A few months ago, an Argentine-founded HR company called Humand raised $66 million in a Series A.
That number caught my attention.
Not $66 million after ten rounds of funding.
A $66 million Series A.
The round was co-led by Kaszek and Goodwater Capital, with participation from Y Combinator and several well-known technology founders. Humand itself called it the largest HR-tech Series A in Latin America.
As an Argentine founder who works in hiring, I obviously wanted to understand what they had built.
And I came away with two thoughts.
First:
Humand has identified a very real problem that most HR software ignored for years.
Second:
I’m not yet convinced that turning the solution into another giant “AI operating system” automatically makes that solution better.
Those two things can be true at the same time.
The original idea is actually very good
Humand was founded in 2020 by Nicolás Benenzon and Gerónimo Maspero and went through Y Combinator’s Winter 2022 batch. YC currently lists the company as active with around 270 employees.
What interested me most wasn't the funding.
It was who Humand originally built for.
Most workplace software has historically been designed by people sitting behind computers for other people sitting behind computers.
Slack.
Notion.
Teams.
HR portals.
Employee intranets.
Performance systems.
Learning platforms.
They generally assume something very basic:
The employee has a laptop and a corporate email account.
But millions of workers don't.
Think about someone working in a warehouse.
A supermarket.
A manufacturing plant.
A hotel.
A restaurant.
A construction site.
A hospital.
They're employees too, but a lot of the infrastructure we associate with “the modern workplace” was never really designed around them.
Humand's pitch was essentially:
Put all of this on the employee's phone.
No corporate email required.
Company announcements, HR requests, documents, onboarding, training, recognition and other workplace functions in one place.
That's not a stupid idea.
It's actually one of the clearer HR-tech problems I've seen.
The company says it now serves more than 2 million users across 2,000+ organizations in 51 countries, including large companies such as OXXO, John Deere and Home Depot. Those are company-reported figures, but even with that qualification, this clearly isn't just a pitch deck looking for product-market fit.
There is a real business here.
Which is exactly why the next part interests me.
But now the product is becoming much bigger
Go through the product today and you find:
internal communications,
documents,
onboarding,
time off,
employee files,
company policies,
performance reviews,
goals and OKRs,
learning,
events,
recognition,
workflows,
roles and permissions,
integrations,
and more.
Then, naturally, AI.
Humand now describes itself as an “AI operating system” for the deskless workforce and says part of the $66 million will be used to embed AI deeper into workplace workflows.
This is where I start asking questions.
Not because I think the product is bad.
Because I've seen this movie before.
A startup identifies one genuinely underserved problem.
It solves it well.
Customers arrive.
Investors arrive.
Then the product expands.
One feature becomes five.
Five become fifteen.
Eventually the website tells you that the product manages your communication, culture, HR, talent, operations and AI.
And suddenly the very thing that made the company interesting becomes harder to see.
The best thing about Humand isn't AI
If I had to explain the company to someone in one sentence, I wouldn't say:
“It's an AI-powered workforce operating system.”
I'd say:
It gives workers who don't sit behind computers one simple place to interact with their employer from their phone.
That's much stronger.
It solves a problem I immediately understand.
A factory worker shouldn't need to chase a manager to find a company policy.
A retail employee shouldn't need a corporate laptop to request leave.
A warehouse employee shouldn't be disconnected from announcements because they don't have an office email account.
Those are problems.
And if Humand solves them well, there is enormous value there.
Whether an AI assistant sits on top of that is secondary.
Yet we're living in a market where apparently nothing can simply be useful anymore.
It has to be:
AI-native.
AI-first.
Agentic.
An operating system.
A copilot.
A platform.
Preferably all of them at once.
I think we're asking the wrong AI question
The question usually becomes:
“How can we add AI to HR?”
I think that's backwards.
Ask the employee.
What does the person working an eight-hour shift actually need?
Maybe they need to know:
When am I working tomorrow?
How many vacation days do I have?
Where is my payslip?
Did my leave request get approved?
What is the policy for this?
What training do I need to complete?
Who do I speak to about this problem?
If AI gets them the correct answer in ten seconds instead of making them search through five menus, excellent.
