public data says one HR person can support 299 employees at the frontier against a median of 53.
0:35
Speaker1: This is a story of how agentic HR actually gets built and why the agents are
0:42
Speaker1: the last step, not the first.
0:54
Speaker1: Hello and welcome everybody. This is StartupRate.io episode 779.
1:00
Speaker1: Startups, tech and venture capital from Germany, Austria and Switzerland.
1:04
Speaker1: I'm Joe Manager joining you from Frankfurter Main. My guest today is Jenny von
1:09
Speaker1: Porter-Wils, co-founder and co-CEO of Leapsom,
1:12
Speaker1: one of the few founders in Europe who bootstrapped a software company for five
1:17
Speaker1: years, raised exactly one round, 60 million US dollars led by inside partners
1:22
Speaker1: and then rebuild the entire product from the ground up before putting AI agents on top.
1:29
Speaker1: Since July, those agents run core HR programs for more than 2,000 organizations,
1:37
Speaker1: including Notion, Spotify, and Mercedes-Benz.
1:41
Speaker1: And this week, Leapsom published research on how the 100 fastest-growing tech
1:46
Speaker1: companies actually run HR
1:49
Speaker1: my thesis today agentic HR is one on data foundations not on agents let's test it
1:57
Speaker1: Jenny welcome back to episode 779 exactly three years after we did together episode da da da da da 401
2:06
Speaker0: Yes, it's great to be here. Thank you for having me.
2:09
Speaker1: Do you feel like 400 episodes older?
2:13
Speaker0: I'm clearly feeling like 400 episodes older. A lot has happened.
2:18
Speaker1: Okay. We talked in 2023 because you were one of the winners of the German Startup Awards.
2:25
Speaker1: And three years ago on this show, you said most performance management fails
2:30
Speaker1: because it is backward looking. three years and one platform rebuilt later?
2:36
Speaker1: Was Leapsome itself getting wrong in 2023?
2:40
Speaker0: Yeah, great question. I mean, I think we're always smart in hindsight, aren't we?
2:46
Speaker0: And I think at the time when we spoke, I would have to go back to when exactly
2:52
Speaker0: we spoke, what we might have not made or were just in,
2:56
Speaker0: the process of taking the decision to go from vertical, best-in-class talent,
3:02
Speaker0: software to becoming a broad platform.
3:06
Speaker0: So we took the bold strategic bets to consolidate the markets to some extent from the other side.
3:14
Speaker0: A lot of platform, core HR players have moved out into talent,
3:19
Speaker0: into recruiting, into adjacent categories.
3:23
Speaker0: But no one had yet successfully shown that you can go from the depths and talent
3:29
Speaker0: to actually becoming that strong, broad HR platform layer that all of HR ultimately runs on.
3:36
Speaker0: And yeah, that's the decision we took. And maybe we should have taken it even a year earlier.
3:44
Speaker0: But yeah, again, you're always smarter in hindsight. And I can share later why
3:48
Speaker0: exactly we decided to take that move.
3:51
Speaker1: Mm-hmm. I see.
3:56
Speaker1: You bootstrapped for five years, reportedly to around 10 million ARR,
4:04
Speaker1: while you as rivals raised hundreds of millions.
4:08
Speaker1: Discipline or necessity, and what did it force into the product that funded
4:15
Speaker1: competitors never built?
4:18
Speaker0: I mean, I think two things. I would say, why did we do it was focus.
4:23
Speaker0: We were never dogmatically bootstrapped but we rather always asked ourselves what ultimately,
4:31
Speaker0: is the core priority right now and for us the core priorities were understanding
4:35
Speaker0: customer needs and building product that solves those needs better than anybody
4:40
Speaker0: else has solved them so that was kind of the rationale for bootstrapping it didn't distract us,
4:45
Speaker0: um in in terms of having to to go fundraise.
4:49
Speaker0: We're radically focused on customers and product.
4:51
Speaker0: At the time when we were bootstrapped, we were growing as fast as VC-backed companies.
4:58
Speaker0: And so focus was the main driver of bootstrapping.
5:03
Speaker0: And why did we then take a round? We took a round because we wanted to expand
5:08
Speaker0: into the US and that requires some upfront investment. We already got some pull demand out of the US.
5:14
Speaker0: So to capture that more effectively, compete in the best software market in
5:18
Speaker0: the world, which makes you a better software player.
5:20
Speaker0: And at the time, we were actually planning to more upmarket,
5:24
Speaker0: but also means like longer sales cycles. You just need to fund more of your
5:28
Speaker0: go-to-market motion upfront.
5:30
Speaker0: And what has that, what DNA do we maintain because of our bootstrapping times?
5:37
Speaker0: And I would say it's that strong focus on the customer because we needed to
5:43
Speaker0: build things that actually generated revenue in the early days.
5:47
Speaker0: And that DNA of really closely listening to our customers, predicting,
5:54
Speaker0: foreseeing what are the next needs that are starting to emerge that no one has
5:58
Speaker0: actually solved really well.
6:00
Speaker0: Right now, for example, AI transformation, like the role in HR department,
6:04
Speaker0: people, leaders play in that transformation is something that isn't solved.
6:08
Speaker0: It actually plays to our strengths.
6:09
Speaker0: But that's really core focus on the customer needs and closing the gap for the
6:15
Speaker0: customer. That's kind of part of our bootstrapping DNA.
6:21
Speaker1: Being focused on customers is usually a pretty good thing.
6:24
Speaker1: But I was wondering, when you stay out of the funding market for half a decade,
6:30
Speaker1: what did that staying out of the market for so long cost you?
6:35
Speaker0: I mean, it's a good question.
6:40
Speaker0: So, I mean, I think the upside is focus.
6:44
Speaker0: The downside is some competitors have more resources to build positioning,
6:51
Speaker0: awareness, have GTM scale in the market.
6:56
Speaker0: At the same time, I think resources are one ingredient in winning the go-to-market game.
7:06
Speaker0: The other is exceptional product, clear differentiation, and to some extent,
7:11
Speaker0: finding the right levers and hacks to grow and start that flywheel of customer referrals.
