Few people have done more to automate the office than Daniel Dines, the software engineer turned billionaire who built UiPath into one of Europe’s biggest technology success stories. His company pioneered robotic process automation (RPA), selling software robots that handle the tedious, repetitive tasks of white-collar work—data entry, invoice processing, contract review. In recent years, UiPath has aggressively pushed into AI agents, most notably by acquiring the compliance-automation firm WorkFusion. Yet on his own company’s podcast, The Path Forward, Dines spent much of his time delivering a message that sounds almost counterintuitive coming from a man who profits from automation: cut jobs too fast, and you destroy the very value you are trying to preserve.
“Everybody feels some sort of anxiety, me included,” Dines said in conversation with UiPath colleague Andrada Morar. “We don’t know how our kids’ career is gonna look like.” His answer to that unease is a line he repeats often. In times of anxiety, action is the answer. But the action he prescribes is not the wholesale replacement of human workers with AI. Instead, he urges companies to think carefully about what they stand to lose when they fire people in a hurry.
No Einstein in the data centre
Dines is impatient with the grandest promise of the moment. Some industry leaders talk of unleashing “50 million Einsteins in the data centre.” Dines thinks that image is only half right. A large language model, he argues, is an average of everything it has been trained on. “An average by definition doesn’t have a taste.” He tested this himself by asking models to write fiction in a given style. The results came back bland. Taste, he says, comes from lived experience, not from memory. He reaches for skiing to make the point: you can memorise every book ever written about the sport, but that will not make you a skier. You have to fall on the slope.
That gap between memorized knowledge and embodied experience matters inside a company. Every enterprise today runs the same handful of frontier models—GPT-4, Claude, Gemini—with the same weights. Feeding them different internal data does not give them a genuine grasp of your customer or your unique processes. “Our memory is not our identity,” Dines said. He believes that true understanding requires the kind of trial-and-error learning that only humans, and perhaps future artificial general intelligence, can perform.
Two ledgers, not one
His warning to executives is blunt. Do not read a job as a single output. Take a lawyer who reviews contracts. The visible outcome is a signed deal, and AI can speed that up. But the hidden outcomes are harder to measure. The same lawyer might mentor juniors, hold a client relationship together, or carry years of unwritten knowledge about the organization’s history and culture. Dines wants firms to keep two separate ledgers: one for visible outcomes and one for hidden ones. “Cut blindly, and you destroy value you never measured,” he says.
It is a pointed message from a man who sells automation. It also lands against a real-world backdrop. Carmakers have shed more than 20,000 white-collar jobs in recent months. A growing chorus of CEOs now pitch AI as a way to “do more with fewer people.” That marks a sharp reversal from two years ago, when most leaders emphasized that AI would augment, not replace, human workers. Dines himself has been navigating this shift. UiPath’s market capitalization, which once topped $35 billion, has fallen sharply as investors grew skeptical of the pure RPA play and demanded AI-driven growth. The company’s pivot to AI agents is an attempt to stay relevant, but Dines’s caution suggests he sees where the road of rash cuts leads.
He also thinks the transformation is slower than the hype suggests. AI agents cannot simply plug into messy, undocumented business processes. Most firms have never mapped who is allowed to approve an invoice, or how to pay one. That knowledge sits in people’s heads and across departments. Documenting it takes years, he says, not a weekend. “You can’t automate what you don’t understand,” Dines remarked. This is a reality that many software vendors gloss over when selling the promise of instant efficiency.
The identity problem
The deepest worry in the conversation is not about tasks. It is about identity. Dines traces his interest in the subject to a conversation with a lawyer friend. She told him her fear was not that her job would vanish. It was that her identity would become irrelevant. Many people build a sense of self around their work. He calls protecting that a shared human interest, and frames the human cost as the thing that enterprises risk losing when they treat workers as interchangeable cogs.
Dines is unconvinced that AI will develop a self of its own. To him, it is a tool—closer to electricity than to a colleague. He borrows an idea from the American philosopher Harry Frankfurt: there are two orders of will. A model can want something (first-order desire). Only a person can want to want something—to want to become better (second-order volition). Chasing a machine that truly reasons, Dines adds, would mean finding a way to inject pain into the system, and risk building a Frankenstein that no one understands. He questions whether society is ready for that.
Curiosity over credentials
Morar picked up the human thread. Models have memory, she said, but they lack the motivation to be excellent. AI can hand you knowledge, but it cannot hand you curiosity or the grit to push through when something breaks. She looks for those traits in her own team. She also argues that companies must still hire and mentor junior staff. Skip that, and there will be no senior leaders in a few years. “If you don’t invest in the next generation, you are eating your own seed corn,” she said.
There is a customer angle too. So much support has moved to chatbots that people now jab at their phones asking for a human. That friction, she suggests, is a clue about what only people offer: empathy, judgment, and the ability to navigate nuance. Dines agreed, noting that the most successful deployments of his company’s technology have been those where automation freed humans to do more meaningful work, not where humans were eliminated entirely.
None of this is disinterested. UiPath sells the agents and robots that make the cuts possible. A message that transformation is long, careful, and human-heavy also happens to describe a long, expensive engagement with clients. Yet coming from an automation billionaire, the caution is worth hearing. Governments are already counting the number of jobs AI touches. Dines’s bet is that the roles left standing will be richer, not poorer. The anxiety, he says, including his own, is the price of not yet knowing what the future of work will look like.
To understand where Dines is coming from, it helps to look at his own journey. Born in Romania, he started coding as a child on a Commodore 64. After studying mathematics and computer science, he worked for Microsoft in the early 2000s, where he experienced the first dot-com bust. He founded UiPath in 2005, originally as a consulting firm called DeskOver. The company pivoted to RPA in 2015 after noticing that clients kept asking for tools to automate repetitive desktop tasks. UiPath’s growth exploded, making Dines a billionaire and putting Bucharest on the global tech map. He has since stepped down as CEO but remains chairman and chief product visionary.
The broader context of Dines’s warning is important. The RPA industry, once hyped as a $30 billion market, has matured. Competitors like Automation Anywhere and Microsoft’s Power Automate have emerged. AI agents represent the next frontier, but they come with risks. If companies automate too aggressively, they may face not only hidden value loss but also public backlash and regulatory scrutiny. The European Union’s AI Act, which classifies high-risk systems, could impose obligations on employers that use AI for hiring, firing, or performance monitoring. Dines’s advice to “keep two ledgers” aligns with this emerging governance framework.
Perhaps the most telling part of the podcast was Dines’s admission of his own anxiety. He is a technologist who has spent his career making software that replaces human labor. Yet he worries about his children’s careers and the societal implications of the tools he sells. That paradox may be the real story here: even the creators of disruptive technology cannot fully predict its consequences. The path forward, as UiPath’s podcast name suggests, is not a straight line to a fully automated workforce. It is a winding road where humans and machines co-evolve, and where the most valuable asset remains the messy, curious, pain-feeling human mind.