The race to redefine automation isn’t a sprint, it’s a decades-long endurance race full of early leads, chaotic spinouts and pre-race strategy being updated on the fly. Yiannis Broustas, vice president of Product Marketing at UiPath, is an AI pioneer who has led go-to-market efforts for established enterprises and guided startups in the technology’s early days. Broustas, who sat down with Automation Today at UiPath’s annual FUSION event in Las Vegas, believes that the automation industry’s latest obsession—agentic AI—isn’t a sudden revolution but the next logical lap in a race that began with robotic process automation (RPA).
Early in his career, Broustas had the opportunity to launch an AI startup focused on video-content personalization–years before Netflix and TikTok normalized it. In his mind, that company didn’t achieve the success and scale it could, primarily because they built what they thought was a great product without making real customer pain points the primary consideration. That focus on building solutions aligned with real customer pain points taught him a valuable lesson, one he’s carried into every role since, including his tenure at Celonis (where he helped introduce AI capabilities to process mining) before joining UiPath just ahead of its IPO.
“When I joined UiPath in 2021, the company was evolving from an RPA vendor into a full automation platform,” he explains. “My mandate was to help build what came next—to extend automation beyond robots.”
From Democratization to Orchestration
For Broustas, RPA democratized automation by making it accessible to non-technical users.
“Before RPA, process automation was largely an IT exercise,” he says. “Scripts became tasks, and tasks became workflows. Suddenly, business users could automate parts of their day.”
That accessibility, he believes, was transformative but also limiting. Quick wins are easy to find with RPA, he notes, but the real challenge is scaling a strategic automation program. UiPath’s evolution mirrored that shift, moving from enabling isolated tasks to enterprise-wide impact.
“As customers matured, they needed more than bots,” Broustas explains. “We added discovery tools like process and task mining, intelligent document processing (IDP) to extract data from unstructured sources, API and test automation, and finally orchestration to tie it all together.”
And, the lesson Broustas learned in his entrepreneurial past—ensure your product is always focused on your customers and their pain points—echoes in UiPath’s C-suite as the company continues to compete for its leadership position in the automation race.
“Ultimately, customers want their problems solved,” Ashim Gupta, UiPath’s COO and CFO, tells Automation Today in a separate interview. “What you call AI, what you call deterministic automation, what you call old or new—it’s all about delivering outcomes. Our platform is differentiated because it incorporates all those elements and orchestrates them to solve real customer problems.”
The company’s focus, Broustas adds, has been to build an intelligent automation platform that integrates deterministic workflows (the rules-based processes bots handle best) with the adaptive, probabilistic power of AI.
“Every workflow has both parts,” he explains. “Rules where precision matters—and judgment calls where flexibility matters. Agentic AI completes that puzzle.”
Bridging the Gap Between Legacy Tools and Next-Gen AI
If automation’s first act was about democratizing technology, Broustas believes its second is about re-humanizing it. “AI is not here to replace anybody,” he stresses. “But someone who uses AI better than you might.”
He draws a historical parallel: “People once feared cars because they were dangerous. Yes, there were risks—but the societal benefits were enormous. The same is true here.”
Executives’ hesitation, he thinks, stems from a familiar tension: demanding more output with fewer resources while worrying that automation could diminish human control.
“It’s always the same people who say, ‘Do 20 percent more this year, but I’m not adding headcount,’” Broustas says. “Agentic AI is precisely what can help achieve that—if implemented responsibly.”
He acknowledges, though, that trust remains a barrier.
“With any powerful technology, people want to know they’re in charge,” he explains. “That’s why orchestration is so critical. Robots and agents can do incredible things—but the value only materializes when they’re coordinated toward business outcomes.”
Asked how enterprises can bridge the gap between proven technologies like IDP and newer concepts such as agentic AI, Broustas offers a pragmatic answer. “You bridge it when value becomes visible,” he says. “When a team lead can see their output improving because robots and agents are working together—and when they still feel in control.”
To Broustas, this orchestration-first approach separates hype from substance and differentiates UiPath from its competitors. To Gupta, it’s a force multiplier that enables businesses to leverage the most advanced technology alongside more foundational tech they know still works, while getting the most out of their total investment.
“The theme for us is where agentic AI meets ROI,” says Gupta. “Customers don’t care whether the solution is AI, RPA, or IDP—they care that it drives productivity, saves hours, and improves quality. Everything we build, from deterministic automation to large-language-model enhancements, is about that connection between investment and return.”
Learning from the “First Lap” of AI
Broustas is frank about the industry’s recent turbulence. “Over the past two years, a lot of organizations put the cart before the horse,” he says. “Executives rushed to experiment with generative AI without clear goals. Budgets flowed freely, and outcomes were uneven.”
He doesn’t see that phase as failure, though. “We’re not burned,” he insists. “We’ve just moved from exploration to execution. We now know which five percent of use cases deliver real value—and that knowledge lets us scale confidently.”
Comparing the period to the early 2000s tech bubble, Broustas sees parallels but also progress. “Think of the dot-com crash,” he says. “After the noise settled, the survivors—Google, Amazon—built sustainable models. The same will happen here. We’re now finding product-market fit for AI.”