7:22
Speaker0: And if you look at our category, HR software is also strongly bought by understanding
7:30
Speaker0: what peers use, what peers are starting to use.
7:33
Speaker0: So, that is something that is also important to us and that,
7:38
Speaker0: again, is also grounded in do you understand your customers' needs and solve
7:43
Speaker0: their problems better than others in the market.
7:46
Speaker1: Jenny, walk me through the two or three decisions that turned a feedback tool into a platform.
7:55
Speaker1: When exactly did you know the point solution was a dead end and what triggered it?
8:03
Speaker0: Great question. So the first phase of LeapSim, we were a talent software, right?
8:10
Speaker0: We brought together strengths and depths in performance management,
8:15
Speaker0: feedback, OKRs, meeting management, things that drive alignment,
8:19
Speaker0: as well as leveraging data to build better organizations, for example,
8:22
Speaker0: through engagement surveys.
8:24
Speaker0: So that's been the site where we started. Vertical, absolute best in class for that talent side.
8:30
Speaker0: And sort of into 2023, we saw more and more consolidation pressures in the market.
8:38
Speaker0: So a lot of pressure to reduce tooling, to consolidate tooling,
8:44
Speaker0: to some extent to have like one hub or central platform that a department truly
8:49
Speaker0: runs on. That was one side of the coin.
8:52
Speaker0: The other side, on the opportunity side, is seeing where AI and generative AI
8:58
Speaker0: in particular was heading.
9:01
Speaker0: We also were very clear that owning the data layer, not just in a particular
9:08
Speaker0: subpart of a function, but actually across a whole department,
9:12
Speaker0: is going to be something that's incredibly valuable.
9:15
Speaker0: Owning the workflows truly end-to-end in a particular department to then leverage.
9:22
Speaker0: AI, I mean, now increasingly genetic AI across the entire department grounded
9:27
Speaker0: in that data and that contextual understanding is something that's incredibly valuable.
9:32
Speaker0: So there was both an opportunity, a technological opportunity,
9:35
Speaker0: as well as a market involvement that we saw that ultimately decided that we.
9:43
Speaker0: How we decided that we actually want to become that platform player and evolve the strategy.
9:49
Speaker0: And at the same time, third reason, that no one actually has the strengths on the talent side.
9:56
Speaker0: And while there were all these consolidation trends in the market,
10:00
Speaker0: we had customers telling us, like, we need that strength on the talent side.
10:04
Speaker0: We cannot actually rely on the
10:07
Speaker0: much weaker, much more superficial talent capabilities of other vendors.
10:11
Speaker0: Like, why don't you actually expand into the broad platform space?
10:15
Speaker0: So that's the third one, like getting very direct customer feedback to some
10:18
Speaker0: extent, like really directly asking us to do that.
10:21
Speaker0: And I think that is even becoming more and more important at a time where being
10:26
Speaker0: a strategic partner to your organization, to your leadership team as an HR function
10:31
Speaker0: is paramount as we speak about the AI transition in the workplace.
10:36
Speaker1: So with this reason, you for the first time took in $60 million from Insight Partners in March 2022?
10:45
Speaker0: So at the time, the main driver behind the decision to fundraise was,
10:50
Speaker0: I mean, one was the milestone we had kind of surpassed the 10 million ARR.
10:55
Speaker0: And I think it's always good to ask yourself, what got you here?
10:58
Speaker0: Will that also necessarily get you there? So what's kind of what gets you from 10 to 100 million?
11:04
Speaker0: So that's kind of was one of the drivers where we decided to go fundraise.
11:08
Speaker0: The second one was we wanted to or were getting ready to open an office in New York.
11:14
Speaker0: And expand our footprint in the US markets.
11:19
Speaker0: That is an endeavor where all other founders tell you it takes longer,
11:22
Speaker0: it is harder, it's more cost intense than you can even foresee.
11:26
Speaker0: So funding the US market entry was the second driver for the funding decision.
11:34
Speaker1: I'm curious, just on a side quest, how big was the competition between the employees
11:43
Speaker1: who can open the New York office?
11:46
Speaker0: Yeah, good question. I mean, there was a lot of interest. We actually sent a
11:50
Speaker0: launch team of, I think, seven people at the time to open the U.S.
11:53
Speaker0: Office, like a mixture of sales, customer success, and had someone from our leadership team.
12:04
Speaker0: And ops and yeah, it was obviously a fun, sexy opportunity for a lot of people,
12:11
Speaker0: and we tried and have ever since had also a lot of travel back and forth between the offices.
12:17
Speaker0: Everybody loves to spend some time on the other side of the Atlantic.
12:23
Speaker1: In October 2024, he launched with, how do you pronounce it? H-R-I-S or HIRIS?
12:33
Speaker0: Yeah, H-R-I-S.
12:34
Speaker1: H-R-I-S, okay. Built from scratch into market, personally, Workday, and ADP already owned.
12:42
Speaker1: Building boring HR infrastructure in the middle of an AIM boom looks a kind of a bit contrarian.
12:49
Speaker1: Why was it the precondition for everything you now call agentic?
12:55
Speaker0: Yes, great question. So we took the decision to move from,
13:01
Speaker0: best-in-class tenant to becoming the single source of truth slayer that owns
13:07
Speaker0: all of the data and all of the processes in an HR function.
13:12
Speaker0: And there's two reasons why we believed and continue to believe in that strategy
13:19
Speaker0: one is you actually have the data across the entire department so you have a
13:24
Speaker0: lot of the contextual data,
13:27
Speaker0: in HR actually because of our depth and talent much beyond around the strategy
13:31
Speaker0: the alignment the manager effectiveness the performance of a team how an organization
13:36
Speaker0: works but we could connect this data and can connect this data with a lot of the compliance data.
13:43
Speaker0: The employee attributes, et cetera.
13:46
Speaker0: So that gives us a really rich contextual data pool, a talent graph we could build.
13:52
Speaker0: And we have the ownership of workflows across an entire department.
