There's a case that gets cited a lot in conversations about artificial intelligence and employment, but it's usually only told halfway.
Klarna, the Swedish fintech payments company, launched an AI-powered virtual assistant that, in its first month, handled 2.3 million conversations — doing the equivalent work of 700 full-time agents. Average handling time dropped from 11 minutes to under 2. The estimated financial impact was a $40 million improvement in profit that year.
It's an extraordinary result. And it's the part everyone tells.
What gets told less is that, in the process, Klarna cut staff and later had to start hiring again. Why? Because it discovered that for certain types of interactions — the ones that require connecting with a customer's complex emotions, navigating delicate situations, and sustaining a real human bond — the technology wasn't solving anything.
That chapter of the Klarna case is the most important one. And it's the one more organizations are ignoring.
Not everything in an organization is automatable
There's a concept we use at Olivia to describe this reality: not everything in an organization is "AI-able." There are tasks, decisions, and interactions AI can do better, faster, and cheaper than a person. And there are others where human presence, with all its complexity and capacity for empathy, is irreplaceable.
The problem is that most organizations aren't making that distinction. The linear logic of "if AI saves hours, I cut headcount" isn't just ethically questionable — it's strategically wrong. It ignores the reskilling and upskilling of the people who remain in the organization, sabotaging its ability to innovate right when it needs it most.
According to the Microsoft 2026 Work Trend Index, the most important shift AI brings isn't the adoption of new tools — it's rebuilding the entire operating model. That's not a technology project. It's a transformation in how an organization understands work, distributes responsibilities, and develops its people.
The real challenge is human
Artificial intelligence is, in effect, the new electricity. But just as in the 19th century it wasn't enough to have access to electric current to transform a factory, today it isn't enough to have access to an AI tool to transform an organization. Processes had to be redesigned, workers retrained, production lines reorganized. Today the logic is exactly the same.
The real challenge isn't technological. It's human. It's organizational. It's cultural.
The organizations that will capture AI's real value are the ones that understand this isn't an IT project. They're the ones that start with mindset before the tool. The ones that design use cases connected to the real business, not to the trend of the moment. The ones that measure before they scale. The ones that invest in developing their people, not just in software licenses.
And above all: the ones that don't automate what's broken.
By Ricardo Niveyro, Director at Olivia.