The tech world is awash in the promise of agentic AI. The prevailing narrative is one of disruption, where autonomous agents will rewrite the rules of business, moving faster and smarter than any human-led process ever could. This vision of radical, uninhibited transformation is celebrated in boardrooms and pitch decks across Silicon Valley.
But what happens when this fast-and-loose technology collides with the unforgiving reality of healthcare and life sciences rooted in legacy systems? In an industry where a misplaced decimal point can trigger a health crisis and a compliance failure can cost billions, the "move fast and break things" ethos doesn't hold. Here, the rules are not suggestions; they are immutable laws governed by patient safety and product quality.
This is the world of Joe Miles, Industry Head of Life Sciences at UiPath. With a career forged over 18 years at SAP and a tenure leading healthcare and life sciences at Google Cloud, Miles offers a perspective that inverts the typical AI narrative. He argues that for AI to succeed in the most critical industries, it isn't breaking the old rules; it's finally being forced to grow up and meet them.
A familiar pattern: "I think there's a lot of parallels to what happened in the on-premise ERP world," he says. To understand how agentic AI can possibly function within these constraints, Miles suggests looking to the past. The current moment, he argues, is not chaos but a new chapter in a familiar story of enterprise evolution. Just as the cloud offered a new layer of efficiency and compliance for regulated industries, agents are now providing a new layer of intelligence.
Flexible intelligence: This is made possible by what he calls an "agnostic" approach, a necessity driven by a messy, inconvenient truth about enterprise IT. "The reality is the vast majority of those offering enterprise AI solutions today aren't as connected as they need to be," Miles notes. "We are a long way away from that. But the best AI agents can be designed to work across fragmented systems, providing a flexible intelligence layer that can orchestrate processes across disconnected systems."
Miles' core mantra for this new era is "agents think, robots act, and people lead." Miles offers a clear distinction: "An agent has agency, autonomy, and a much more goals-based approach to its role, as opposed to a robot, which is much more deterministic and rules-based in its decision-making."
Finding where to deploy this powerful combination requires a deliberate strategy. While many companies start by building on existing automations, Miles pushes his clients to think bigger.
Beyond low-hanging fruit: "My advice is to look for areas that are fairly strategic, with a lot of manual processes, where we can have a significant impact," he said. "Let's look for somewhere we can drive a lot of value, not necessarily just capturing the lowest-hanging fruit."
As he looks to the future, Miles doesn't see this acceleration leading to the rise of new, monolithic giants. Instead, he sees a different pattern emerging. "I see it as very difficult to envision where the giants like Epic or SAP will come again," he argues. "I'm always amazed at the depth—this 'inch wide and a mile deep' perspective—where so many companies are doing so much with very specific functional enhancements."
In this new era, the greatest challenge for leaders will shift from technical execution to strategic wisdom. "Technical teams can always do the development," Miles concludes. "But it's understanding what to develop that really becomes a difficult challenge."