14:01
Speaker0: If something happens here that can trigger a whole lot of additional things
14:04
Speaker0: across the entire employee life cycle across the organization and having that
14:09
Speaker0: connection in the data and in the workflows and now,
14:13
Speaker0: in the genetic capabilities that enable that is something that generates real value to our customers.
14:18
Speaker0: So ultimately power is how an organization or HR organization and a company,
14:24
Speaker0: interacts with data and people processes.
14:27
Speaker1: Hmm? You rebuilt. I was wondering what broke during the rebuild and what would you not do again?
14:41
Speaker0: I mean, I think it's what's broke.
14:46
Speaker0: I mean, I think there's a lot of things you know. You do really diligent customer
14:51
Speaker0: research, user research, product research.
14:54
Speaker0: And then there is maybe the things you don't know yet.
15:02
Speaker0: And I think the part that,
15:08
Speaker0: I, in hindsight, probably would have invested into earlier is changing the perception
15:15
Speaker0: of the platform in the market.
15:17
Speaker0: If you enter the market with a new product, you have the benefit of the curiosity
15:21
Speaker0: of like, oh, who are they? What do they do?
15:24
Speaker0: And if you are an established and actually even like beloved brand,
15:29
Speaker0: the challenge is like, how do you get people to reevaluate and rethink what
15:33
Speaker0: you actually do and change the perception of who you are?
15:37
Speaker0: So that is actually something that is something that, again,
15:43
Speaker0: requires some time, resources to re-educate the market.
15:46
Speaker0: So I think that is something in hindsight I would have started potentially even
15:50
Speaker0: earlier alongside the build.
15:53
Speaker1: Yeah that's always a good idea because then uh people call you hey look i'm
15:58
Speaker1: interested in your product uh we're doing something
16:01
Speaker1: really different now oh i didn't know that's usually a sign your way past that
16:06
Speaker1: but i do believe a lot of companies out there start to communicate
16:12
Speaker1: something like a pivot like a change of strategy way too late to the outside
16:17
Speaker1: they're waiting for the big moment And then from one moment to another, you just
16:21
Speaker1: tap the switch and that's it.
16:23
Speaker1: But it really doesn't work like that, especially in the minds of the clients, right?
16:29
Speaker0: Especially if people have like, they already think they understand what you
16:33
Speaker0: do, right? So it's almost like re-education is harder than initial education.
16:39
Speaker1: Yeah, people don't like to unlearn, yeah.
16:45
Speaker1: In a minute, the public numbers say one HR person can support 299 employees.
16:51
Speaker1: And why Jenny's own research says most companies buying AI see nothing like that. Stay with us.
17:03
Speaker1: Jenny, since July 5, you hosted AI agents run on the platform.
17:09
Speaker1: Name one HR program an agent executes end-to-end today, like every step.
17:17
Speaker1: And tell me where the human signs off.
17:20
Speaker0: I mean, I think a favorite one of mine is we actually see when we speak to HR
17:28
Speaker0: operators that there's still so many manual processes running in an HR department.
17:33
Speaker0: Like people move things from A to B or they remind other humans to like complete
17:39
Speaker0: a step, send something, et cetera. So there's a lot of manual work.
17:44
Speaker0: And to some extent, a lot of that could have already been automated by workflows before.
17:51
Speaker0: However, setting up a workflow takes a bit of time. Maybe some people are also
17:57
Speaker0: like afraid, how do I actually do that?
17:59
Speaker0: What are the right triggers? What are the right events? Can I get it wrong setting
18:03
Speaker0: this up? And one of my favorite.
18:08
Speaker0: Agentic capabilities is actually Leapsome is building the workflows for you.
18:13
Speaker0: So if you have, I mean, either the platform can also help you to draft the process
18:19
Speaker0: you actually want to run, or maybe if you already have a process you want to
18:23
Speaker0: run, let's say an onboarding process.
18:25
Speaker0: There's a whole bunch of things that need to happen, and tool lock-ins that need to be created.
18:31
Speaker0: Equipment that needs to be set up, access right that needs to set up,
18:35
Speaker0: but also maybe the more cultural things like onboarding, maybe a buddy in the
18:39
Speaker0: team, a check-in conversation with a manager, meaningful peers.
18:44
Speaker0: If you have that onboarding process set up, you basically drag and drop it into
18:48
Speaker0: our agentic workflow builder, and the platform builds the process the way you
18:53
Speaker0: have envisioned it and from a technical perspective in the tool. So it's,
18:58
Speaker0: basically from your process, understand what needs to happen on which day,
19:02
Speaker0: who are the respective countable person, what are the rules behind that,
19:07
Speaker0: what are the steps that need to be set up,
19:10
Speaker0: what are some of the steps, can they happen in parallel, do they need to happen,
19:15
Speaker0: in a particular sequence, and all of that is set up immediately and.
19:21
Speaker0: With a much lower technical barrier because everyone can drag and drop a document, right?
19:27
Speaker0: I still believe there's value in the human kind of double-click,
19:31
Speaker0: double-checking if the way it's being set up actually is exactly the way they
19:36
Speaker0: wanted to set it up to quickly audit that,
19:39
Speaker0: but then they can activate this workflow and it's running on autopilot as well.
19:44
Speaker0: But I think this is a really nice example where natural language,
19:50
Speaker0: everybody can define a process in natural language is then being set up to run,
19:56
Speaker0: in an automatic way so much easier than we used to be able to do that in the
20:01
Speaker0: past. That's, I think, one example.
20:04
Speaker0: I could give you others, but I think that's one that's solving a real manual pain point in HR.
20:09
Speaker1: Actually, I also love the same thing. For example, a lot of people already noticed
20:15
Speaker1: that we have a semi-automated way with agents right now to post on social media for Startup Radio.
20:22
Speaker1: They notice and they like it, but you know what my first experience was with
20:27
Speaker1: an agent? I wanted them to post on social media.
20:30
Speaker1: It turns out no post arrived at the end because they just wrote it in a ledger.
20:39
Speaker1: So, when you have the bugs figured out and you have it made very simple,
20:46
Speaker1: you have a framework and understanding what it does, it makes it easier to have agents in the running.