That's useful AI.
If it automates repetitive HR administration and saves an HR team hundreds of hours, excellent.
Use it.
But if we're adding AI because software investors currently expect an AI story, that's different.
The employee doesn't care whether the answer was delivered by an LLM.
They care whether the answer is correct.
That distinction sounds obvious.
The technology industry regularly forgets it.
And $66 million doesn't answer that question
Huge funding rounds create a strange effect.
People start treating investment as proof.
It isn't.
A $66 million Series A means sophisticated investors believe the opportunity could become very large.
That is significant.
It does not mean every product decision is correct.
It doesn't tell me:
How many of those 2 million registered users use Humand every week?
How deeply customers use the different modules.
How many companies replace existing HR systems with Humand versus adding it on top.
What retention looks like after several years.
Which modules employees actually use.
Whether adding more features improves adoption or makes the product more complicated.
How much of the AI functionality creates measurable improvements.
Those are much less exciting numbers than “$66M Series A.”
They're also the numbers I'd eventually want to know.
This isn't specific to Humand.
We have become strangely obsessed with funding announcements in technology.
Company raises $20 million.
Company raises $50 million.
Company valued at $2 billion.
Everyone congratulates them.
Six months later we're talking about the next company.
Capital raised is an input.
It isn't the result.
There is another risk with the “all-in-one” strategy
Every founder understands the temptation.
A customer says:
“Can you also do X?”
Another asks for Y.
Enterprise customer wants Z.
You build them.
Now your communication product has HR features.
Your HR product has learning.
Your learning product has performance management.
Your performance platform has workflows.
Your workflow platform has AI.
Eventually you've built twenty products inside one login.
Sometimes that creates an incredible platform.
Sometimes it creates twenty average products.
That's the difficult part.
Humand's original positioning around deskless workers is specific.
“Everything every employee might ever need from HR” is much less specific.
And HR isn't exactly a market lacking all-in-one software already.
There are enormous, deeply established companies selling HR systems.
The battle gets much harder once your proposition becomes:
We do everything too.
Humand's strongest moat may therefore not be how many HR modules it can add.
It may be something much simpler:
It understands a type of worker most HR software originally ignored.
I would protect that obsessively.
And I would probably stop calling everything an operating system
This is becoming one of my minor technology annoyances.
Apparently every sufficiently ambitious SaaS product eventually becomes an “operating system.”
Sales operating system.
Recruiting operating system.
Marketing operating system.
People operating system.
AI operating system.
Maybe I'm old-fashioned, but sometimes software can just be very good software.
Calling Humand an operating system doesn't make it more valuable to someone working in a distribution center.
Making that person's work life simpler does.
That's the metric.
So is Humand overhyped?
Too early to say.
And I don't think dismissing it would be fair.
They have a genuinely interesting wedge.
They have significant reported adoption.
They've attracted serious customers.
They've raised an extraordinary amount of capital.
And unlike many AI startups I look at, I can explain the original problem without mentioning AI once.
That's a compliment.
What I'm skeptical about is what happens next.
Can Humand maintain that clarity while becoming a much broader HR platform?
Can it demonstrate that its AI features improve outcomes rather than improve the pitch?
Can it turn millions of users into deeply engaged users?
Can it expand aggressively without becoming another HR suite containing twenty modules that companies only half use?
That's what I'd watch.
Not the funding announcement.
There's a bigger lesson here for founders
I actually think Humand represents both sides of what is happening in software right now.
The good side:
Someone noticed a group of people being poorly served by existing technology and built specifically for them.
That's exactly how great companies often begin.
The dangerous side:
Once something starts working, there is enormous pressure to make the story bigger.
More products.
More markets.
More AI.
Bigger terminology.
Bigger fundraising rounds.
Sometimes that ambition builds the next enormous technology company.
Sometimes it takes a very clear product and slowly makes it look like everyone else.
If I were building Humand, the question I would keep asking isn't:
“How much more can we add?”
It would be:
“Why did these companies choose us in the first place?”
Because there is a very good answer.
They built workplace software for the employee everyone else forgot.
I'd be careful not to lose that underneath the AI.