20:53
Speaker1: But I do love the idea with other companies, with you, for example, also Moss,
20:59
Speaker1: is there no complete automation, but the human in the loop, because there's
21:03
Speaker1: some processes like your salary, where you don't want an AI agent to go crazy
21:09
Speaker1: or your performance review? Or
21:15
Speaker1: are there non-important processes in HR? I don't think so.
21:19
Speaker0: No, I think it's more about getting the aspects that the agent can do well and
21:27
Speaker0: the human kind of collaborating most effectively together.
21:31
Speaker0: So the agent can quickly set something up.
21:35
Speaker0: They can see maybe if we think about payroll preparation, maybe see an unexpected,
21:44
Speaker0: pattern they would expect differently. But then the human looks at the data,
21:49
Speaker0: is the workflow set up or looks at the interface of the workflow set up correctly?
21:53
Speaker0: Or if there is an anomaly in a payroll, maybe there's a reason behind it that the human knows, right?
22:01
Speaker0: So I can, it's a signal that helps me, like it's auditing, but then it's also
22:06
Speaker0: giving me the signal where as a human, I might want to duck deeper and build
22:12
Speaker0: additional context or start an additional sort of review process,
22:16
Speaker0: that I think is the best of combining the best of both worlds.
22:21
Speaker1: In preparation of this interview, I talked to a lot of lovely people on your
22:25
Speaker1: team and they insist Leapsom is a platform, not a tool.
22:30
Speaker1: Strip away the language. What is the operational difference for,
22:34
Speaker1: let's say, a nine-person team?
22:37
Speaker0: I mean, I think it's Leapsam as a platform is becoming the central layer that
22:43
Speaker0: enables a small HR team to run like a much more mature, much bigger,
22:51
Speaker0: people or HR team, right?
22:53
Speaker0: A lot of the things that don't have to be done manually can be delegated to the platform.
23:01
Speaker0: The platform also gives the team the insight and the data points where to double click on.
23:07
Speaker0: Or let's make that really specific. We have resilient signals in the platform.
23:12
Speaker0: They give you a signal, a data point where you maybe want to dive deeper to
23:18
Speaker0: have a retention risk there.
23:21
Speaker0: It's a conversation starter. it's orienting my attention as a human to double
23:25
Speaker0: click on something and figure out like what's the backstory to his do i need
23:28
Speaker0: to focus some of my attention there
23:30
Speaker0: also like it's solving real pain points right um i have a board meeting coming
23:34
Speaker0: up the next day all of a sudden i'm getting a data request that i would have
23:38
Speaker0: been scrambling to to respond to
23:41
Speaker0: in the past but now i can use
23:43
Speaker0: the functionality the genetic functionality in the platform and vibe code a
23:46
Speaker0: dashboard to actually pull together the right information in the moment to answer
23:50
Speaker0: that question even like a nice polished,
23:53
Speaker0: report and dashboard right that's solving a real pain and that enables the team to act.
23:59
Speaker0: Much more strategically have much more impact in an organization and save themselves
24:05
Speaker0: a lot more of the sort of like pain the scrambling the manual work and thereby operate on on this layer,
24:13
Speaker0: across all of the data, across all of the processes, the workflows in a really effective way.
24:18
Speaker1: You're actually, during your rebuild, you built an EU-hosted,
24:22
Speaker1: an EU AI Act compliant to a platform from the start.
24:29
Speaker1: A constraint that slowed you down or do you see it as an asset now selling in the US?
24:39
Speaker0: I'd actually say it's it's i mean we are working with highly sensitive data
24:47
Speaker0: compliance is essential as paramount it's the baseline we operate on and,
24:53
Speaker0: the regulation we have in europe i think is is guiding us to provide and build
25:00
Speaker0: at that level of security at that level of um,
25:05
Speaker0: of care and,
25:09
Speaker0: be intentional about also the agentic and AI products we build.
25:13
Speaker0: And I think this is something that we also bring into the US market.
25:18
Speaker0: And I think it also gives us a competitive edge, right? We can,
25:24
Speaker0: and it's actually part of also how we educate the market that with maybe traditional software.
25:33
Speaker0: Some of the judgments your vendors put into how they build authentic capabilities
25:39
Speaker0: are actually really, really important, but you don't even necessarily see them on first side, right?
25:44
Speaker0: So understanding kind of the product principles of your technology partners,
25:51
Speaker0: understanding how they build technology is really, really essential,
25:56
Speaker0: especially at this day and age.
25:59
Speaker0: So, being a vendor with a strong European and a strong U.S. presence,
26:05
Speaker0: I think, is something that can actually speak to the diligence and the care,
26:10
Speaker0: that we put into our capabilities across the board.
26:16
Speaker1: We're also talking about the report...
26:23
Speaker1: You made. Can you define talent density in economics, not cultural terms for
26:32
Speaker1: us, and what does it change on a company's P&L?
26:37
Speaker0: So what does talent density mean? Talent density means the share of your workforce
26:44
Speaker0: that is truly effective in the context of your organization, right?
26:50
Speaker0: Different organization, run slightly differently, have different preferences,
26:54
Speaker0: have different operating rhythms.
26:56
Speaker0: So talent density is about finding to some extent the recipe of who is successful
27:03
Speaker0: in your particular company.
27:04
Speaker0: What do you understand about them? How do you make sure you hire more people
27:09
Speaker0: that will be that level of effective in your organization? And how do you actually
27:13
Speaker0: make it more successful?
27:16
Speaker0: Tangible also to your existing workforce of what exactly your expectations are
27:22
Speaker0: for them to be effective.
27:24
Speaker0: So it's very much about managing or setting clear expectation,
27:29
Speaker0: but then also managing against those in hiring and with the existing team.
27:33
Speaker0: And why does it matter? Because in the end, that means you'll have a higher
27:39
Speaker0: performing, more effective team.
27:42
Speaker0: And in many organizations we work with talent and humans are still a lot of the deciding factor um.
27:50
Speaker0: As to how successful companies actually are and um if we want to kind of like
27:56
Speaker0: go into the reports um what we actually did is we analyzed um in our 2027,
28:02
Speaker0: workforce trends report that you also find on Leapsom's website,
28:07
Speaker0: we analyzed data from the 100 most successful, fastest growing companies across
28:16
Speaker0: a lot of publicly available data points.
28:20
Speaker0: We complemented them with some in-depth calls to kind of complement some of
28:26
Speaker0: the sort of broader data points we analyzed.
28:29
Speaker0: So it was a larger data analysis. Historically, we've actually often done these survey-based.
28:36
Speaker0: This one we've done data-based, about 5,000 data points we analyzed,
28:41
Speaker0: partially with AI with a lot of manual human facts validation.
28:47
Speaker0: And I think what's interesting, if we want to go into those directly,
28:52
Speaker0: and basically the question we ask ourselves is, how does HR in these best companies actually work?
28:59
Speaker0: And I think the three things that really stood out was the HR teams weren't
29:04
Speaker0: actually necessarily getting smaller, which is interesting because we often
29:09
Speaker0: ask ourselves the question, does AI actually kill jobs?
29:14
Speaker0: So in these companies, we actually, with two outliers, haven't really seen that.
29:19
Speaker0: The HR teams were roughly the same size as they are in kind of the industry average.
29:24
Speaker0: However, what work is done in these organizations has shifted and roles have shifted.
29:31
Speaker0: And they shifted away from the more sort of like mundane administrative tasks
29:37
Speaker0: to the more talent-oriented strategic tasks of building the right organizations.
29:43
Speaker0: Retaining, well, like hiring the right talent, retaining the right talent,
29:48
Speaker0: making managers effective, building the right org structures.
29:51
Speaker0: So it's the automation of the core layer has enabled the more strategic work
29:57
Speaker0: that ultimately, again, creates to some extent that flywheel of more talent, more success.
30:04
Speaker0: And what's also been interesting is that in these organizations,
30:10
Speaker0: HR tends to be – or the HR leadership tends to be set higher up in the organizations.
30:18
Speaker0: Which is also interesting.
30:19
Speaker0: So HR actually has a more strategic seat at the table.
30:22
Speaker0: And those were the most fundamental changes that we've seen in that data set.
30:27
Speaker1: I've also looked into the report. By the way, I do believe we will link the
30:32
Speaker1: website you'll share with me after the interview. And we'll link it where everybody
30:36
Speaker1: can find it, at least on our blog.
30:39
Speaker1: And go straight from there to you and download it there.
30:43
Speaker1: Sorry for the detour. Your report puts the median...
30:47
Speaker1: At 53 employees per HR professional and 11 labs at 299 with people team of just four.
31:00
Speaker1: Can a company outside of an AI lab reproduce this one to 299 for a normal company,
31:09
Speaker1: or is it just an AI lab anomaly that just happened?
31:16
Speaker0: I mean, it's an interesting question. The two outliers were,
31:18
Speaker0: as you mentioned, 11 labs and the other one was OpenAI.
31:22
Speaker0: That was where like the massive outliers was like 250 to 200 or 300 employees per HR headcount.
31:32
Speaker0: I mean, I think at this stage, I would say these are rather outliers and the
31:38
Speaker0: 98 other companies in the data set, again, behaved a lot more like the standard.
31:47
Speaker0: Company in terms of numbers, not in terms of what people do.
31:50
Speaker0: Again, like very different, important differentiation.
31:55
Speaker0: I mean, I think neither of us has a crystal ball, right?
32:00
Speaker0: But I think what we're actually rather seeing from the conversations we're having
32:04
Speaker0: with people leaders in our community, in the markets, in our ecosystem is.
32:13
Speaker0: HR's role has actually become more important in becoming a strategic partner
32:18
Speaker0: to the C-level, the CEO in navigating companies through the AI transformation.
32:23
Speaker0: It's partially a technological transformation, but it's also very much a people transformation.
32:28
Speaker0: So what we're seeing is HR is not just tasked to rebuild their own department
32:35
Speaker0: and make HR truly like agentic and set up for the,
32:42
Speaker0: AI era in the narrow sense, like that's the own HR department strategy, right?
32:46
Speaker0: That's what we empower with SleepSense agentic HR platform.
32:49
Speaker0: But beyond that, it's also a key partner to the rest of the business to navigate
32:53
Speaker0: the rest of the organizations through the change, right?
32:55
Speaker0: We've seen massive change in engineering organizations. Like one of the fastest
33:00
Speaker0: evolutions of a job that we've seen in this decade, like massive change of how people work.
33:08
Speaker0: And HR is still in many organizations the most experienced change management
33:15
Speaker0: professional that can be.
33:18
Speaker0: The partner to a CTO to manage their organization through that change.
33:21
Speaker0: How do we rethink hiring? How do we rethink management? How do we rethink compensation,
33:28
Speaker0: at the time where the job of an engineer has changed so much and maybe converges
33:33
Speaker0: with product roles and design roles, right?
33:35
Speaker0: So I think that's kind of the whole change management track record or change
33:39
Speaker0: management work stream.
33:40
Speaker0: Another one we've seen is like, how do we need to rethink org designs?
33:44
Speaker0: Are teams becoming flatter?
33:47
Speaker0: Are we increasing the span of control for managers, right? This is again one
33:51
Speaker0: where the strategic partner to the leadership team is often the HR persona.
33:57
Speaker0: They're the ones that also then implement these changes.
34:00
Speaker0: So all of this really essential strategic talent work and org work,
34:05
Speaker0: And it's also something that has become even more valuable, even more urgent,
34:11
Speaker0: even more in demand, and again, also needs to have the resources and also the
34:15
Speaker0: tooling to get this job done right.
34:18
Speaker0: So this is why I would see these, I would probably right now predict that the
34:26
Speaker0: 11 labs and open AI ratios are more of the outliers.
34:29
Speaker0: And because all of these challenges are real and a lot of work,
34:34
Speaker0: I would expect the HR team sizes to be staying relatively similar,
34:42
Speaker0: but with a radical change in where the priorities lie and what expertise is actually in demand.
34:48
Speaker1: Jenny, your team was careful to tell us that 40% admin time and 20% performance
34:54
Speaker1: numbers are customer outcomes since adopting Leapsome,
35:00
Speaker0: Not an AI before or after.
35:01
Speaker1: Honestly, respect for that. So what is the AI before or after inside Leapsome?
35:10
Speaker0: Inside Leapsome as an organization, right? Not for our customers,
35:13
Speaker0: but for us as a company ourselves. I think it's...
35:18
Speaker0: We actually hired a new VP, People and Strategy, earlier this year.
35:26
Speaker0: I think she started in March, who is also operating from a very AI-first mindset.
35:34
Speaker0: And it's really exceptional. She's actually going to run a masterclass,
35:39
Speaker0: for a cohort of customers in Q1. So I think kind of the six months to Gentic,
35:46
Speaker0: HR. So, we have a people leader in place who's really exceptional on rethinking HR.
35:54
Speaker0: And we have been very, I mean, to some extent, we're always our customer number
36:02
Speaker0: one on everything we build anyways.
36:04
Speaker0: So, what has changed is we have introduced a whole lot of.
36:13
Speaker0: Or we've reinvented a whole lot of processes that used to be slow,
36:19
Speaker0: manual, painful. And I think we haven't done a great job to sum up of what all of these compound to.
36:28
Speaker0: But I think if I give you specific examples, for example, our calibration process
36:32
Speaker0: used to be something that was probably a net one week lift for the people team.
36:39
Speaker0: That is probably now in the late last form with AI-generated pre-reads probably
36:47
Speaker0: condensed down to less than a day of work.
36:50
Speaker0: So cutting that from one week to one day is like a very.
36:58
Speaker0: Very significant lift in how that process is actually now much easier.
37:05
Speaker0: And I think calibration is one of the really, really valuable parts,
37:10
Speaker0: of a good performance process and not every organization is able to do it and
37:15
Speaker0: run it because it used to be quite resource intensive, right?
37:19
Speaker0: And I think this is to the point I made earlier, a way how a small people team
37:26
Speaker0: can now actually act like only the best, most mature people team could in the past, right?
37:33
Speaker0: Because I now have technology as partner that can help me get these lifts done well.
37:38
Speaker0: And I think this is something we're feeling internally.
37:43
Speaker0: It leads them across calibration, across,
37:48
Speaker0: coaching managers to make merit and performance cases, to prepare them right,
37:53
Speaker0: to challenge them right, where they aligned with the principle of what grants
37:58
Speaker0: a promotion, what grants...
38:02
Speaker0: A merit. It's making sure we really,
38:07
Speaker0: automate as many of the internal processes, make it even easier to be self-service,
38:15
Speaker0: as the user, as the manager, to have a sparing partner to prepare me for a hard
38:20
Speaker0: conversation or to make a conversation more effective.
38:26
Speaker0: I think for our managers, preparing reviews,
38:31
Speaker0: has probably like I always used to take me about 45 minutes to prepare a really
38:36
Speaker0: strong feedback conversation it's probably now down to like 15 minutes,
38:41
Speaker0: using sort of leveraging pulling context through AI more effectively challenging
38:47
Speaker0: myself on where I have provided enough example I haven't so again I think that's a 3x lift,
38:54
Speaker0: on my time so there's all these very specific ways.
38:58
Speaker0: It hasn't just made the lives of our HR team simpler, but actually also significantly
39:05
Speaker0: for our managers while increasing the transparency and the fairness for employees.
39:13
Speaker0: And also the ability to own your own development to some extent.
39:16
Speaker1: I also found that 81% of companies with available data, put AI into people functions.
39:27
Speaker1: Yet your research says most fail to get the expected productivity gains.
39:35
Speaker1: Where does the gain actually die? Is it in hiring, in management spans, skills, or trust?
39:43
Speaker0: I would say it's because we use AI to compartmentalize, like in many organizations,
39:54
Speaker0: it's still like a chat here, a quick,
40:00
Speaker0: LLM conversation there. But I think the biggest lift comes from making sure
40:07
Speaker0: you really have a strong connection of all of the relevant contextual layer.
40:13
Speaker0: So, for example, all the context in Leapsam, Leapsam integrates through an MCP
40:17
Speaker0: with Slack, with Notion, right?
40:19
Speaker0: Like connecting all of the data that is needed to give a most relevant,
40:28
Speaker0: or use AI in the most relevant ways. And then really rethinking,
40:35
Speaker0: processes and being more ambitious
40:37
Speaker0: around what the AI or the identity capabilities should be doing for you.
40:42
Speaker0: So moving maybe more away from like piecemeal to having a strong technological setup with,
40:50
Speaker0: access to a broad range of processes and deep contextual data that is then the unlock for higher,
41:01
Speaker0: productivity, but actually also like ultimate impact.
41:12
Speaker1: Only 30% of AI Labs' own people team job postings mention AI skills.
41:20
Speaker1: I found this pretty funny.
41:21
Speaker1: The loss of the 11 segments you measured.
41:25
Speaker1: The companies building AI do not hire for it in HR.
41:29
Speaker1: What does that tell you?
41:32
Speaker0: I think that was an interesting one. And we deep dived into it.
41:36
Speaker0: And allow me a quick diversion.
41:41
Speaker0: Do you remember the ads I don't know what was that in the 90s of,
41:47
Speaker0: am I in bin ich drin for internet,
41:51
Speaker0: providers and I think at the time like it was such a like novel thing to like
41:57
Speaker0: go onto the internet so we celebrated
41:59
Speaker0: right and we would say I go onto the internet now I shop on Amazon
42:03
Speaker0: I listen to music on Spotify I watch I don't know shows on Netflix right So
42:08
Speaker0: we don't actually talk about the internet anymore.
42:11
Speaker0: It's such a like table stakes.
42:13
Speaker0: And to some extent, what we're seeing there is the most AI native and AI forward
42:19
Speaker0: companies don't even necessarily speak about AI in the broad terms anymore because
42:24
Speaker0: it's so table stakes already.
42:26
Speaker0: What they specifically mention and look for is the right kind of judgment.
42:31
Speaker0: It's asking good questions. It's almost like the next layer of showing the clear,
42:38
Speaker0: capabilities and skills that make working with AI most effective.
42:44
Speaker1: Let's teach you a little bit about the future of work.
42:48
Speaker1: What has the AI transition changed about how you personally lead?
42:54
Speaker1: We already talked about the feedback conversations, where your time goes.
42:59
Speaker1: Which decisions do you no longer make?
43:05
Speaker0: I mean, especially with...
43:08
Speaker0: Capabilities like co-work. There are some tasks where I previously asked somebody
43:15
Speaker0: else to do them for me that I can actually much easier set up for myself to run quickly.
43:23
Speaker0: And that almost means like no passing, unnecessarily having to pass on context.
43:31
Speaker0: Or waiting for someone else finishing that. So I think it allows me to actually,
43:35
Speaker0: get certain things done more simultaneously and faster.
43:41
Speaker0: I think that's maybe one very specific change.
43:47
Speaker0: I think it is challenging me to find time to experiment, explore,
43:54
Speaker0: learn, just because the space is actually moving so quickly.
43:57
Speaker0: And I think that's the challenge for any leader. There's not a lot of spare
44:01
Speaker0: time on the calendar and at the same time, it is important to really stay on
44:05
Speaker0: the forefront of what is happening.
44:08
Speaker0: And in terms of.
44:12
Speaker0: How it changes how I lead or how we function on the leadership team.
44:18
Speaker0: I think to some extent it has meant massive acceleration of velocity across the org.
44:26
Speaker0: We ship product faster.
44:28
Speaker0: We can run more experiments in parallel.
44:32
Speaker0: And to some extent we also compete with other players that are similarly moving faster.
44:39
Speaker0: So I think it also generates a new operating speed.
44:43
Speaker0: We always moved fast, but I think right now it's getting even faster,
44:46
Speaker0: which means you need to, as a company, need to,
44:50
Speaker0: continue to evolve your own operating rhythm to make sure like you both move
44:57
Speaker0: really fast, but we never want to compromise in quality either.
45:00
Speaker0: So finding that sweet spot of the right operating rhythm for speed with quality
45:08
Speaker0: and also probably being honest that that level of speed might not actually resonate
45:13
Speaker0: with everyone. And that's maybe also OK.
45:17
Speaker0: But I think it is also part of the new reality of running a technology company in this day and age.
45:25
Speaker1: This day and age. Curious. What happens to junior roles? Your own job postings
45:30
Speaker1: now require automation fluency.
45:34
Speaker1: Where does the first run of the career letter go?
45:38
Speaker0: I mean, I think folks who are entering the job market now have a disadvantage and an advantage.
45:47
Speaker0: I think the advantage is there's no preconceived learned ways of the old working.
45:57
Speaker0: Which to some extent can be an advantage, right? I try new things,
46:00
Speaker0: I start directly with grounded in the technology that's available at my fingertips right now.
46:07
Speaker0: And at the same time, the disadvantage is,
46:12
Speaker0: that I think sometimes you have to do things a certain, a few times yourself,
46:18
Speaker0: to get the pattern recognition, the muscle memory that enables you to be most successful.
46:24
Speaker0: For example, junior engineers, like it's harder to probably review codes when
46:28
Speaker0: you don't have the same kind of years of experience writing code yourself.
46:33
Speaker0: So how do you, how do you QA some white coded codes?
46:37
Speaker0: And I think in the end, it's a matter of pairing these two strengths and teams,
46:42
Speaker0: the folks that come with a wide blank canvas.
46:46
Speaker0: And the folks who have the mentorship, the coaching abilities to still continue
46:52
Speaker0: to actually advance folks in their career.
46:56
Speaker0: So I think it's not a nice role. It's like really combining the strengths of both.
47:02
Speaker0: But if I look at us internally at Leapsome, for example, we used to have support
47:09
Speaker0: roles, right? Now, a lot of the
47:11
Speaker0: sort of first level support tickets can actually be responded to by AI.
47:17
Speaker0: So this means the people on the support team actually get to work on maybe the
47:23
Speaker0: more interesting, but also more hairy tickets earlier, which requires there's
47:28
Speaker0: maybe slightly longer onboarding.
47:30
Speaker0: However, there's also technologies to support them, but it also actually makes
47:33
Speaker0: the job more interesting, a higher learning curve.
47:38
Speaker0: It's almost becoming a product specialist than a support person.
47:41
Speaker0: So I think there's, again, there's new challenges, but also quite a lot of new opportunity.
47:46
Speaker1: Your report says AI made the old split, HR managing people, while real decision
47:52
Speaker1: happened elsewhere, impossible to sustain.
47:55
Speaker1: Which companies will the industry study five years from now and what did they do differently?
48:04
Speaker0: I mean, I think what we're seeing in some of the best leading-edge companies is that,
48:13
Speaker0: next to their kind of traditional HR.
48:21
Speaker0: To-dos and roles and responsibilities, there's a lot more
48:27
Speaker0: focus on the strategic side of talent work the right org design the right manager
48:31
Speaker0: effectiveness the right kind
48:33
Speaker0: of understanding of who succeeds should be retained should be hired and,
48:38
Speaker0: then we've actually seen a,
48:41
Speaker0: whole stack of senior HR leaders now becoming also the AI transformation leaders
48:48
Speaker0: again like we're in a massive transformation and shift in work and how we work.
48:55
Speaker0: And part of that is underlying technology, but part of that is a massive people innovation effort.
49:05
Speaker0: Like, do we understand, like, how do we manage the change? How do we build new capabilities?
49:12
Speaker0: How do we redesign design org models, leadership models, compensation models.
49:18
Speaker0: So we see the best companies being in the forefront of that.
49:22
Speaker0: And I think that is the direction and the blueprint that other companies will
49:27
Speaker0: follow, where again, some of the core and the mundane is going to be increasingly
49:31
Speaker0: automated with judgment through humans at the right points.
49:37
Speaker0: But there's going to be more focus and roles on the more strategic.
49:42
Speaker0: Work and really supporting organizations finding the right operating rhythm,
49:49
Speaker0: in the era of AI or agents and humans working together.
49:54
Speaker1: So, I would take away from that statement that
50:00
Speaker1: the demand for people who also read and correct the footnotes is getting lower
50:05
Speaker1: and the demand for people who could think ahead, even in terms of what other
50:10
Speaker1: people do in terms of the strategy, the demand will increase there.
50:15
Speaker0: I mean, I think it's about asking the right questions, like interpreting signals,
50:21
Speaker0: where to go deeper, where to build better human contacts, taking,
50:24
Speaker0: making those decisions and,
50:28
Speaker0: making sure or making a call in a decision, not just what is good everywhere,
50:33
Speaker0: but what is the right and best for your organization, for your people.
50:37
Speaker0: So it's these human traits, the curiosity,
50:46
Speaker0: the human relationships, the judgments that are really, really paramount,
50:51
Speaker0: and those won't be replaced.
50:56
Speaker1: Jenny, you're already holding up very well here for 55 minutes of recording
51:02
Speaker1: time. Awesome. Thank you very much.
51:04
Speaker1: I have three more questions.
51:07
Speaker1: One on the capital and team, you raised once in 10 years.
51:12
Speaker1: Are you open to talk to new investors or was Series A the last round you will ever need?
51:18
Speaker0: I mean, we're not in an active funding process.
51:23
Speaker0: We are cash efficient. We have a lot of capital in the bank.
51:28
Speaker0: And at the same time, I think it's always good to speak to the right people, get to know one another.
51:34
Speaker0: And at the same time, we're also someone who's very laser focused.
51:37
Speaker0: So I think that's a good German Jain.
51:40
Speaker1: Jain, exactly. By the way, every investor interested go on our blog there is
51:46
Speaker1: your linkedin profile linked and
51:49
Speaker1: who are you hiring right now and what does ai fluency required actually screen for
51:56
Speaker0: Yeah great question um so we're actually hiring across um i think almost all
52:01
Speaker0: teams um so engineering products um marketing sales cs people I feel like I
52:09
Speaker0: forgot something, BizOps.
52:11
Speaker0: So I think there are some open roles across the board. And we are looking for,
52:18
Speaker0: I mean, we're first of all also looking for people who are aligned with our
52:22
Speaker0: operating systems, our culture.
52:24
Speaker0: We care about impact. We care about speed. We care about ownership without ego,
52:28
Speaker0: which is important to learning.
52:31
Speaker0: And we look for folks who are curious and capable in interacting with AI.
52:41
Speaker0: So being curious, asking the right questions, also being,
52:48
Speaker0: careful to evaluate where's the output right, where is it wrong.
52:55
Speaker0: So like that level of caution is also very important. And we actually tested
53:01
Speaker0: through case studies that are now AI enabled.
53:05
Speaker0: So finding mistakes is also important. And then it's important for us that as
53:10
Speaker0: people who are adaptable, who are willing to learn because the technology is evolving so quickly.
53:16
Speaker0: And that level of curiosity, adaptability and a learning mindset is really,
53:21
Speaker0: really quite essential for us.
53:24
Speaker1: Hmm, that sounds pretty interesting.
53:28
Speaker1: Let's close with your forecast. One falsifiable prediction on the record.
53:35
Speaker1: Where is the medium HR professional to employee ratio in your 2028 report?
53:43
Speaker1: And what happens to HR software spending per employee by then?
53:53
Speaker1: Predictions are always difficult, especially concerning the future. I know, I know.
53:58
Speaker0: So, where's the ratio? I actually think in next year's report,
54:04
Speaker0: the employee to HR persona ratio is going to be stable.
54:12
Speaker0: Because on the one hand manual work is going to be reduced and at the same time the importance of,
54:21
Speaker0: talent work and ai transformation work is essential right now for the companies
54:27
Speaker0: for companies to succeed so that's where the the priorities are shifting to so um,
54:35
Speaker0: same size of team, very different priorities and focus and impact,
54:41
Speaker0: and spend,
54:44
Speaker0: on HR tooling.
54:46
Speaker0: I think that's an interesting one because on the one hand,
54:52
Speaker0: I think that I would expect tooling.
54:59
Speaker0: Investments into tooling to increase.
55:03
Speaker0: And at the same time, through especially agentic capabilities,
55:08
Speaker0: that might also at times mean there's less need for external,
55:16
Speaker0: resources, certain consultants, certain recruiters, because more of the work
55:21
Speaker0: can also be done by the internal team with the help of technology,
55:26
Speaker0: which again means there's less loss of context.
55:30
Speaker0: It's faster it's a bit related to the example i also use i can now get more
55:34
Speaker0: done myself with technology so those would probably be my two predictions.
55:40
Speaker1: So that should scare a lot of free hr professionals out there working as recruiters
55:46
Speaker0: Headhunters and so on i mean i think at the same time there is also a lot of,
55:52
Speaker0: work to be done to help organizations make that transition so i think again,
55:57
Speaker0: the expertise isn't lost.
56:00
Speaker0: I think, again, it's going to shift, I think, where it's being deployed.
56:06
Speaker1: Jenny, thank you. The one insight I want everybody to keep, the agents are downstream.
56:12
Speaker1: LeapSum spent eight years building the data foundation before the first agent shipped.
56:18
Speaker1: And the density gap 53 versus 299 employees per HR professional is the number
56:25
Speaker1: that tells you whether AI is actually changing how a company runs.
56:30
Speaker1: You'll find LeapSum's 2027 Workforce Trends report and every source in the blog post.
56:36
Speaker1: Link here down in the show notes. If this episode was useful,
56:39
Speaker1: send it to one founder who still thinks AI transformation starts with buying AI tools.
56:45
Speaker1: Rate us, reviewers, share us. It's genuinely helpful.
56:49
Speaker1: This interview is in partnership with Leapsome.
56:52
Speaker1: This podcast goes live on Tuesday, September 29, 2026, the audio podcast earlier that same day.
56:59
Speaker1: I'm Joe Manager. This was . Thank you for listening.
57:04
Speaker0: Thank you so much.